active to anti-a2-macroglobulin also was detected. Its level in the aqueous humor was 3.2 ±. 1.3Mg/ml. Conclusions. The gelatinases, serine-like proteinases, ...
MMPs and Proteinase Inhibitors in the Human Aqueous Humor Hideo Ando,* Sally S. Twining,^ Beatrice Y. J. T. Yue,* Xiaoye Zhou,-\ M. Elizabeth Fini,% Tadayoshi Kaiya,% Eve]. Higginbotham,\\ and Joel Sugar*
Purpose. This study was performed to examine the gelatinolytic and caseinolytic activities and the levels of two proteinase inhibitors, a 1-proteinase inhibitor (al-antitrypsin) and «2-macroglobulin, in the human aqueous humor. Methods. Aqueous humor samples were collected during elective surgery in patients with cataracts. Zymography with gelatin- and casein-containing gels was performed. The inhibitors were examined by Western blot analyses, enzyme-linked immunosorbent assay, and dot blot assays. Results. The aqueous humor contained a major band of gelatinolytic activity at a molecular weight of 66 kD and minor bands at 125, 95, and 62 kD. These gelatinases were inhibited by 10 mM ethylenediaminetetraacetic acid (EDTA) or 1,10-phenanthroline. After extended incubation (48 hours), zymography on casein-containing gels showed proteinase bands with molecular weights in the 80- to 84-kD range. Additional bands at 68 and 48 kD also were observed. All the caseinase activities were inhibited by 10 mM phenylmethylsulfonyl fluoride and 1 Mg/m' aprotinin. No inhibition was observed with 5 mM EDTA, 5 /xM E-64, or 1 nM pepstatin. These results indicated that the caseinases are serine proteinases. Western blot analysis showed a 53-kD a\-proteinase inhibitor band in the aqueous humor. The concentration was 32.2 ± 9.9 Mg/ml, constituting approximately 15% of the total protein. A 360-kD protein band immunoreactive to anti-a2-macroglobulin also was detected. Its level in the aqueous humor was 3.2 ± 1.3Mg/ml. Conclusions. The gelatinases, serine-like proteinases, and proteinase inhibitors found in the aqueous humor may participate in the remodeling of extracellular matrices in the trabecular meshwork and other tissues bordering the anterior chamber. Invest Ophthalmol Vis Sci. 1993;34:3541-3548.
JVecent studies have shown the roles of a class of enzymes known as matrix metalloproteinases (MMPs) in connective tissue degradation and remodeling.1 The MMPs have been grouped into three subfamilies—collagenase, gelatinase, and stromelysin—based on the
From the * Department of Ophthalmology and Visual Sciences, University of Illinois at Chicago College of Medicine, Chicago, Illinois; ^Departments of Biochemistry and Ophthalmology, Medical College of Wisconsin, Milwaukee, Wisconsin; XCutaneous Biology Research Center, Massachusetts General Hospital, and Department of Dermatology, Harvard Medical School, Boston, Massachusetts; ^Department of Ophthalmology, Seirei Hamamatsu General Hospital, Hamamatsu, Japan; and ^Department of Ophthalmology, University of Michigan, Ann Arbor. Dr. Ando is currently affiliated with the Department of Ophthalmology, Niigata University, School of Medicine, Niigata, Japan. Supported by grants EY03890 (BYJTY), EY05628 (BYJTY),EY06663 (SST), and EY08408 (MEF) and core grant EYOI792 from the National Institutes of Health, Bethesda, Maryland. Dr. Hideo Ando received a fellowship from Kakegawa Eye Clinic, Shizuoka, Japan. Submitted for publication October 5, 1992; accepted May 28, 1993. Proprietary interest category: N. Reprint requests: Beatrice Y.J. T. Yue, 1855 W. Taylor Street, Chicago, IL 60612.
Investigative Ophthalmology & Visual Science, December 1993, Vol. 34, No. 13 Copyright © Association for Research in Vision and Ophthalmology
specificity for different extracellular matrix components.1 For instance, the interstitial collagenase, designated as MMP-1, degrades native collagen types I, II, and III.1 Stromelysin, MMP-3, cleaves proteoglycans, fibronectin, laminin, gelatins, and casein, whereas the 72-kD gelatinase, MMP-2, and the 92-kD gelatinase, MMP-9, show activities against gelatins of all types and native collagen types IV, V, and VII.1 The MMP family members share considerable sequence similarity, and the different proteinases appear to have evolved from a common ancestral gene. 23 These enzymes are secreted as proenzymes and are activated by cleavage of a polypeptide from the amino terminal portion of the molecule. They have been implicated in normal remodeling processes, such as those involved in developmental events45 and wound healing.6-7 Abnormal expression of MMPs has been associated with pathologic conditions including rheumatoid arthritis,8 corneal ul-
3541
3542
Investigative Ophthalmology & Visual Science, December 1993, Vol. 34, No. 13
ceration,910 and invasion and metastasis of cancer cells.11 a 1-Proteinase inhibitor (formerly called al-antitrypsin) and a2-macroglobulin are two major proteinase inhibitors in the plasma.1213 a-1-Proteinase inhibitor reacts with serine proteinases, and a2-macroglobulin reacts with all classes of proteinases including MMPs.1213 Their mechanisms of inhibition are different: al-proteinase inhibitor reacts at the active site of the proteinases,13 and a2-macroglobulin is cleaved by the proteinases and undergoes a conformational change, trapping the proteinases.1415 As a result of the cleavage, internal thiolesters in a2-macroglobulin become exposed and may react with the proteinase, forming covalently bound complexes. The inhibitorproteinase complexes then are cleared rapidly in blood by a receptor-mediated mechanism.15 Both a l proteinase inhibitor and a2-macroglobulin are key defense proteins that act to control the degradation of tissue constituents.1315 The aqueous humor is an extracellular fluid contained in the eye.16 It is produced by the ciliary body and percolates and drains through the pathway located at the angle of the anterior chamber. The aqueous humor provides nutrition to the surrounding avascular tissues, including the cornea, lens, and trabecular meshwork, and functions as a reservoir, accommodating metabolites from the various types of cells. The physical and chemical properties of the aqueous humor have been thought to play an important role in influencing the dynamics of its outflow17 and the development of pathologic conditions.18 Investigators have detected many components of normal serum, including transferrin,19 transforming growth factor-/?,20-21 and fibronectin,22 in the aqueous humor. In addition, type I collagenolytic activity23 and tissue-type24 and urokinase-type plasminogen activator25 also have been found in the aqueous humor. In view of the importance of the MMPs and other proteinases, we sought to determine whether the human aqueous humor contains these molecules. In addition, we extended previous investigations26"28 by examining the levels of al-proteinase inhibitor and a2-macroglobulin in the aqueous humor. MATERIALS AND METHODS Aqueous Humor Sampling Human aqueous humor was collected from 11 patients undergoing elective surgery for senile cataract by a previously described method.29 Their ages ranged from 60 to 84 years (mean ± SD, 69.5 ± 6.5 years). The patients had no ocular diseases other than cataract. In addition, none of them had general diseases such as type II diabetes mellitus or hypertension. Ap-
proximately 100 to 130 ii\ aqueous humor was obtained from each subject. The collected samples were stored at -20°C and thawed just before the experiments. The tenets of the Declaration of Helsinki were followed. Informed consent was obtained from all patients, and approval was granted by the institutional human experimentation committee. Protein Determination The protein concentration in the sample was determined by the method of Bradford30 with bovine serum albumin as the standard. Aliquots of the human aqueous humor (2 ix\) were pipetted into flat-bottom 96-well plates (Costar, Cambridge, MA), and Coomassie brilliant blue G-250 dye reagent (Bio-Rad, Richmond, CA) was added to each well. The optical density of each well was measured at 600 nm with a microplate reader (Biotek, Burlington, VT). The protein content of the aqueous humor samples was determined from the standard albumin curve. Zymography Zymography was performed as previously described31 in sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis with gels containing substrates, such as gelatin (for gelatinase) and casein (for stromelysin and serine proteinases), to analyze the content of specific proteinases in the aqueous humor samples. In brief, 10% or 11% SDS-polyacrylamide gels were prepared, and 1.5 mg/ml of either bovine skin gelatin (type A, Sigma Chemical Company, St. Louis, MO) or casein (Nutritional Biochemicals, Cleveland, OH) was copolymerized in the gel. Human aqueous humor samples (12.5 to 20 fi\ each) were mixed with an equal volume of 2X sample buffer, loaded onto the gel without boiling, and electrophoresed at 4°C under nonreducing conditions. The gel then was shaken for 1 hour in a 2.5% solution of Triton X-100 (Sigma) at room temperature to remove the SDS and incubated at 37°C in a reaction buffer (50 mM Tris, pH 7.5; 10 mM CaCl2) overnight to allow proteinase digestion of the substrates. Some casein gels also were incubated for 48 hours in 50 mM Tris (pH 8.0), 5 mM CaCl2, and 0.02% azide. After staining with Coomassie brilliant blue R-250 (Bio-Rad), the positions of enzymatic activities were identified as clear bands in the stained background. Molecular weight standards (Bio-Rad) and trypsin (as a positive control; Sigma) were run in parallel. In some gelatinase experiments, one of the metalloproteinase inhibitors, ethylenediaminetetraacetic acid (EDTA) (dissolved in water; Sigma) or 1,10-phenanthrolene (dissolved in methanol; Sigma), was included in the 2.5% Triton X-100 solution for SDS/Triton exchange, as well as in the Tris-CaCl2 buffer dur-
Proteinases and Inhibitors in the Aqueous Humor
ing proteinase reactions. The final concentration of each inhibitor was 10 mM. To further examine the caseinolytic enzymes, inhibitors were added to a final concentration of 5 mM for EDTA, 10 mM for phenylmethylsulfonyl fluoride (dissolved in isopropanol; Sigma), 1 Mg/ml for aprotinin (dissolved in water; Sigma), 5 /iM for E-64 (Ltrans-epoxysuccinyl-leucylamido(4-guanidino)-butane or N- [N-(L-3-trans-carboxyoxiran-2-carbonyl)-L-leucyl]agmatine, dissolved in water; Sigma), or 1 /xM for pepstatin (dissolved in methanol; Sigma) to the samples before electrophoresis and to the gel incubation solutions used after electrophoresis. Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis and Immunodetection of Proteinase Inhibitors Human aqueous humor samples (20 fil) were electrophoresed on 5% to 20% gradient gels under reduced conditions. The proteins then were electroblotted overnight onto a nitrocellulose membrane.32 After blocking with 5% Carnation nonfat dry milk, the membrane was incubated with rabbit anti-al-proteinase inhibitor (1:1000; DAKO, Santa Barbara, CA) and biotinylated goat antirabbit IgG. Immunodetection was performed with the Vectastain ABC system (Vector Laboratories, Burlingame, CA) with 3,3-diaminobenzidine (Sigma) to develop the color. Prestained molecular weight marker (Bio-Rad) and normal human serum samples were run in parallel. Human aqueous humor samples (12.5 /xl) and a2-macroglobulin (Athens Research Technology, Athens, GA) were electrophoresed on 6% SDS gels under nonreducing conditions. The gels were electroblotted, and the nitrocellulose membranes were blocked and washed in the same manner as those for a 1-proteinase inhibitor. The blots were allowed to react with either rabbit anti-a2-macroglobulin (1:2000; Accurate Corporation, Westbury, NY) or rabbit IgG (1:2000; Sigma), and then they were incubated with goat antirabbit IgG conjugated to horseradish peroxidase (1:2000; Bio-Rad). Immunoreactive bands were visualized with the luminol-based ECL system (Amersham, Arlington Heights, IL) and recorded on radiographic film (Konica, Tokyo, Japan). The lanes containing molecular weight standards (Sigma) were stained with India ink. Enzyme-Linked Immunosorbent Assay for alProteinase Inhibitor Aqueous humor samples in appropriate dilutions or al-proteinase inhibitor standard (1 to 15 ng; Sigma) was added to each well of 96-well enzyme-linked immunosorbent assay plates (Costar). After the wells were coated at 4°C overnight, excess solution was removed and the wells were incubated sequentially with
3543
3% bovine serum albumin, rabbit anti-al-proteinase inhibitor (1:1000; DAKO), goat anti-rabbit IgG (1:500; Antibodies, Inc., Davis, CA), and rabbit peroxidase-antiperoxidase complex (1:1000; DAKO). Each incubation at 37°C lasted 30 minutes. The peroxidase substrate O-phenylenediamine-2-HCl (Sigma) then was added to each well, and the color developed was measured in a microplate reader (Biotek) at 490 nm.32 The condition selected for the enzyme-linked immunosorbent assay was optimized to provide a linear curve for the standard, and the standard curves (correlation coefficient > 0.95) were used to determine the al-proteinase inhibitor content in the samples. Dot Blot Assay for a2-Macroglobulin Fifty microliters of aqueous humor samples (diluted 1:5) or «2-macroglobulin standard (4 to 64 ng) were applied to nitrocellulose membranes with a 96-well filtration apparatus. After drying, the membranes were washed, blocked with nonfat dry milk, and incubated sequentially with rabbit anti-human a2-macroglobulin and horseradish peroxidase-labeled goat anti-rabbit IgG.33 The dots were visualized by the ECL system and quantified with a microplate reader. RESULTS The total protein concentration in the aqueous humor samples was 151 to 286 fig/m\ (mean ± SD, 218.1 ± 54.2 /ig/ml). The electrophoretic pattern of proteins by Coomassie blue staining was similar in all samples (photographs not shown). Enzymes in the Aqueous Humor Zymography of gelatin-containing gels showed a major gelatinase band at a molecular weight of 66 kD (Fig. 1). Minor bands were also apparent at molecular weights of 125, 95, and 62 kD in the human aqueous humor (Fig. 1). Trypsin, run as a positive control, showed bands at 29 and 24 kD. The gelatinase activities were inhibited by treatment with 10 mM of metal chelators EDTA (Fig. 2) and 1,10-phenanthroline, indicating that all these bands are products of metalloproteinase activities. The gel containing casein that was incubated overnight showed only the bands of trypsin (Fig. 3). When the casein-containing gels were incubated for 48 hours (Fig. 4), caseinase activities at the molecular weight range of 80 to 84 kD appeared. A band at 68 kD (Figs. 4, 5) and a minor band at 48 kD also were observed (Fig. 4). For additional examination of the caseinases present, the aqueous humor samples and gels were incubated in the presence of various proteinase inhibitors. The caseinolytic activities seen in the untreated gel
3544
kD
Investigative Ophthalmology & Visual Science, December 1993, Vol. 34, No. 13
1 2 3 4 5 6 7 8 9
kD
1 2 3 4 5 6 7 8 9
1166645-
31-
FIGURE 1. Zymography of gelatinolytic activities in the human aqueous humor. Aqueous samples (20 /xl) were electrophoresed on a gelatin-containing 11% gel. One major band at 66 kD and minor bands at 125, 95, and 62 kD were seen in eight human aqueous samples from patients with cataracts (lanes 1 to 6, 8, and 9). Trypsin (0.5 ng, lane 7) showed major bands at 29 and 24 kD.
kD
1 2 3 4 5 6 7 8 9
FIGURE 2. Zymography of gelatinolytic activities in the human aqueous humor in the presence of EDTA. Gelatinase activities in the eight human aqueous samples (lanes 1 to 8) used for Figure 1 were inhibited by pretreatment of 10 mM EDTA and inclusion of 10 mM EDTA during incubation of gels in 2.5% Triton X-100, as well as in the Tris-CaCl2 buffer. The activity of the major trypsin band (0.5 ng, lane 9) seen at 24 kD and that of the minor band at 29 kD also was inhibited significantly by EDTA.
FIGURE 3. Zymography of caseinolytic activities in the human aqueous humor. Aqueous samples (20 fi\) were electrophoresed on a casein-containing 11% gel and incubated for 18 hours in the Tris-CaCl2 buffer. The major trypsin bands (0.5 ^g, lane 1) were seen at 31 and 24 kD. No activity was visible at approximately 52 kD, where stromelysin is expected in eight human aqueous samples from patients with cataracts (lanes 2 to 9). (Fig. 5, gel A) were essentially unaffected by the cysteine proteinase inhibitor E-64 (Fig. 5, gel B) and aspartic proteinase inhibitor pepstatin (Fig. 5, gel F). The 68-kD protein band was enhanced by EDTA (Fig. 5, gel C), methanol (Fig. 5, gel E), and isopropanol (Fig. 5, gel G). The 48-kD activity was partially reduced in isopropanol (Fig. 5, gel G). Both serine proteinase inhibitors, aprotinin (Fig. 5, gel D) and phenylmethylsulfonyl fluoride (Fig. 5, gel H), inhibited all caseinase bands. These results suggest that the observed caseinases are serine proteinases.
FIGURE 4.
Zymography of caseinolytic activities in the human aqueous humor. Five aqueous samples (12.5 pd each) from patients with cataracts were electrophoresed on a caseincontaining 10% gel and incubated for 48 hours in 50 mM Tris (pH 8.0), 5 mM CaCl2, and 0.02% azide. (a) The position of the 84- to 80-kD protein bands, (b) The position of the 68-kD band, (c) The position of the 48-kD band.
Proteinases and Inhibitors in the Aqueous Humor A
a
* •
O
W
3545 W
ti
r
w u
illlllil
FIGURE 5. Zymography of caseinolytic activities in a representative human aqueous humor sample in the absence or presence of proteinase inhibitors. Aqueous humor (12.5 n\) from one patient with a cataract was electrophoresed with or without protease inhibitors on caseincontaining 10% gels and was incubated for 48 hours in incubation buffer (A) without protease inhibitors or containing (B) 5 fiM E-64, (C) 5 mM EDTA, (D) 1 fig/ml aprotinin, (E) 2.5% methanol, (F) 1 pM pepstatin in methanol, (G) 5% isopropanol, and (H) 10 mM phenyl methyl sulfonyl fluoride in isopropanol. (a) The position of the 84- and 80-kD protein bands, (b) The position of the 68-kD band, (c) The position of the 48-kD band.
Proteinase Inhibitors in Aqueous Humor A major 53-kD protein band corresponding to alproteinase inhibitor molecule (Fig. 6) was seen with Western blot immunodetection in all aqueous humor samples and serum. With the enzyme-linked immunosorbent assay, the aqueous humor concentration of al-proteinase inhibitor was found to be 32.2 ± 9.9 Mg/ml (range, 16.0 to 50.4 /ag/ml; n = 11). This amounted to 14.9 ± 4.5% of the total protein in human aqueous humor. a-2-Macroglobulin was detected on Western blots when the SDS gels were run under nonreducing conditions as a band co-migrated with the purified serum a2-macroglobulin (Fig. 7). The calculated molecular weight was 360 kD. The level of a2-macroglobulin in the human aqueous humor determined by a dot blot assay was 3.2 ± 1 . 3 Mg/ml (range, 1.3 to 4.6 jug/ml; n = 6).
DISCUSSION All aqueous humor samples used were collected during elective surgery in patients with cataracts because previous studies28 have verified that the postmortem aqueous humor is not a reliable source of samples for analyses of components in vivo. The protein concentration in our samples was well within the range of those reported in the literature (98 to 360 M g / m l )
29,34-36
The abundant 66-kD gelatinase found by zymography is similar in molecular size to that extracted from the normal corneal stroma7-10'31 and that pro-
duced by corneal keratocyte31 and trabecular meshwork cultures.37 This enzyme emerged as a 72-kD protein band on reduction and Western blot analysis and was identified as the proenzyme form of 72-kD gelatinase MMP-2.31'37'38 The electrophoretic mobility of the 62-kD gelatinase suggested that it might be an activated form of MMP-2. The 95-kD gelatinase is in the size range to be the proenzyme form of MMP-9, a 92to 95-kD gelatinase produced by cultures of corneal epithelium31 and corneal fibroblast39 and by trabecular meshwork37 cultures after phorbol ester induction. There is also another possibility that the lower molecular weight bands in the aqueous humor could represent breakdown products of the 125-kD band, which might be related to the 130-kD species found in the plasma.40 The gelatinases in the aqueous humor are likely to be contributed in part by the cornea and other surrounding cells and tissues; however, most of the aqueous humor gelatinases still may be derived from the human plasma, in which gelatinases with molecular weights of 92, 66, and 62 kD (besides the 130-kD gelatinase) were detected.40 Two major bands at 29 and 24 kD were observed for the trypsin used. The higher molecular weight band could be a contaminating protease or another form of the enzyme that is folded differently, migrating more slowly than the 24-kD form. This latter alternative is possible because the sample was not boiled before electrophoresis, nor was a reducing agent added.
3546
kD
Investigative Ophthalmology & Visual Science, December 1993, Vol. 34, No. 13
12 3 4 5 6 7
844 7 — "—~—"• • » — •—"•"
3324-
FIGURE 6. Western blot analysis of a 1 -proteinase inhibitor. The a 1-proteinase inhibitor standard (500 ng, lane 1), normal human serum (5 /rg protein, lane 2), and five human aqueous humor samples from patients with cataracts (lanes 3 to 7) all showed a major protein band of 53 kD immunoreactive to an anti-cd-proteinase inhibitor antibody.
In agreement with a previous article,25 proteinase bands in human aqueous humor samples were not found in casein-containing gels when incubated overnight. Caseinolytic activities, however, were shown after extended incubation of casein-containing gels. This suggests that the caseinases are not present in appreciable amounts. The caseinases observed were inhibited by the serine proteinase inhibitors aprotinin and phenylmethylsulfonyl fluoride, but not by the cysteine proteinase inhibitor E-64, aspartic proteinase inhibitor pepstatin, or metalloproteinase inhibitor EDTA. These inhibitor results indicate that the caseinases are serine proteinases. The enhancement of the 68-kD proteinase band by EDTA, methanol, and isopropanol suggests that this enzyme is more active after a cation is removed or when a conformational change occurs. a-1-Proteinase inhibitor was shown to represent approximately 15% of the total protein in the human aqueous humor. The reported value of this inhibitor in the human serum was 2.5 ± 0.4 mg/ml,41 constituting approximately 3% of the total serum protein. The
greater proportion of a 1-proteinase inhibitor relative to total protein in the aqueous humor compared with serum suggests that, in addition to filtration from the serum, this inhibitor may be degraded to a lesser degree in the aqueous humor, secreted preferentially by the ciliary processes, or actively produced locally by corneal and other types of cells. The latter possibility is supported by previous studies in which this inhibitor was localized in the cornea33 and trabecular meshwork.42 Western blot analysis of the aqueous humor indicated the presence of a 360-kD dimer form of a2-macroglobulin held together by disulfide bonds.4344 It is unknown whether the «2-macroglobulin molecule actually exists as a dimer (360 kD) or tetramer (720 kD) because the ionic bonds between disulfide-bridged dimers are broken in the presence of SDS. Both forms of the inhibitor are active,43'44 with the tetramer being the form found in serum. The a2-macroglobulin level in the aqueous humor is approximately 1000-fold less than that in serum.45 This level was still greater than that expected because filtration of high molecular weight proteins from the serum is not favored36; however, the inhibitor may be synthesized and secreted by tissues surrounding the anterior chamber. Corneal cells, for example, have been shown to contain46 and produce47 a2-macroglobulin. The gelatinases, serine-like proteinases, and proteinase inhibitors in the aqueous humor may partici-
FIGURE 7. Western blot analysis of a2-macroglobulin. The «2-macroglobulin standard (500 ng, lane 1) and two aqueous humor samples from patients with cataracts (lanes 2 and 3) all showed a major protein band of immunoreactivity at 360 kD (arrowhead). By dot blot assays, the samples in lanes 2 and 3 were found to contain 4.6 Mg/ml a n d 1.3 Mg/m' a2-macroglobulin, respectively. The blots incubated with nonimmune serum in place of the primary antibody served as controls, in which the 360-kD band was not observed.
Proteinases and Inhibitors in the Aqueous Humor pate in the remodeling of extracellular matrices in the trabecular meshwork and in other tissues bordering the anterior chamber. The level and balance between the enzymatic action and inhibitory defense conceivably are essential in the turnover of tissue constituents and the normal physiology. Activity changes in the enzyme or inhibitor may upset the balance, resulting in excessive or insufficient degradation processes, and, in turn, they may affect the structure and integrity of cells and tissues. It is also noteworthy that al-proteinase inhibitor, which forms a 1:1 complex with its target enzymes, such as elastase, cathepsin G, and plasmin,13 may be cleaved and inactivated by metalloproteinases, including gelatinases.48"50 Our data suggest that gelatinase in the aqueous humor under normal conditions is mostly the inactive 66-kD proenzyme form of MMP2. Under conditions such as inflammation, the proenzyme may be activated, rendering the inhibitor inactivated. These scenarios may trigger uncontrolled proteolysis, leading to tissue damage and serious pathologic consequences. Key Words aqueous humor, gelatinase, caseinase, zymography, proteinase inhibitor Acknowledgments The authors thank Isabel Chang and Roger William for technical assistance. References 1. Woessner JF Jr. Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEBJ. 1991;5:2145-2154. 2. Fini ME, Plucinska IM, Mayer AS, Gross RH, BrinckerhoffCE. A gene for rabbit synovial cell collagenase: Member of a family of metalloproteinases which degrades the connective tissue matrix. Biochemistry. 1987;26:6156-6164. 3. Huhutala P, Tuuttila A, Chow LT, Lohi J, Keski-Oja J, Tryggvason K. Complete structure of the human gene for 92 kDa type IV collagenase: Divergent regulation of expression for the 92- and 72-kilodalton enzyme genes in HT-1080 cells. J Biol Chem. 1991;266: 16485-16490. 4. Grillo HC, Gross J. Collagenolytic activity during mammalian wound repair. Dev Biol. 1967; 15:300317. 5. Librach CL, Werb Z, Fitzgerald ML, et al. 92 kD type IV collagenase mediates invasion of human cytotrophoblasts./ Cell Biol. 1991 ;113:437-449. 6. Brenner CA, Adler RR, Rappolee DA, Pedersen RA, Werb Z. Genes for extracellular matrix-degrading metalloproteinases and their inhibitor, TIMP, are expressed during early mammalian development. Genes Dev. 1989;3:848-859. 7. Matsubara M, Girard MT, Kublin CL, Cintron C, Fini ME. Differential roles for two gelatinolytic enzymes of
3547 the matrix metalloproteinase family in the remodelling cornea. Dev Biol. 1991;147:425-439. 8. Okada Y, Yakeuchi N, Tomita K, Nakanishi I, Nagase H. Immunolocalization of matrix metalloproteinase 3 (stromelysin) in rheumatoid synovioblasts (B cells): Correlation with rheumatoid arthritis. Ann Rheum Dis. 1989;48:653-654. 9. Berman MB. Collagenase and corneal ulceration. In: Woolley DE, Evanston JM, eds. Collagenase in Normal and Pathological Connective Tissues. New York: John
Wiley and Sons; 1980:140-174. 10. Matsubara M, Zieske J, Fini ME. Mechanism of basement membrane dissolution preceding corneal ulceration. Invest Ophthalmol Vis Sci. 1991;32:92-107. 11. Liotta LA, Rao CN, Wewer UM. Biochemical interactions of tumor cells with the basement membrane. Annu Rev Biochem. 1986;55:1037-1057. 12. Barrett AJ, Starkey PM. The identification of a2-macroglobulin with proteinases: Characteristics and specificity of the reaction and a hypothesis concerning its molecular mechanism. Biochem]. 1973;133:709-724. 13. Heidtmann H, Travis J. Human al-proteinase inhibitor. In: Barrett AJ, Salvesen G, eds. Proteinase Inhibitors. Amsterdam: Elsevier; 1986:441-456. 14. Roche PA, Moncino MD, Pizzo SV. Independent analysis of bait region cleavage dependent and thiolester bond cleavage dependent conformational changes by crosslinking of a2-macroglobulin with cis-dichlorodiammineplatinum (II) and dithiobis(succinimidyl propionate). Biochemistry. 1989;28:7629-7636. 15. Borth W. «2-Macroglobulin, a multifunctional binding protein with targeting characteristics. FASEB J. 1992;6:3345-3353. 16. Cole DF. Ocular fluids. In: Davson H, ed. The Eye. Vol 1A. New York: Academic Press; 1984:269-390. 17. Tripathi RC. The functional morphology of the outflow systems of ocular and cerebrospinal fluids. Exp Eye Res. 1977;25(suppl):65-116. 18. Epstein DL, Jedziniak JA, Grant WM. Obstruction of aqueous outflow by lens particles and by heavy-molecular weight soluble lens proteins. Invest Ophthalmol Vis Sci. 1978;17:272-277. 19. Tripathi RC, Millard CB, Tripathi BJ, Norohna A. Tau fraction of transferrin is present in human aqueous humor and is not unique to cerebrospinal fluid. Exp Eye Res. 1990;50:541-547. 20. Jampel HD, Roche N, Stark WJ, Roberts AB. Transforming growth factor-/? in human aqueous humor. Curr Eye Res. 1990;9:963-969. 21. Granstein RD, Staszewski R, Knisely TL, et al. Aqueous humor contains transforming growth factorj8 and a small (
