Materials and Methods

Isolation of bovine granulocytes and production of conditioned media—Whole blood from healthy cows (450 mL) was obtained by jugular venipuncture into 450-mL containers preloaded with anticoagulant citrate dextrose solution (anticoagulant-to-blood ratio of 1:9). The blood was centrifuged at 1,000 × g for 20 minutes and the plasma removed. The remaining buffy coat that contained residual RBCs was treated with sterile distilled water for 30 seconds to lyse RBCs.²⁷ Tonicity was restored by use of 2.7% saline solution that contained 10mM phosphate buffer with 2% BSA.²⁷ The suspension was diluted to 50 mL with PBS solution and recentrifuged at 400 × g for 8 minutes. Cell pellets were washed in PBS solution, and any contaminating RBCs were removed by 1 additional cycle of hypotonic lysis.²⁷ White blood cells were counted, and cell differential was performed to determine the percentage of granulocytes. On average, the isolation procedure resulted in approximately 2 × 10⁹ viable cells, of which approximately 40% were granulocytes. Cell pellets were resuspended in 200 mL of Dulbecco modified Eagle medium without fetal bovine serum and stimulated with PMA (20 ng/mL). Cells were incubated at 37°C in a water bath for 30 minutes with occasional stirring, after which time the cells were centrifuged and the conditioned media harvested and buffered with 20mM Tris-HCl buffer (pH, 7.5) and stored frozen (–70°C) until protein purification studies were performed. Previous data indicate that bovine mononuclear cells and alveolar macrophages cultured in vitro require induction of MMP-9 gene expression, which is maximal at 24 hours of incubation.² We therefore ignored the potential contribution of monocytes and lymphocytes in our incubations lasting 30 minutes. For preparative purification of MMP-9 molecular species, several aliquots of conditioned media (400 to 4,000 mL) were used.

To investigate the time course of MMP-9 release from neutrophils into culture medium after phorbol ester stimulation, aliquots of cells and conditioned media were removed from the incubation after 10 to 60 minutes (1 to 24 hours after induction). The cell suspension was centrifuged to pellet the cells, and cell lysates and supernatants were subjected to zymographic evaluation of MMP-9.

Sera of cows with inflammatory disease—Cows admitted to the Food Animal Clinic at the Ohio State University and with septic conditions were bled (30 mL) immediately prior to humane euthanasia by pentobarbital overdose (90 mg/kg, rapidly IV). Cows were evaluated clinically and diagnostics performed as part of a plan developed after discussion with the owner. Client consent to collect blood to obtain serum samples (30 mL) was provided, and the protocol for blood collection was approved by The Ohio State University Veterinary Teaching Hospital executive committee and the Institutional Laboratory Animal Care and Use Committee of The Ohio State University. Diseases diagnosed were chronic pyogenic mastitis (n = 1), chronic pyogenic pneumonia (1), acute diffuse septic peritonitis (4), and severe acute myositis (1). From clinically normal cows, sera and pooled sera were also obtained from a commercial source.^a

Purification of molecular species of MMP-9 released by granulocytes—Purification of MMP-9 was based on a method described by Goldberg et al,²⁸ with addition of a gel filtration step. Serum-free conditioned Dulbecco modified Eagle medium from PMA-stimulated bovine granulocytes (400 to 4,000 mL) was buffered with Tris-HCl buffer (pH, 7.5) and applied to a reactive red-120 agarose column^b (1 mL of resin/100 mL conditioned medium), equilibrated with 20mM Tris-HCl buffer (pH, 7.5) that contained 150mM NaCl and 2mM CaCl₂. After washing the column with the same buffer, bound proteins were eluted with 20mM Tris-HCl buffer (pH, 7.5) that contained 2M NaCl and 2mM CaCl₂. Material eluted from this column was dialyzed against 20mM Tris-HCl buffer (pH, 7.5) that contained 0.5M NaCl, 2mM CaCl₂, and the nonionic surfactant 0.01% Brij 35 and was applied on a gelatin-agarose^c column (1 to 3 mL). After washing, the gelatin-agarose column was eluted with 20 volumes of a 0% to 7% DMSO gradient. Elution was monitored by UV absorption at a wavelength of 280 nm, and peaks were collected, concentrated,^d and analyzed by gradient zymography, western blot, and protein staining. Total yield of DMSO eluted material was approximately 100 μg of protein/L of conditioned media. For final purification, concentrated peaks were subjected to gel filtration^e on a 1 × 95-cm column equilibrated with 20mM Tris-HCl buffer that contained 200mM NaCl, 2mM CaCl₂, and 0.01% Brij 35. Elution was monitored by UV absorption, and peaks were collected, concentrated, and stored in 5mM Tris-HCl buffer (pH, 7.5), 0.5mM CaCl₂, and 0.01% Brij 35 buffer at −80°C.

Purification of molecular species of MMP-9 from bovine sera—Twenty-five to fifty percent of the ammonium sulfate saturation fraction of sera (30 mL) was dialyzed against 20mM Tris-HCl buffer (pH, 7.5) that contained 0.5M NaCl, 2mM CaCl₂, and 0.01% Brij 35. Sera were fractionated by gelatin-agarose affinity chromatography as described for conditioned media.

Hemoglobin affinity chromatography—The procedure used was based on published methods with modification.²⁹ Briefly, 150 mg of bovine hemoglobin^f in 50mM HEPES buffer (pH, 7.3) was coupled to 10 mL of affinity media^g over 90 minutes at room temperature (approx 22°C) according to the manufacturers instructions. After blocking of residual active esters, the hemoglobin-affinity-gel column was equilibrated with TBS containing 0.01% Brij 35. Fifty micrograms of MMP-9-Hp eluted from gelatin-agarose at 3.5% DMSO and dialyzed against TBS that contained 0.01% Brij was applied to 1.5 mL of a hemoglobin-affinity-gel column. After washing, bound material was eluted by a 0M to 6M linear gradient of urea in 150mM NaCl (pH, 11). Fractions of 0.5 mL were collected in tubes that contained 50 μL of 1M Tris-HCl buffer (pH, 7.5). Elution was monitored by UV absorption at a wavelength of 280 nm. Pooled peaks were dialyzed, concentrated, and analyzed by discontinuous gradient zymography.

Discontinuous gradient zymography—Three layers of 7%, 5%, and 4% acrylamide that contained all components required for acrylamide polymerization were consecutively poured in gel casting cassettes.^h To avoid convection, the 7% mix contained 60% sucrose instead of water, and the 5% mix contained 40% sucrose instead of water. Concentrations of gelatin in 7%, 5%, and 4% layers were 0.1%, 0.2%, and 0.25%, respectively. Zymography gels were run in the presence of SDS and developed as described.³⁰

Western blot analysis—Proteins were electrophoresed through prepoured SDS-PAGE gels and electrophoretically transferred to nitrocellulose membranes.^31,32 After washing in TBS that contained 0.5% Tween-20, the membranes were blocked with TBS-Tween-20 that contained 5% nonfat dry milk powder.² Goat anti-human MMP-9,ⁱ goat anti-human MMP-9,^j rabbit anti-human MMP-9,^k and goat anti-bovine haptoglobin^l antibodies were used at a concentration of 1 μg/mL as primary antibodies for western blot analysis.² Horseradish peroxidase–conjugated goat anti-rabbit antibody^m at a dilution of 1:1,000 and horseradish peroxidase–conjugated donkey anti-goat antibodyⁿ at a dilution of 1:5,000 were used as secondary antibodies. Labeled protein was detected by use of enhanced chemiluminescence.^o

SDS-PAGE band observation and protein determination assays—Routine Coomassie blue and silver staining^p were used for PAGE. A protein ladder q and standards^r were used as molecular mass markers. Three- to eight-percent Trisacetate, 14% Trisglycine, and 4% to 12% Trisglycine gels^s were used. Protein concentrations in samples were measured by use of UV absorption or a commercially available protein quantitation kit^t or by scanning gels and quantitation of PAGE band intensities in precalibrated linear range of density with imaging software.^u

MMP-9 activation—Bovine MMP-9 molecular species were activated with activated stromelysin^v in stromelysin–to–MMP-9 molar ratios of 1:15 to 1:30 during 3 to 8 hours at 37°C. Stromelysin activation was achieved by incubation with tosyl phenylalanyl chloromethyl ketone–treated trypsin^w at a trypsin-to-stromelysin molar ratio of 1:15 over 20 to 40 minutes at room temperature. Trypsin activation was stopped by addition of a 4-fold molar excess of soybean trypsin inhibitor.^x Activation process (disappearance of proforms and appearance of N-terminal truncated active species) was monitored by SDS-PAGE.

Assays of enzymatic activity of activated MMP-9—Assays were performed in a concentration range of 10 to 100 ng of enzyme/mL. Concentrations of proforms and activated forms of MMP-9-Hp in the assay were calculated based on the quantity of MMP-9 released from complexes at the time of dithiothreitol reduction. The reaction mixture contained a fluorescein conjugate (12.5 μg/mL),^y and assays were performed between 5 and 60 minutes at room temperature. Fluorescence was measured with a microplate reader set for excitation at a wavelength of 495 nm and emission detection at a wavelength of 530 nm. For each experimental point, values for spontaneous degradation of the fluorescein conjugate (sample that contained no enzyme) were subtracted.

Mass spectral peptide sequencing—Purified MMP-9-Hp were reduced with 20mM dithiothreitol and run on a 14% SDS-PAGE gel. The gel was stained with Coomassie blue, and protein bands were submitted for peptide sequence analysis by use of capillary liquid chromatography-nanospray tandem mass spectrometry. Briefly, gels were digested with sequencing grade trypsin^z with a commercially available digestion kit^aa following the manufacturer's recommended protocols. Bands were trimmed of excess polyacrylamide. Gels were dried with acetonitrile and reconstituted with 75 μL of dithiothreitol (32.5mM dithiothreitol) in 100mM ammonium bicarbonate solution. Iodoacetamide (75 μL of 80mM) dissolved in 100mM ammonium bicarbonate was added to alkylate cysteine residues, and the gel was washed again with sequential cycles of acetonitrile and ammonium bicarbonate. Trypsin was added and the digestion conducted at room temperature overnight. Peptides were extracted several times from the polyacrylamide gel with 50% acetonitrile and 5% formic acid and pooled. The pooled extracts were concentrated under a vacuum to approximately 25 μL.

Protein identification—Capillary liquid chromatography–nanospray tandem mass spectrometry was performed on a mass spectrometer^bb equipped with a nanospray source operated in positive ion mode. The chromatographic system^cc had an autosampler and column switcher. Solvent A was water that contained 50mM acetic acid, and solvent B was acetonitrile. Five microliters of each sample was first injected on to a trapping column and washed with 50mM acetic acid. The injector port was switched to inject, and the peptides were eluted off the trap onto the column. A C18 column^dd (5 cm long, 75 μm internal diameter) packed directly in the nanospray tip was used for chromatographic separations. Peptides were eluted directly off the column into the LTQ system by use of a gradient of 2% to 80% of solvent B over 30 minutes, with a flow rate of 300 nL/min and a total run time of 58 minutes. The scan sequence of the mass spectrometer was based on a triple play method.³³ The analysis was programmed for a full scan to determine the charge of the peptide, and double mass spectrometry was used to generate the product ion spectra. The amino acid sequence in consecutive instrument scans of the most abundant peak in the spectrum was determined. Dynamic exclusion was used to exclude multiple double mass spectrometry of the same peptide.

Database searches—Sequence information from the double mass spectrometry data was processed to generate a peak list file by use of a software program.^ee Data processing was performed following established guidelines.³⁴ Assigned peaks have a minimum of 10 counts (signal-to-noise ratio of 3). The mass accuracy of the precursor ions was set to 1.5 days to accommodate accidental selection of the C13 ion, and the fragment mass accuracy was set to 0.5 days. Considered modifications (variable) included methionine oxidation and carbamidomethyl cysteine. Protein identifications were checked manually, and proteins with a Mascot score of ≥ 40 with at least 2 peptides having a y or b ion sequence tag of ≥ 3 residues were accepted.³⁴

Results

The time course of MMP-9 release from neutrophils into culture medium after phorbol ester stimulation—Matrix metalloproteinase-9 was present within the cellular homogenate prior to induction but was not detectable at 30 minutes. Only trace amounts of MMP-9 were detectable in serum-free culture media that contained WBCs prior to addition of phorbol ester. After addition of phorbol ester (20 ng/mL), the amount of MMP-9 in the media increased rapidly at 10 minutes and was maximal at 30 minutes (data not shown).

Molecular species of MMP-9 released by granulocytes—Cell media conditioned with bovine granulocytes and stimulated with PMA were initially subjected to affinity chromatography on gelatin-agarose columns, and peaks were eluted with a 0% to 8% DMSO gradient and analyzed by SDS-PAGE gelatin zymography and western blot analysis. Typically, 7 to 10 peaks identified by UV absorption were observed in chromatographic elution profiles in each experiment (Figure 1). Monomeric (105 kd) and dimeric (210 kd) forms of pro-MMP-9 are major bands on zymograms of fractions eluted at 5% and 6.5% DMSO, respectively. In addition, several gelatinolytic bands with molecular masses in the range of 300 to 500 kd and > 500 kd were observed in granulocyte-conditioned media. A ladder of gelatinolytic bands with molecular mass between 300 kd and 500 kd and > 500 kd eluted early in the gradient (approx 3.5% DMSO), demonstrating reduced binding affinity for gelatin. Two other pronounced high–molecular mass gelatinolytic bands were eluted with higher concentrations of DMSO. The bulk of gelatinolytic activity with molecular masses of approximately 300 kd coeluted with monomer and a major 500-kd gelatinolytic band coeluted with dimer forms of bovine MMP-9. Silver staining of the same material demonstrated a pattern similar to that observed by zymography, although ratios of band intensity revealed by zymography and by silver staining varied.

View Full Size

Molecular species of MMP-9 released by PMA-stimulated bovine granulocytes as evaluated on a gelatin zymogram (A), western immunoblot with anti-human MMP-9 polyclonal antibody AF911 (B), silver-stained electrophoretogram (C), and western immunoblot with anti-human MMP-9 polyclonal antibody AB805 (D). By use of chromatography, peaks were eluted with a 0% to 8% DMSO gradient. Lanes 1 to 6 in panels A and C correspond to eluted peaks 4 to 9. Peaks 4, 6, and 8 were eluted with 3.5%, 5%, and 6.5% of DMSO, respectively. Lanes 1 to 10 in panels B and D correspond to eluted peaks 4 to 8; lanes 1, 3, 5, 7, and 9 are unreduced samples, and lanes 2, 4, 6, 8, and 10 are samples reduced with 20mM dithiothreitol. Positions of molecular mass markers at the left side of gels are in kilodaltons. *Position of a monomer of bovine MMP-9. **Position of a dimer of bovine MMP-9.

Citation: American Journal of Veterinary Research 68, 9; 10.2460/ajvr.68.9.995

• Download Figure
• Download figure as PowerPoint slide

Antibody AF911 (antisera to whole molecule of MMP-9; Figure 1), AB805 (N-terminal region of active enzyme), and C20 (peptide mapped to C-terminus of MMP-9; not shown) recognized monomeric and dimeric forms of bovine MMP-9. Antibody AF911 also identified most of the high–molecular mass complexes expressing gelatinolytic activity in zymograms. Antibodies AB805 and C20 were poorly reactive against complexes with molecular mass > 300 kd (results not shown for C20). The amount of gelatinolytic and MMP-9 immunoreactive material with molecular masses greater than MMP-9 dimer comprised approximately 10% to 30% of total MMP-9 released. After sample reduction with dithiothreitol, MMP-9 containing high–molecular mass bands of all peaks collapsed to monomeric MMP-9 and was detected with all 3 antibodies.

Identification of components of high– molecular mass MMP-9 complexes—For identification of components of MMP-9 complexes, peaks eluting from gelatin-agarose at a low DMSO concentration containing multiple gelatinolytic bands of high molecular mass were further purified by gel filtration chromatography. All bands representing high–molecular mass complexes were eluted as a single peak in the void volume of the column. Unreduced SDS-PAGE of purified material had 3 bands with apparent molecular masses of 320 kd, 360 kd, and 450 kd and at least 4 bands with molecular masses between 600 and 1,000 kd (Figure 2). Reduction of samples with dithiothreitol led to complete conversion of high–molecular mass bands to 3 major bands detected at 105 kd, 38 kd, and 18 kd. Densitometry scans revealed an approximate molar ratio of these 3 bands to be 1:13:13 (MMP-9, β-haptoglobin, α-haptoglobin, respectively). A minor band with a molecular mass of approximately 58 kd was occasionally observed and was unidentified. The protein band at 105 kd migrated with purified monomeric bovine MMP-9.

View Full Size

Purification of high–molecular mass MMP-9-Hp (peak eluted with 3.5% DMSO) and identification of its components before and after reduction with 20mM dithiothreitol (A; lane 1, silver-stained electrophoretogram before reduction; lane 2, silver-stained electrophoretogram after reduction; lane 3, gelatin zymogram; and lane 4, western immunoblot with anti-bovine haptoglobin antibody GHPT-10A). Silver-stained electrophoretograms with 3% to 8% (B) and 14% (C) gels were evaluated (lane 1, unreduced purified complexes; lane 2, dithiothreitol-reduced purified complexes; lane 3, unreduced mixture of monomer and dimer forms of bovine MMP-9; and lane 4, dithiothreitol-reduced mixture of monomer and dimer forms of bovine MMP-9). See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 68, 9; 10.2460/ajvr.68.9.995

• Download Figure
• Download figure as PowerPoint slide

Mass spectral analysis of the peptide digests of bands of 105 kd, 38 kd, and 18 kd identified them as bovine MMP-9, β-chain of bovine haptoglobin, and α-chain of bovine haptoglobin, respectively (MMP-9 > 99% similarity, 20.4% of protein; β-haptoglobin > 99% similarity, 27% of protein; α-haptoglobin > 99% similarity, 40% of protein). Western blot analysis with antibovine haptoglobin antibody confirmed the presence of haptoglobin in high–molecular mass complexes (Figure 2).

Enzymatic properties of MMP-9-Hp—Incubation of bovine MMP-9-Hp with activated stromelysin neither altered mobility of haptoglobin complexes nor produced any new protein bands below 300 kd in nonreduced gels. However, upon reduction with dithiothreitol, truncated forms of MMP-9 with molecular masses of approximately 10 kd lower than the proenzyme were detected (Figure 3). Electrophoretic mobility of this activated form of MMP-9 originating from the haptoglobin complexes was identical to the mobility of the truncated form of haptoglobin-free MMP-9 activated with stromelysin. Both haptoglobin chains were not susceptible to cleavage by activated stromelysin (not shown).

View Full Size

Activation of bovine haptoglobin-free MMP-9 and high–molecular mass MMP-9-Hp with stromelysin evaluated on a silver-stained electrophoretogram with a 14% gel (lane 1, MMP-9 from complex not treated with stromelysin; lane 2, MMP-9 from complex treated with stromelysin; lane 3, monomeric untreated MMP-9; lane 4, monomeric MMP-9 treated with stromelysin). See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 68, 9; 10.2460/ajvr.68.9.995

• Download Figure
• Download figure as PowerPoint slide

Gelatinolytic activity of pro-MMP-9 within haptoglobin complexes and haptoglobin-free pro-MMP-9 was inactive in this assay. Activated MMP-9-Hp had the same time-curve of activity and similar specific activity as activated haptoglobin-free MMP-9 (Figure 4). Activity of activated MMP-9-Hp was linear and dependent upon concentration of MMP-9.

View Full Size

Kinetics of enzymatic activity of stromelysin-activated MMP-9-Hp based on relative fluorescence intensity versus time. Graph represents activated free monomer at 30 ng/mL (solid line, filled circle), nonactivated complex at 100 ng/mL (small dotted line, square), and activated complex at 100 ng/mL (dashed line, upward triangle), 50 ng/mL (dot-dash line, downward triangle), and 13 ng/mL (dashed line, box with crosshair symbol). Inset represents relative fluorescence intensity versus enzyme complex concentration.

Citation: American Journal of Veterinary Research 68, 9; 10.2460/ajvr.68.9.995

• Download Figure
• Download figure as PowerPoint slide

Binding of MMP-9-Hp to hemoglobin—To determine whether MMP-9-Hp retain the ability to bind to hemoglobin, column fractions eluted with 1M and 2.8M urea were analyzed by zymography. Fractions contained MMP-9-Hp with identical patterns of high–molecular mass gelatinolytic bands in the region between 300 kd to 1,000 kd (Figure 5). These results indicate that MMP-9-Hp are capable of binding hemoglobin.

View Full Size

Binding of MMP-9-Hp to hemoglobin (A) as evaluated by discontinuous gradient zymography (lane 1, column load; lane 2, material passed through column; lane 3, peak eluted with 1M urea; and lane 4, peak eluted with 2.8M urea). Comparison of MMP-9-Hp in sera of clinically normal cows to cows with inflammation (B) by SDS-PAGE gelatin zymography and western blot analysis. Sera originated from a cow with acute diffuse peritonitis (lane 1, anti-haptoglobin western immunoblot; lane 2, 4%, 5%, and 7% discontinuous gradient gelatin zymogram), cow with chronic pneumonia (lane 3, 4%, 5%, and 7% discontinuous gradient gelatin zymogram), and clinically normal cattle (pooled sera; lane 4, 4%, 5%, and 7% discontinuous gradient gelatin zymogram). See Figure 1 for remainder of key.

Citation: American Journal of Veterinary Research 68, 9; 10.2460/ajvr.68.9.995

• Download Figure
• Download figure as PowerPoint slide

MMP-9-Hp in sera of cows with inflammation—To determine the presence of MMP-9-Hp in sera of clinically ill and healthy cows, serum proteins were subjected to gelatin agarose affinity chromatography. Chromatographic profiles of bovine serum proteins eluted from gelatin agarose with DMSO gradients consisted of 4 to 5 major peaks and some minor peaks that were pooled and analyzed by discontinuous gradient zymography and western blot analysis. In a cow with severe acute septic myositis (n = 1) and cows with acute diffuse peritonitis (4), zymography profiles of fractions eluted from gelatin agarose were similar to those of media conditioned by bovine granulocytes stimulated with PMA. Specifically, some gelatinolytic bands between 300 and 500 kd and a series of gelatinolytic bands > 500 kd were observed (Figures 1 and 5). High–molecular mass activity on zymograms indicated that sera obtained during acute inflammation comprised approximately 10% to 15% of total gelatinolytic material (not shown). Immunoblots with anti-haptoglobin and anti-MMP-9 antibodies confirmed identity of these gelatinolytic bands as MMP-9-Hp. No high–molecular weight gelatinolytic bands were found on zymograms of fractions eluted from gelatin agarose when sera of cows with chronic pyogenic pneumonia or chronic mastitis were used or when pooled sera or individual serum samples of clinically normal cows were used.

Sera of sick cows had a more complex protein composition of MMP-9-Hp; results of SDS-PAGE of bulk affinity purified bovine haptoglobin under nonreducing conditions reveal a series of proteins with molecular masses of > 240 kd and substantial unresolved protein material near the top of gels (Figure 6). Reduction of the sample with dithiothreitol converted high–molecular mass bands into 3 bands with molecular masses of 18 kd, 38 kd, and 68 kd. These proteins were confirmed to be α- and β-chains of haptoglobin and BSA, respectively, by peptide mass spectrometry.

View Full Size

Comparison of nonreduced (A) and reduced (B) samples of purified bovine haptoglobin (lane 2, panel A; lane 3, panel B) and purified haptoglobin subfraction (lane 3, panel A; lane 2, panel B) by SDS-PAGE. Bands 1, 2, 3, 4, 5, and 6 in lanes 2 and 3 of reduced samples (B) are MMP-9, immunoglobulin heavy chain, BSA, β-chain of haptoglobin, immunoglobulin light chain, and α-chain of haptoglobin, respectively. Proteins were identified by mass spectral peptide sequencing. Coomassie blue staining (lane 1 of panels A and B)molecular mass markers in kilodaltons.

Citation: American Journal of Veterinary Research 68, 9; 10.2460/ajvr.68.9.995

• Download Figure
• Download figure as PowerPoint slide

For analysis of protein composition of bovine haptoglobin subfractions that had gelatin-binding properties, 1 mg of a nonreduced haptoglobin preparation was applied to a 2-mL gelatin agarose column and eluted with a linear DMSO gradient. Two peaks were observed; 1 peak was observed at a DMSO concentration of 1.8% (peak 1) and another peak at a DMSO concentration of 3.4% (peak 2). Calculations made based on quantitation of the UV absorption at a wavelength of 280 nm of eluted peaks revealed that the amount of bound and eluted protein was equivalent to approximately 1% of total haptoglobin loaded. Results of SDS-PAGE analysis of the eluted gelatin agarose peaks in nonreduced conditions revealed high–molecular mass band patterns characteristic for haptoglobin (Figure 6). Reduction of this material with dithiothreitol converted these high–molecular mass bands into free α- and β-chains of haptoglobin and a series of protein bands specific to each of these 2 fractions. Identities of these bands were determined by peptide mass spectrometry. In peak 2, eluted from gelatin agarose, besides α- and β-chains of haptoglobin, mass spectral peptide sequencing identified MMP-9, IgG heavy chain-like protein, IgG light chain protein, and BSA. In peak 1, no MMP-9 was found bound to haptoglobin. This peak comprised haptoglobin bound to fibronectin in addition to several other proteins (not shown).

Discussion

Despite cellular heterogeneity of WBC cultures that we used as a source of bovine MMP-9 molecular species, including MMP-9-Hp, we suggest that the MMP-9 molecules are neutrophil derived. Quick disappearance (within 30 minutes) of zymographically detected MMP-9 from cell lysates in conjunction with reciprocal accumulation of the enzyme in culture medium, described in our study, strongly argues in favor of releasing of presynthesized protein and against protein secretion upon gene induction. In other words, kinetics of the process correspond to kinetics of neutrophil degranulation, but not to kinetics of production of MMP-9 by other WBCs.³

We found that PMA stimulated bovine granulocytes, in addition to monomeric (105 kd) and dimeric (210 kd) forms of pro-MMP-9, and released a previously uncharacterized reduction-sensitive MMP-9-Hp with molecular mass ranging from 320 kd to 1,000 kd in the presence of 2% SDS. It has been reported that native bovine haptoglobin exists as a series of multimers with compositions of (α₂β₂)_n.³⁵ Assuming this model is correct, the first member of the MMP-9-Hp family described in our study, having a molecular mass of approximately 320 kd, may represent haptoglobin (α₂β₂)₂ in complex with 1 molecule of bovine pro-MMP-9 (105 kd). Because the typical molecular ratio of α-haptoglogin to β-haptoglobin to MMP-9, after complete reduction of complex, is 13:13:1, formation of higher–molecular mass members of the family probably occur mostly by increasing the number of α- and β-chain components rather than MMP-9 molecules within the complexes. Lower affinity of MMP-9-Hp for gelatin (elution with lower concentration of DMSO) associated with higher molecular mass is consistent with this hypothesis. Only quantitative determination of N-terminal sequences of each individual band would allow us to determine the precise stoichiometry of MMP-9-Hp. In the absence of SDS, all members of MMP-9-Hp, including molecules with apparent molecular mass of 320, 360, and 450 kd on SDS-PAGE, behave as molecules with molecular masses > 750 kd on gel filtration columns. These data suggest the formation of large aggregates of MMP-9-Hp in vivo. A minor band with molecular mass of approximately 58 kd, present in Coomassie blue–stained gels after dithiothreitol treatment of purified MMP-9-Hp, was not identified by peptide mass spectrometry, but may represent BSA in association with haptoglobin polymers as has been reported previously.³⁶

Results of our study indicate that the MMP-9 molecule found within MMP-9-Hp is in the proenzyme form and is enzymatically inactive toward the fluorescein conjugate^x in vitro. Thus the binding of MMP-9 to haptoglobin does not induce enzymatic activity of the proenzyme as has been observed in cases of binding pro-MMP-9 to gelatin and collagen type IV or to dentin matrix protein.^37,38

We have shown, however, that activated stromelysin converted the haptoglobin complexed pro-MMP-9 to active N-terminal truncated forms that are indistinguishable from activated form of free enzyme by either molecular size or enzymatic activity. These results support our contention that interactions of MMP-9 with haptoglobin within MMP-9-Hp do not alter its susceptibility to proteolytic attack by activated stromelysin and that without reduction of disulfide bridges the activated form of MMP-9 remains a constituent of the complexes. Whether MMP-9 in complex with haptoglobin acquires some particular properties in regard to substrate specificity remains to be determined. We have also found that MMP-9 that contained haptoglobin polymers retain their ability to bind hemoglobin so that MMP-9-Hp retain biochemical properties of both their protein constituents.

Formation of complexes of haptoglobin with MMP-9, BSA, and other proteins most probably occurs within cells prior to release, and we believe production, storage, or both at specific subcellular compartments is part of the mechanism of specificity of haptoglobin heteromeric protein complex formation. This suggestion is supported by findings in our laboratory of MMP-9-Hp inside granulocytes. We have also found that only a third of haptoglobin molecules (affinity purified with anti-BSA chromatography) of sera obtained during acute inflammation contain covalently bound serum albumin, with the remainder being BSA-free. It is hard to explain such asymmetry without an assumption that haptoglobin-BSA complexes form in specific cellular compartments, but not in plasma where haptoglobin and stoichiometrically excessive amounts of BSA are homogeneously distributed. In sera obtained during acute inflammation, beside MMP-9-Hp, free haptoglobin and haptoglobin-free MMP-9 are present in substantial concentration, suggesting complex formation outside intracellular compartments is a remote possibility.

We have found that high–molecular mass MMP-9-Hp, produced by stimulated granulocytes, are present as specific components of sera of cows with acute inflammatory conditions and are absent from sera of healthy cows. Identification of the molecular composition of the complexes was performed by mass spectral analysis of the peptide digests in addition to zymography and western blot analysis. Although MMP-9 and haptoglobin have been implicated in inflammation and specifically in acute phase of the inflammatory response, their functions are not fully understood.^18,19,39,40

Matrix metalloproteinase-9 is expressed by various cell types, playing important roles in the process of WBC extravasation and cell recruitment during inflammation, although some disagreement exists over the importance of these functions.^41–43 Proteolytic capabilities of MMP-9 toward cleavage of extracellular matrix proteins, growth factors, interleukins, interleukin receptors, and proteinase inhibitors are thought to be the molecular basis of MMP-9 biological activity, although substrates critical for enzyme function and precise mechanisms in most instances remain obscure. It has been found recently that human haptoglobin-free MMP-9 can bind to cell surface lipoprotein receptor–related protein 1 (CD91), megalin, and the DNA repair protein Ku via the hemopexin domain of the enzyme.^44,45 Formation of haptoglobin polymers harboring MMP-9 may alter targeting of the enzyme during the acute phase of the inflammatory response and result in its binding to CD163, the cellular receptor of haptoglobin.¹³

We have also found that MMP-9-Hp polymers in sera obtained from septic cattle with acute inflammation, in contrast to MMP-9-Hp in granulocyte-conditioned medium, contain an IgG-like component, which could target these complexes to other cellular receptors (eg, Fc receptor). Results of our study indicate that subfractions of haptoglobin molecules purified from sera of cows with acute inflammation contain disulfide bound MMP-9 along with other serum proteins that are not components of MMP-9-Hp released by granulocytes in culture. The presence of MMP-9-Hp in sera of cows with acute inflammation, but not in sera of cows with chronic disease or healthy cows, suggests that the formation of MMP-9-Hp is part of the acute inflammatory response of cattle.

Scavenging functions of haptoglobin in hemoglobin clearance from plasma by endocytosis of hemoglobin-haptoglobin complexes bound to macrophage CD163 is well described.¹³ Results of our study indicate that MMP-9-Hp retains the ability of haptoglobin to bind hemoglobin. In this regard, MMP-9-Hp formation may have a role in clearance of MMP-9 from the circulation. In addition to potential roles in scavenging, binding of hemoglobin-haptoglobin complexes to CD163 is suggested to play immunomodulatory roles by increasing the expression of heme oxygenases and secretion of IL-6 and IL-10.⁴⁶ Whether the presence of MMP-9, as a constituent of hemoglobin-haptoglobin complexes, influences alterations of gene expression remains to be elucidated. Physiologic roles of soluble forms of CD163, which are capable of binding hemoglobin-haptoglobin complexes and have plasma concentrations that increase manyfold during reactive hemophagocytic syndrome, sepsis, and some other conditions, is not yet understood.^47,48 The mechanism of CD163 ectodomain shedding is also unknown, but appears to involve activation of toll-like receptor 4, resulting in an innate immune response to extracellular pathogens.⁴⁹ As a component of hemoglobin-haptoglobin complexes, MMP-9 hypothetically may participate in the cleavage of the extracellular domain of CD163, leading to formation of soluble CD163.⁵⁰

The function of haptoglobin in bovine ovary and oviduct and rabbit oviduct and endometrium, where it was found in blastocyst extraembryonic matrix, is unknown^51,52 It is noteworthy that MMP-9 is also thought to be involved in implantation and trophoblast function.^53,54 It has been reported that the major matrix metalloproteinase activity in the preimplantation uterus of mice originated from pro-MMP-9 bearing neutrophils attracted by seminal plasma.⁵⁵ This finding indicates that mechanisms of delivering MMP-9 to target tissue during acute inflammation and implantation are virtually identical. It is plausible that MMP-9 and haptoglobin are physiologically active in reproduction in the same way as in acute inflammation, perhaps as MMP-9-Hp.

In conclusion, covalent MMP-9-Hp, described in our study, combine properties of both components. We suggest that multifunctionality of MMP-9-Hp is used to transport and target MMP-9 during the acute phase of the inflammatory response. The presence of neutrophil-derived haptoglobin in complex with MMP-9 may also provide information regarding neutrophil degranulation during acute inflammation. Assays specific for MMP-9 bound to haptoglobin may be more sensitive than total serum haptoglobin concentrations in the detection of acute inflammation in cattle.