Expression of matrix metalloproteinases and tissue ...

GLIA 33:205–216 (2001)

Expression of Matrix Metalloproteinases and Tissue Inhibitors of Metalloproteinases in Human Optic Nerve Head Astrocytes OLGA A. AGAPOVA,1 CYNTHIA S. RICARD,1 MERCEDES SALVADOR-SILVA,1 1,2 AND M. ROSARIO HERNANDEZ * 1 Department of Ophthalmology and Visual Sciences, Washington University School of Medicine, St. Louis, Missouri 2 Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, Missouri

KEY WORDS

MMPs; TIMPs; glaucoma; optic nerve head; lamina cribrosa; axons; extracellular matrix; glia

ABSTRACT Glaucomatous optic neuropathy is a common blinding disease characterized by remodeling of the extracellular matrix (ECM) and loss of retinal ganglion cell (RGC) axons at the level of the optic nerve head (ONH). Astrocytes, the major cell type in ONH, may participate in this process by production of matrix metalloproteinases (MMPs) and their inhibitors (TIMPs). In normal and glaucomatous ONH, we detected MMP and TIMP expression by immunohistochemistry. Cultured astrocytes were used to characterize expression of MMPs and TIMPs by zymography, Western blot, and RNase protection assay. MMP production was stimulated with phorbol 12-myristate 13-acetate (PMA). Astrocytes expressed MMP1, MT1-MMP, MMP2, TIMP1, and TIMP2 in normal and glaucomatous ONH. MMP2, TIMP1, and TIMP2 localized to RGCs and their axons. Increased MMP1 and MT1-MMP expression was demonstrated in glaucoma. Cultured astrocytes constitutively expressed MMP2, MT1-MMP, TIMP1, and TIMP2, whereas MMP3, MMP7, MMP9, and MMP12 were not detectable in tissues or in cultured astrocytes. Our findings demonstrate the presence of specific MMPs and TIMPs in the ONH that may participate in the homeostasis and remodeling of the ECM in glaucoma. Expression of the same MMPs and TIMPs in cultured ONH astrocytes will allow further studies on the mechanisms regulating these enzymes. GLIA 33:205–216, 2001. ©

2001 Wiley-Liss, Inc.

INTRODUCTION Matrix metalloproteinases (MMPs) or matrixins are a family of Zn– dependent metalloproteinases, which have the specialized function degrading the extracellular matrix (ECM) components such as collagens, proteoglycans, elastin, laminin, fibronectin, and other glycoproteins in normal and pathological conditions (Woessner, 1991). A multigene family encodes for secreted and membrane-associated MMPs. MMPs are synthesized in preproenzyme form and most of them are secreted from cells as proenzymes and activated in the extracellular compartment. Currently, more then ©

2001 Wiley-Liss, Inc.

20 MMPs are described, which are involved in normal and pathological processes connected with remodeling and destruction of the ECM (Nagase and Woessner, 1999). Some MMPs are expressed constantly at certain level and in most cell types, while others are inducible and tissue-specific (Fini et al., 1998). The expression

Grant sponsor: NIH; Grant number: EY-06416 and EY-02687. *Correspondence to: M. Rosario Hernandez, Department of Ophthalmology and Visual Sciences, Washington University School of Medicine, 660 S. Euclid Avenue, Box 8096, St. Louis, MO 63110. E-mail: [email protected] Received 28 September 2000; Accepted 20 November 2000

206

AGAPOVA ET AL.

and activity of MMPs is tightly regulated (Borden and Heller, 1997; Nagase, 1997). Specific proteins known as the tissue inhibitors of metalloproteinases (TIMPs) are the physiological regulators of these enzymes. The TIMP family comprises four members (TIMP1-4) with high gene and protein homology (Douglas et al., 1997). The optic nerve head (ONH) is thought to be the site of retinal ganglion cell (RGC) damage in primary open angle glaucoma (POAG), a common blinding neurodegenerative disease. POAG is characterized clinically by cupping or excavation of the optic disk and in many patients by elevated intraocular pressure (IOP) (Quigley et al., 1983). At the microscopic level, disk cupping in POAG is due to loss of the RGC axons and extensive remodeling of the ECM at the level of the lamina cribrosa of the ONH (Hernandez and Pena, 1997). In humans and nonhuman primates, the lamina cribrosa consists of several fibroelastic connective tissue lamellae, the cribriform plates. They are arranged in register, leaving channels to allow the exit of the nonmyelinated axons of the RGCs. Astrocytes, oriented horizontally across the nerve head perpendicular to the axons, line the cribriform plates (Anderson, 1969). Collagen fibers, elastic fibers, basement membranes, and proteoglycans form the ECM of the cribriform plates. Basement membranes separate the astrocytes from the underlying ECM (Hernandez and Pena, 1997). In addition, during development of the retina and optic nerve, the lamina cribrosa is thought to constitute a barrier to prevent migration of myelinating cells and myelination of the retina (Perry and Lund, 1990). In POAG, there is extensive remodeling of the ECM of the lamina cribrosa with increased expression of ECM, cell adhesion molecules, and neurotoxic mediators by reactive astrocytes (Hernandez, 2000). Astrocytes detach from their basement membranes and migrate into the nerve bundles (Varela and Hernandez, 1997). There are marked changes in elastic fibers at the level of the lamina cribrosa, which lead to elastotic degeneration of the ECM (Pena et al., 1998). In addition, there is a decrease in the density of collagen fibers in the tissue (Quigley et al., 1991; Hernandez, 1992). However, in POAG, despite all these changes, there is no neovascularization, no breakdown of the bloodnerve barrier, no obvious sign of inflammation, and no formation of a typical glial scar. This pathophysiology suggests a very specific and tightly regulated activity of ECM-degrading enzymes and their inhibitors. The important role of MMPs and TIMPs preserving the homeostasis of the ECM in most tissues led us to identify and localize the major MMPs and TIMPs in the normal and glaucomatous ONH and in cultured type 1B astrocytes. In this study, we used immunohistochemistry to screen human normal and glaucomatous tissues for expression and distribution of several MMPs and TIMPs. In addition, we characterized the expression, synthesis, and enzymatic activity of MMPs and TIMPs in cultured human ONH astrocytes (type 1B) using casein and gelatin zymography, Western blots, and RNase protection assays.

MATERIALS AND METHODS Tissue Preparation Human eyes from nine donors, ages 54 – 85 (70 ⫾ 12 years, mean ⫾ SD), with no history of eye disease, diabetes, or neurodegenerative disease, were obtained from eye banks throughout the United States through National Disease Research Interchange (NDRI) and the Mid-America Transplant Services (Saint Louis, MO). Eyes from 14 donors, ages 47– 89 (71 ⫾ 14 years), with well-documented POAG, were obtained through the Glaucoma Research Foundation and NDRI. The eyes with glaucoma had extensive histories, which were evaluated by an ophthalmologist to ascertain POAG and degree of damage. In addition, cross-sections of the myelinated nerves stained with p-phenylenediamine were routinely checked to confirm axonal loss (Pena et al., 1998). The eyes with POAG were separated into three groups according to the clinical histories into mild, moderate, and advanced POAG. Due to the nature of this study, there was no history of sepsis or infections in any of the donor used. The cause of death for all donors was myocardial infarction or cardiopulmonary failure. The eyes were enucleated shortly after death (2– 4 h) and fixed in 10% neutral buffered formalin and transported to the laboratory on ice within 24 h after death. The optic nerves were dissected, washed several times in phosphate buffered saline (PBS) containing 20 mM of glycine, and paraffin-embedded according to standard protocol. Four eyes with POAG and two normal age-matched controls were embedded in O.C.T. Compound (Miles, Elkhart, IN) and flash-frozen in methyl butane cooled in liquid nitrogen.

Immunochistochemistry Six ␮m sagittal sections were used for MMP1, MMP2, MMP3, MMP7, MMP9, MMP12, MT1-MMP, and TIMP1 immunodetection. For immunoperoxidase detection, sections after blocking endogenous peroxidase activity were incubated with primary antibodies (Table 1) and then with appropriate biotinylated secondary antibodies, avidin– biotinylated peroxidase, and DAB (3⬘,3⬘-diaminobenzidine) substrate from Vestastain ABC Kit (Vector Laboratories, Burlingame, CA); nuclei were counterstained with hematoxylin. For TIMP1 detection, antigen retrieval in 1 mM EDTA, pH 8.0, at 100°C 20 min was performed. For TIMP2 and TIMP1 detection, we used 15 ␮m sagittal frozen sections followed by immunofluorescence staining. For double immunofluorescence staining, we used mouse monoclonal antibody against human glial fibrillary acid protein (GFAP) and rabbit polyclonal antibodies against MMP1, MT1-MMP, TIMP1, and TIMP2 (Sigma, St. Louis, MO) and Alexa 488 and Alexa 568 – labeled secondary antibodies (Molecular Probes, Eugene, OR). For negative control, the primary antibody was replaced for nonimmune serum. To control for

207

MATRIX METALLOPROTEINASES TABLE 1. Antibodies for immunohistochemistry and western blots Antigen TIMP1 TIMP1 TIMP2 TIMP3 MMP1 MMP2 MMP2 MMP3 MMP3 MMP7 MMP9 MMP12 MT1-MMP GFAP

Antibodies

Source and catalog number

Mouse mono Rabbit poly Rabbit poly Rabbit poly Rabbit poly Rabbit poly Mouse mono Mouse mono Rabbit poly Rabbit poly Mouse mono Rabbit poly Rabbit poly Mouse mono

Neomarkers, MS-608 Sigma, T8187 Sigma, T8062 Sigma, T7812 Sigma, M4177 Chemicon, AB809 Neomarkers, MS-567 Neomarkers, MS-810 Chemicon, AB810 Gift from Dr. Parks Chemicon, MAB3300 Gift from Dr. Shapiro Sigma, M3927 Sigma, G3893

cross-reactivity in double immunofluorescence, sections were incubated with primary antibody followed by the wrong species secondary antibody. Serial sections of normal and glaucomatous eyes were stained simultaneously to control for variations in immunostaining. Formalin-fixed, paraffin-embedded human placenta or human breast carcinoma sections, known to contain most of the MMPs and TIMPs, served as a positive control for immunostaining.

Visualization and Photography Slides were examined by microscopy (BHS, Olympus, Tokyo, Japan) and images were recorded using a digital camera (Spot, Diagnostic Instruments, Sterling Heights, MI) and stored as computer files in Photoshop (Adobe, version 5.0). Double immunofluorescent images were obtained by a Zeiss LSM 410 laser-scanning confocal microscope (Carl Zeiss, Thornwood, NY). Images were stored as TIF files and subsequently analyzed by Adobe.

Cell Culture Three pairs of human eyes from donors without history of eye disease (ages 42– 46) were used to obtain primary cultures of type 1B astrocytes from the explanted human optic nerve heads (Kobayashi et al., 1997). Type 1B astrocytes (passages 3–5) were plated at 2 ⫻ 105 cells per 100 mm dish and grown to confluence in DMEM/F12 with 10% FBS at 37°C and 5% CO2. Cells were preincubated (20 h) in DMEM/F12 with Insulin-Transferrin-Selenite (ITS; Sigma) and treated in fresh media with PMA (5 ⫻ 10⫺7 M) 6 or 24 h. Control cells were preincubated and grown in fresh DMEM/F12 with ITS 6 or 24 h. Conditioned media were collected in test tubes and stored at ⫺80°C. For Western blot and casein zymography, media were concentrated 20 times with the Centricon 10 Centrifugal Filter Devices (Amicon, Millipore, Bedford, MA)

Immunohistochemistry 1:50 1:100 1:100 1:200

Western blot 1:1,000 1:1,000 1:1,000 1:1,000 1:1,000

1:100 1:25 1:1,000 1:1,000 1:200 1:350 1:400 1:400

1:1,000 1:1,000

Protein Extraction Cells were washed twice in cold 1 ⫻ PBS and incubate in 500 ␮l of ice-cold RIPA buffer (50 mM Tris HCl, pH 7.5, 150 mM NaCl, 1 mM EGTA, 1% IGEPAL CA630, 0.5% deoxycholate, Roche protease inhibitors, Roche Molecular Biochemicals, Indianapolis, IN). After 15-min incubation, cells were scraped with disposable cell lifters and centrifuged for 15 min at 4°C and 14,000 RPM. The supernatant was recovered and protein concentration in cell lysates determined by Bio-Rad (Hercules, CA) Protein Assay Kit (Bradford method). Samples were stored at ⫺80°C.

Zymography Proteins (20 ␮l of conditioned media for gelatin and 20 ␮l 20 ⫻ media for casein zymography) were separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) on 8% separating gels with 0.1% gelatin, or with 0.2% casein, without reduction. After electrophoresis, gels were washed in 2.5% Triton X-100 two times on the shaker for 15 min at room temperature to remove the SDS. Gels were incubated overnight at 37°C in incubation buffer (50 mM Tris pH 8.2, 5 mM CaCl2, 0.5 ␮M ZnCl2), stained with 0.25% Coomassie blue, and destained in 5% acetic acid and 10% methanol in H2O. Areas of proteolytic activity were detected as transparent bands. Some samples were treated 1 h at 37°C with 10 mM APMA (p-aminophenylmercuric acetate–MMP activator) in 50 mM NaOH, or with 50 mM NaOH as negative control of activation. In a few experiments, 10 mM EDTA was added to incubation buffer to inhibit MMP activity. Gelatinase zymography standard (Chemicon, Temecula, CA) and MMP control 1 (Sigma) were used as positive controls for gelatin and casein zymography.

Western Blot Samples (10 ␮g of cell lysates and 20 ␮l 20 ⫻ conditioned media) were run on 4%–15% gradient SDS poly-

208

AGAPOVA ET AL.

acrylamide gel under reduction conditions and transferred to Bio-Rad nitrocellulose membrane. Membrane was blocked for 1 h in blocking solution (PBS, 0.05% Tween-20, 5% Amersham blocking agent, Amersham Pharmacia Biotech, Piscataway, NJ) and incubated 1 h with primary antibody (Table 1) diluted in PBS with 0.05% Tween-20 and 2.5% blocking agent. After washing with PBS/0.05% Tween-20, membrane was incubated with goat anti-rabbit or goat anti-mouse horseradish peroxidase-labeled secondary antibody (Amersham) diluted in blocking solution. After additional washes, binding of the peroxidase-labeled antibody was visualized by using ECL Western blotting detection system (Amersham).

RNA Isolation Total cytoplasmic RNA was isolated as previously described (Favaloro et al., 1980). Cells were rinse twice in saline and then scraped into saline. Cells were recovered by centrifugation and resuspended in TSM (10 mM Tris, pH 7.4, 150 mM NaCl, 2 mM MgCl2) ⫹ 0.5% IGEPAL CA-630. After incubation on ice (2–3 min), cells debris were pelleted. An equal volume of TSE⫹S (10 mM Tris, pH 7.4, 150 mM NaCl, 5 mM EDTA, 0.2% SDS) was mixed with supernatant. The samples were extracted twice with 50% phenol ⫹ 50% chloroform/ isoamyl alcohol (24:1) and twice with chloroform/ isoamyl alcohol (24:1). The supernatant was adjusted to 0.1 M NaCl, 2 volumes of ethanol were added and precipitated overnight at ⫺20°C. RNA was quantified by measuring absorbence at 260 nm and purity was assessed by calculating the absorbence ratio 260/280 nm.

Plasmid Constructs for Riboprobes Probe cDNA fragments for MMP2, MMP3, MMP7, and MMP9 were cut from cDNA (a gift from Dr. W.C. Parks) with appropriate restriction enzymes and cloned into vectors for in vitro transcription. Different vectors were used for cloning of different cDNA fragments due to restriction enzyme site compatibility. MMP3 and MMP7 cDNA fragments were cloned into pBluescript KS (Stratagene, La Jolla, CA), MMP9 into pBluescript SK (Stratagene), and MMP2 into pGEM3Zf(⫹) (Promega, Madison, WI). Probe cDNA fragments for MT1-MMP, MMP1, TIMP1, and TIMP2 were synthesized by RT-PCR of total RNA from cultured ONH astrocytes treated with PMA. MMP1, MT1MMP, and TIMP1 PCR fragments were cloned into pGEM(R)-T Easy (Promega), cut with EcoRI, and cloned into pBluescript KS. TIMP2 PCR fragment was cloned directly into pBluescript KS into SmaI site. Sequencing proved fragment orientation and sequence. The specificity of these sequences was confirmed by computer-assisted alignment with the reported cDNA sequences for MMPs and TIMPs (see accession num-

TABLE 2. Positions of the specific target sequences used to generate riboprobes for MMPs and TIMPs mRNA detection and probe and protected fragment sizes

MMP1 MMP7 MMP2 TIMPI MMP3 TIMP2 MMP9 MT1-MMP 18SrRNA

Sequence

Accession number

Probe size (base)

Protected fragment size (base)

1053-1342 403-679 1502-1752 302-491 1585-1757 571-739 1982-2123 1364-1486

NM_002421 NM_002423 J03210 S68252 NM_002422 S48568 NM_004994 NM_004995

388 320 308 316 227 255 206 185 128

290 277 251 190 173 169 142 123 80

bers in Table 2). One ␮g of each linearized plasmid was used for making antisense transcripts (riboprobes). The linearized Ambion pTRI RNA18S plasmid (Austin, TX) was used to produce 18SrRNA antisense probe that serve as internal control for the amount of RNA in each sample. Positions of the specific target sequences used to generate riboprobes for MMP and TIMP mRNA detection and also probe and protected fragment sizes are detailed in Table 2.

RNase Protection Assay Radioactively labeled antisense transcripts (riboprobes) were produced utilizing MEGAscript (for lowspecific-activity 18S RNA probe, ⬃ 5 ⫻ 104 cpm/␮g) or MAXIscript (for high-specific-activity MMPs and TIMPs probes, ⬃ 109 cpm/␮g) in vitro transcription Kits (Ambion). ␣-[32P] UTP (ICN, Costa Mesa, CA) and appropriate RNA polymerase (T3 or T7) were used for in vitro transcription. RNase protection assay was performed using RPA III Kit from Ambion. The riboprobes (5 ⫻ 104 cpm) were annealed to 5 ␮g total RNA 5 min at 85°C and then at 54°C overnight in Hybridization III Buffer. RNase digestion was performed with RNaseA/ RNaseT1 diluted 1:50 in RNase Digestion III Buffer at 37°C 30 min. After RNase inactivation and RNA duplexes precipitation, samples were dissolved in Gel Loading Buffer II, denatured at 90°C for 5 min, and run on 8% polyacrylamide sequencing gel. Probe sets (2 ⫻ 103 cpm each) and transcribed RNA Century Marker Templates (Ambion; 3 ⫻ 103 cpm) were run on the separate lanes as size standard.

RESULTS MMP and TIMP Expression in Normal and Glaucomatous Optic Nerve Head The distribution of MMPs and TIMPs in normal and glaucomatous human optic nerve heads were analyzed using immunoperoxidase staining for MMP1, MMP2, MMP3, MMP7, MMP9, MMP12, MT-MMP1, and TIMP1. Immunofluorescence staining was used for TIMP1 and TIMP2 detection. Double immunofluores-

MATRIX METALLOPROTEINASES

209

Fig. 2. Colocalization of MMP1 and MT1-MMP with GFAP in the glaucomatous lamina cribrosa. Double immunofluorescence staining for GFAP (green) and MMP1 (red) (A and B) and for GFAP (green) and MT1-MMP (red; C and D). Arrows point to the cells that show double staining. Arrowheads point to the extracellular localization of MMP1. NB, nerve bundles; CP, cribriform plates; PS, pial septa. Scale bar, 25 ␮m.

Fig. 1. MMP1 immunoperoxidase staining in normal (A–C) and glaucomatous (D–F) ONH. A and D, prelaminar region; B and E, lamina cribrosa; C and F, postlaminar region. In glaucomatous ONH (D–F), significantly higher level of MMP1 immunoreactivity than in control (A–C) was detected in laminar (E) and postlaminar (F) regions in association with astrocytes (arrows) and extracellular (arrowheads) in glial columns, cribriform plates, pial septa, and nerve bundles. GC, glial columns; CP, cribriform plates; PS, pial septa; NB, nerve bundles; V, blood vessels. Scale bar, 20 ␮m.

cence staining for MMPs or TIMPs and the astrocyte marker, GFAP, was performed to determine whether the cells expressing MMPs and TIMPs were astrocytes. Three distinct histological regions of the optic nerve head, the prelaminar region, lamina cribrosa, and postlaminar region, were analyzed.

MMP1 In the normal adult ONH, very low levels of MMP1 immunoreactivity was observed in association with astrocytes in the glial columns in the prelaminar region (Fig. 1A) and in the cribriform plates in the lamina cribrosa (Fig. 1B). There were few labeled astrocytes in the postlaminar myelinated optic nerve (Fig. 1C). In the glaucomatous ONH, immunoreactivity for MMP1 was markedly stronger than in normals in the laminar and postlaminar regions (Fig. 1). Immunostaining for MMP1 was associated with astrocytes and ECM in the cribriform plates. In glaucoma, MMP1 was also localized to the reactive astrocytes and axons in nerve bundles in the laminar and postlaminar regions (Fig. 1E and F). MMP1 immunoreactivity was detected in small vessels throughout the ONH. Double immunofluorescent staining for GFAP and MMP1 showed localization

of MMP1 to astrocytes in the cribriform plates and in the nerve bundles. MMP1 also stained the ECM in the cribriform plates and pial septa (Fig. 2A and B).

MT1-MMP In the normal adult ONH granular immunostaining for MT1-MMP was associated with blood vessels (Fig. 3A and C) and with a few astrocytes lining the cribriform plates in the lamina cribrosa (Fig. 3B). No MT1MMP immunoreactivity was detected in axons and ECM. In the glaucomatous ONH, high level of MT1MMP immunoreactivity was observed in the lamina cribrosa and the postlaminar optic nerve (Fig. 3E and F). Double immunofluorescence staining for GFAP and MT1-MMP demonstrated that MT1-MMP was associated with astrocytes in the optic nerve head (Fig. 2C and D).

MMP2 MMP2 was expressed in the ONH and retina. There was no difference in the level of expression between normal and glaucomatous ONH (Fig. 4). Retinal ganglion cell bodies and their axons in the nerve fiber layer were MMP2-positive (Fig. 4A and E). In the ONH, MMP2 was associated with the astrocytes in the glial columns and axons in the prelaminar region (Fig. 4B and F), astrocytes in the cribriform plates, and axons in the lamina cribrosa (Fig. 4C and G). MMP2 immunoreactivity in the nonmyelinated axons in the prelami-

210

AGAPOVA ET AL.

Fig. 3. MT1-MMP immunoperoxidase staining in normal (A–C) and glaucomatous (D–F) ONH. A and D, prelaminar region; B and E, lamina cribrosa; C and F, postlaminar region. Positive immunoreactivity to MT1-MMP was localized to blood vessels (A, C, and D). MT1-MMP immunoreactivity is remarkably higher in glaucomatous OHN in lamina cribrosa (E) and postlaminar region (F) compare to normal (B and C). Arrows point to immunoreactive astrocytes. V, blood vessels; GC, glial columns; NB, nerve bundles; CP, cribriform plates; PS, pial septa. Scale bar, 20 ␮m.

nar and laminar regions was stronger than in the postlaminar myelinated nerve (Fig. 4D and H).

MMP3, MMP7, MMP9, MMP12 In normal and glaucomatous adult ONH, no detectable immunoreactivity for MMP3 was observed in association with astrocytes, ECM, or nerve bundles. However, MMP3-positive staining was detected in the blood vessels throughout the region. Immunoreactivity for MMP3 was associated with the perivascular cells (data not shown). No immunoreactivity for MMP7, MMP9, and MMP12 localized in normal and glaucomatous adult ONH (data not shown).

Fig. 4. MMP2 immunoperoxidase staining in normal (A–D) and glaucomatous (E–H) ONH and retina. A and E, retina; B and F, prelaminar region; C and G, lamina cribrosa; D and H, postlaminar region. MMP2 immunoreactivity localized in RGCs (arrows) and their axons in nerve fiber layer in retina (A and E), in glial columns, cibriform plates, pial septa, and nerve bundles in prelaminar (B and F), laminar (C and G), and postlaminar regions (D and H). NFL, nerve fiber layer; GC, glial columns; CP, cribriform plates; PS, pial septa; NB, nerve bundles; V, blood vessels. Scale bar, 20 ␮m.

immunofluorescence staining for TIMP2 and GFAP demonstrated TIMP2 localization in the ONH similar to TIMP1 (Fig. 6). There were no apparent differences in the level of staining for TIMP1 and TIMP2 in normals and glaucomatous samples using either fluorescence or peroxidase detection (Figs. 5 and 6). Clear positive immunoreactivity for both TIMP1 and TIMP2 was localized to the soma of RGCs and their axons in nerve fiber layer (Fig. 7). Astrocytes in the nerve fiber layer did not express TIMP1 or TIMP2 (Fig. 7).

TIMP1 and TIMP2 In the normal and glaucomatous ONH, immunostaining for TIMP1 was associated with astrocytes and axons in the prelaminar region (Fig. 5A and D) and in the lamina cribrosa (Fig. 5B and E). In the postlaminar nerve, TIMP1 was associated with the axons and with astrocytes lining the pial septa (Fig. 5C and F). Double

MMP and TIMP Expression in Cultured Human Type 1B Astrocytes Zymography Gelatin and casein zymography were performed for detection of the proteolytic activity in conditioned me-


211

Western blots

Fig. 5. TIMP1 immunoperoxidase staining in normal (A–C) and glaucomatous (D–F) ONH. A and D, prelaminar region; B and E, lamina cribrosa; C and F, postlaminar region. TIMP1 immunoreactivity in normal and glaucomatous ONH is localized in glial columns, cribriform plates, pial septa, and nerve bundles in prelaminar (A and D), laminar (B and E), and postlaminar (C and F) regions. GC, glial columns; CP, cribriform plates; PS, pial septa; NB, nerve bundles; V, blood vessels. Scale bar, 20 ␮m.

dia from cultured ONH astrocytes (Fig. 8). Gelatinolytic activity at 68 kDa (Fig. 8A, lane 1) and light caseinolytic activity at 50 kDa (Fig. 8B, lane 1) were detected in 24 h conditioned media from nontreated cells. Media from cells treated 24 h with PMA demonstrated the same level activity at 68 kDa and a new activity at 92 kDa by gelatin zymography (Fig. 8A, lane 2). Caseinolytic activity at 50 kDa was apparently increased after PMA treatment and an additional caseinolytic activity at 78 kDa was detected (Fig. 8B, lane 2). To determine whether proteolytic activities detected by zymography were indeed MMPs, conditioned media from astrocytes were treated with p-aminophenylmercuric acetate (APMA) for MMP activation or with EDTA to inhibit MMP activity. After treatment with APMA, the gelatinolytic and caseinolytic bands at 92, 68, and 50 kDa partly converted to species of lower molecular weight at 86, 64, and 47 kDa, respectively (Fig. 8A and B, lanes 3). However, the band at 78 kDa did not change after APMA treatment (Fig. 8B, lane 3), indicating that this proteolytic activity does not correspond to MMPs. EDTA inhibited protease activity at 92, 68, and 50 kDa (data not shown). According their mobility, these MMPs are MMP9 (gelatinase B, 92 kDa), MMP2 (gelatinase A, 68 kDa), and MMP1 (interstitial collagenase, 50 kDa).

In order to confirm our zymography results, Western blot detection of MMPs and TIMPs in conditioned media and cell lysates was carried out. MMP2 was detected in media and cell lysates from control and PMA-treated cells. No additional induction of MMP2 expression was observed after PMA treatment (data not shown). MMP9 was not detectable in conditioned media and cell lysates in control samples. However, after 24 h of PMA treatment, a light band about 92 kDa appeared in media and cell lysates (data not shown). Low expression of MMP1 was found in conditioned media from control astrocytes (Fig. 9, lanes 1 and 2), but a clear induction of MMP1 both in cell lysates and media was observed after PMA treatment (Fig. 9, lanes 3 and 4). MMP3 was not detectable by Western blot of cell proteins or media from control astrocytes or after PMA treatment (data not shown). MT1-MMP was detected as a 65 kDa band in cell lysates from control astrocytes (Fig. 9, lane 5). A slight upregulation of MT1-MMP was observed after 24 h PMA treatment (Fig. 9, lane 7). TIMP1 and TIMP2 antibodies recognized bands at 30 and 24 kDa in conditioned media (Fig. 9). TIMP2 expression did not change after PMA treatment. A slight increase of TIMP1 in the media was detected after PMA treatment for 24 h compared with untreated controls (Fig. 9, lanes 2 and 4). In control cells, TIMP1 was identified as a band at approximately 35 kDa, but after PMA treatment, an additional band 30 kDa was appeared in cell lysates (Fig. 9, lanes 5–7).

MMP and TIMP m-RNA Expression MMP and TIMP m-RNA expression in cultured type 1B astrocytes was characterized using RNAse protection assay. Five ␮g of total RNA from ONH astrocytes was hybridized with antisense riboprobes for MMP1, MMP2, MMP3, MMP7, MMP9, MT1-MMP, TIMP1, TIMP2, and 18SrRNA (internal control; Table 2). Protected fragments of expected size for MMP2, MT1MMP, TIMP1, TIMP2, and 18SrRNA were visible in cells grown in control conditions (Fig. 10, lanes 1, 4, 5). No detectable bands for MMP1, MMP3, MMP7, and MMP9 mRNA were visible in control cells. There were no differences in band intensity and detection between 6- and 24-h controls (data not shown). When total RNA from cells treated with PMA was used for hybridization, we detected the MMP1 mRNA after 6 and 24 h of treatment (Fig. 10, lanes 2 and 3). The amount of MT1-MMP mRNA was clearly increased after 6- and 24-h PMA treatment (Fig. 10, lanes 6 and 7). TIMP1, MMP2, and TIMP2 mRNA were detected at the same level as in control, untreated cells (Fig. 10, lanes 1–3, 5–7). MMP3, MMP7, and MMP9 mRNA were not detectable after 6- or 24-h PMA treatment, as in control (Fig. 10).

212

AGAPOVA ET AL.

Fig. 6. Double immunofluorescence staining for GFAP (green) and TIMP2 (red) in lamina cribrosa in normal (A) and glaucomatous (B) ONH. TIMP2 immunostaining of nerve bundles and colocalization

(yellow) of GFAP and TIMP2 in astrocytes (arrows) in cribriform plates and nerve bundles. CP, cribriform plates; NB, nerve bundles. Scale bar, 25 ␮m.

Fig. 7. Double immunofluorescence staining for TIMP1 (red) and GFAP (green; A), and TIMP2 (red) and GFAP (green; B) of retina in glaucomatous eye. TIMP1 and TIMP2 staining of RGC bodies (arrows) and their axons in nerve fiber layer and GFAP-positive astrocytes (green) in nerve fiber layer (A and B). NFL, nerve fiber layer; VS, vitreal surface. Scale bar, 25 ␮m.

DISCUSSION This study reports the immunohistochemical localization of MMPs and TIMPs in normal and glaucomatous ONH. MMP and TIMP expression was further characterized at the protein and mRNA level in type 1B astrocytes cultured from human ONH. The results indicate that, as in the central nervous system (CNS), a restricted group of MMPs is expressed in normal and glaucomatous ONH and in cultured type 1B astrocytes. MMPs were expressed at low level in normal ONH. The expression of some MMPs was increased in glaucoma.

However, TIMP expression was detected at high level in both normal and glaucomatous tissues. Interestingly, TIMP1 and TIMP2 localized to the RGCs and their axons in the retina and ONH, suggesting that inhibition of MMP activity plays an important role in maintaining the intactness of these tissues in normal and pathological conditions. The main role of MMP1 or interstitial collagenase is degradation of fibrillar collagen type I, II, and III (Imper and Van Wart, 1998). The presence of increased amounts of MMP1 in reactive astrocytes and in the ECM of the glaucomatous lamina cribrosa is in agree-


213

Fig. 8. Gelatinolytic (A) and caseinolytic (B) activity of conditioned media (CM) from cultured ONH astrocytes. Lane 1, untreated cells; lanes 2– 4, CM from cells PMA-treated for 24 h; lane 3, CM incubated with 10 mM APMA (p-aminophenylmercuric acetate) in 50 mM NaOH for 1 h at 37°C for MMP activation; lane 4, CM incubated in 50 mM NaOH for 1 h at 37°C as a negative control of MMP activation. Prestained Molecular Weight Standard (Bio-Rad) of low range was run in each gel for estimating the molecular weights of sample proteins. 92 kDa, proMMP9; 86 kDa, active MMP9; 68 kDa, proMMP2; 64 kDa, active MMP2; 78 kDa, protease, but not MMP; 50 kDa, proMMP1; 47 kDa, active MMP1. Fig. 10. RNase protection assay of total RNA from control cells and cells treated with PMA for 6 or 24 h with riboprobes for MMPs and TIMPs. M, RNA Century Size Markers a set of five in vitro transcripts of 100, 200, 300, 400, and 500 bases (Ambion). Lanes 1, 4, 5: hybridization with 5 ␮g of total RNA from 24-h control cells. Lanes 2 and 6 and lanes 3 and 7, hybridization with 5 ␮g of total RNA from 6- and 24-h PMA-treated cells, respectively. Lanes 1–3, hybridization with riboprobes for MMP1, MMP3, MMP7, MMP9, TIMP1, and 18SrRNA; lane 4, hybridization with riboprobes for MMP1, MMP2, MMP3, MMP7, MMP9, MT1-MMP, TIMP1, TIMP2, and 18SrRNA; lanes 5–7, hybridization with riboprobes for MMP2, MT1-MMP, TIMP2, and 18SrRNA. Protected fragments for MMP1 (290 bases), MMP2 (251, 235 bases), TIMP1 (190 bases), TIMP2 (169 bases), MT1-MMP (140, 123 bases), and 18SrRNA (80 bases) are indicated. No protected fragments for MMP3 (173 bases), MMP7 (277 bases), and MMP9 (142 bases) are visible.

Fig. 9. Western blots of ONH astrocyte conditioned media (CM) and cell lysates (CL) with antibodies to TIMP2, TIMP1, MMP1, and MT1MMP. Lanes 1– 4, CM; lanes 5– 6, CL. Lanes 1 and 2: 6- and 24-h CM from control cells. Lanes 3 and 4: 6- and 24-h CM from PMA-treated cells. Lane 5, untreated cells. Lanes 6 and 7: 6- and 24-h PMA-treated cells. Prestained Molecular Weight Standard (Bio-Rad) of broad range was run in each gel for estimating the molecular weights of sample proteins. Seven Western blots are shown and 65 kDa, 52 kDa, 30 kDa, and 24 kDa indicate MT1-MMP, MMP1, TIMP1, and TIMP2 molecular weights, respectively.

ment with previous ultrastructural qualitative and quantitative studies that demonstrated loss of collagen fibers in the ECM of the cribriform plates (Quigley et al., 1991; Hernandez, 1992). Our study indicates that type 1B astrocytes are most likely to be the source of MMP1 in glaucoma. As was shown recently, cultured human brain astrocytes produced increased amounts of MMP1 after stimulation with IL-1␤ (Vos et al., 2000). The authors also reported that MMP1 was cytotoxic to organotypical spinal cord and neuronal cultures and

hypothesized that MMP1 exerted its effects through the destruction of ECM or by activation of membranebound receptors/cytokines as well as soluble cytokines (Vos et al., 2000). MT1-MMP is an integral membrane MMP that can participate in ECM degradation by activation of proMMP2 on the cell surface (Strongin et al., 1995). Recently, it was shown that MT1-MMP might cleave fibrillar collagens, proteoglycans, and various ECM glycoproteins in vitro (Ohuchi et al., 1997). The association between MT1-MMP and invasion of astrocytic tumors in the CNS has been studied in detail (Belien et al., 1999; Nakada et al., 1999). MT1-MMP enables invasive migration of glioma cells not only through proMMP2 activation, but also by direct degradation of proteins that inhibit cell migration in CNS (Belien et al., 1999). MT1-MMP is expressed constitutively at low level in normal ONH tissues and in type 1B astrocytes in culture. Increased expression of this enzyme was

214

AGAPOVA ET AL.

evident in the glaucomatous ONH and was associated with reactive astrocytes. MT1-MMP may be a component of the transition of quiescent astrocytes to the reactive phenotype in glaucoma, which requires a change in cell shape, detachment from the basement membrane, and migration through the ECM. MT1MMP activities are likely to be involved in this transition by degrading cell surface adhesion molecules and, through disruption of the extracellular domains, by altering the cytoskeleton (Werb, 1997). Expression of MT1-MMP and proMMP2 may be required to localize activity of MMP2 to the cell surface to accomplish this transition (Chen and Wang, 1999). MMP2 or gelatinase A degrades gelatin, type IV collagen, and elastin (Matrisian, 1992; Yu et al., 1998). The ubiquitous distribution, lack of significant transcriptional regulation, and specific cellular mechanism for the control of MMP2 activation make this enzyme unique among matrixins (Yu et al., 1998). As in other tissues, MMP2 is expressed constantly in brain. Low level of MMP2 mRNA expression was demonstrated in the mouse brain (Pagenstecher et al., 1998), cultured rat brain astrocytes (Wells et al., 1996), and in human brain neurons and astrocytes (Del Bigio and Jacque, 1995; Anthony et al., 1997). MMP2 expression in brain neurons and astrocytes did not change after experimental cerebral focal ischemia in rats or in human brains with multiple sclerosis or stroke (Del Bigio and Jacque, 1995; Anthony et al., 1997; Romanic et al., 1998). MMP2 was detected in astrocytes and axons in ONH tissues and in RGC in retina at the same level in normal and glaucomatous eyes. The similar levels of MMP2 in normal and glaucomatous ONH may imply that MMP2 in ONH, as in other tissues, plays important role in tissue homeostasis in normal and pathological conditions. A previous report described the localization of MMP1, MMP2, and MMP3 in human eyes with glaucoma (Yan et al., 2000). Our results confirm the presence of MMP2 and MMP1 in human normal and glaucomatous tissues. In our study, however, the presence of MMP3 or stromelysin 1 was not detectable in neural tissues. MMP3 immunoreactivity was limited to perivascular cells in the ONH, perhaps smooth muscle cells, or microglia (Neufeld, 1999). In addition, MMP3 activities and mRNA were also undetectable in cultured ONH astrocytes with or without PMA stimulation. MMP9 or gelatinase B has substrate specificity similar to MMP2 (Matrisian, 1992) and was first identified as a product of neutrophils and macrophages (Hibbs, et al., 1987). In the CNS, MMP9 has been detected in endothelial cells and in infiltrating neutrophils and macrophages after experimental rat brain injury (Rosenberg et al., 1998) and in human brain in multiple sclerosis and stroke (Anthony et al., 1997). MMP9 expression in the CNS correlates with the opening of the blood-brain barrier and inflammation (Mun-Bryce and Rosenberg, 1998; Pagenstecher et al., 1998). During CNS inflammation, MMP9 and MMP12 were expressed within the inflammatory lesions, most likely by

infiltrating leukocytes or activated microglia (Pagenstecher et al., 1998). MMP7, MMP9, and MMP12 were not detectable in normal and glaucomatous ONH or in cultured type 1B astrocytes. The absence of these enzymes suggests the lack of inflammatory cells and neovascularization in glaucomatous ONH. TIMPs are the major endogenous regulators of MMP activities in tissues. TIMP1 inhibits active MMP1, MMP3, and MMP9 and also binds to proMMP9, whereas TIMP2 preferentially binds to MMP2. TIMPs are multifunctional proteins that also can act as growth factors and inhibitors of angiogenesis (Gomez et al., 1997). In adult rat brain, TIMP1 and TIMP2 mRNA but not TIMP3 or TIMP4 were markedly increased after stab injury. In situ hybridization demonstrated increased expression of TIMP1 mRNA in reactive astrocytes and TIMP2 mRNA in microglia and neurons under these experimental conditions (Jaworski, 2000). TIMP1 neuronal/axonal localization was reported in comparable levels in both the control and ischemic cortical region of rat brain by immunohistochemistry (Romanic et al., 1998). In our study, TIMP1 and TIMP2 were expressed at high levels in the ONH and in cultured astrocytes. In ONH, TIMP1 and TIMP2 are present in both astrocytes and the nonmyelinated axons of the RGCs. Furthermore, their localization to the RGC soma suggests that both TIMPs are synthesized in the RGC bodies and transported to the axons. TIMP1 and TIMP2 expressed by RGCs can protect nerve axons in the ONH from structural damage, due to MMP proteolysis, in normal and pathological conditions. In ONH astrocytes, TIMP2 may be important for activation of membrane-bound proMMP2. Cultured type 1B astrocytes expressed only proMMP2 and active MMP2 was detected in media only after APMA activation in vitro. Considerably high level of TIMP2 expression by type 1B astrocytes may explain the lack of active MMP2 in culture. Growth factors and cytokines regulate MMP and TIMP expression, whereas MMPs can release growth factors from the storage sites in the ECM and activate the growth factor/cytokines or their receptors (Gottschall and Deb, 1996; Alexander et al., 1998; Qin et al., 1998; Rooprai et al., 2000; Yu and Stamenkovic, 2000). TGF␤2 is a multifunctional cytokine that is present at high level in the glaucomatous optic nerve head (Pena et al., 1999). Latent TGF␤ activation can be mediated by proteolysis of ECM-binding proteins by MMPs (Yu and Stamenkovic, 2000). Perhaps MMP2 in ONH is required to release and activate TGF␤2, which in turn may increase ECM synthesis and inhibit MMPs and thus regulate tissue homeostasis. TNF␣ has also been detected in the glaucomatous ONH (Yan et al., 2000; Yuan and Neufeld, 2000). TNF␣ is a potent cytokine that may play a neurodestructive role in glaucoma. In a rat model of experimental brain damage, TNF␣ induced marked upregulation of MMPs that cause neuronal damage, which can be reduced by treatment with MMP inhibitors (Leib et al., 2000). Thus, the regulation and interplay of MMPs and TIMPs with


growth factors and cytokines in the glaucomatous ONH represent important areas for future research. In conclusion, the data described here represent the initial characterization of the role of MMPs and TIMPs in the pathophysiology of glaucomatous optic neuropathy. It is well known that elevated IOP is a major risk factor in glaucoma. IOP-related stress has been shown to be considerable at the tissue level (Bellezza et al., 2000). Induction of MT1-MMP in response to stretch has been reported using cultured cardiac fibroblasts (Tyagi et al., 1998). Our laboratory demonstrated that hydrostatic pressure applied in vitro and that elevated IOP in a monkey model of glaucoma induced increased expression and deposition of elastin by astrocytes in the optic nerve head (Hernandez et al., 2000; Pena et al., 2000). Perhaps elevated IOP in glaucoma can induce increased expression of MT-MMP1 leading to detachment of astrocytes from the underlying basement membranes and migration through the ECM. Subsequently, MMP1 permits migration of astrocytes throughout the ECM of the lamina cribrosa into the nerve bundles, where MMP1, if not counterbalanced by TIMP1, will continue to degrade the scant ECM around the axons and interfere with axon survival. Our data suggest that feedback regulation of MMP activity or expression in reactive astrocyte occurs by interactions between MMPs, TIMPs, growth factors/cytokines, and ECM substrates with astrocyte membrane or intracellular components. Our results are consistent with the pathophysiology of POAG, a chronic neurodegenerative disease that spans years, in which focal areas of damage to the optic nerve axons expand slowly, leading to RGC lost and blindness. ACKNOWLEDGMENTS The authors thank Mrs. Belinda McMahan and Ms. Ping Yang for excellent technical assistance. The National Disease Research Interchange (NDRI) and the Glaucoma Foundation (San Francisco, CA) have provided the human eyes used in this study. The authors also thank Dr. William C. Parks, Washington University School of Medicine, for helpful discussion of our experimental results and for the gift of the MMP7 antibodies and human MMPs cDNA; Dr. Steven D. Shapiro, Washington University School of Medicine, for the gift of the MMP12 antibodies; and Dr. Ted S. Acott, Oregon Health Sciences University, for the gift of the MT1-MMP and TIMP2 PCR primers. Supported by an unrestricted grant from the National Research to Prevent Blindness (RPB) to the Department of Ophthalmology and Visual Sciences, Washington University School of Medicine. REFERENCES Alexander JP, Samples JR, Acott TS. 1998. Growth factor and cytokine modulation of trabecular meshwork matrix metalloproteinase and TIMP expression. Curr Eye Res 17:276 –285.

215

Anderson DA. 1969. Ultrastructure of human and monkey lamina cribrosa and optic nerve head. Arch Ophthalmology 82:800 – 814. Anthony DC, Ferguson B, Matyzak MK, Miller KM, Esiri MM, Perry VH. 1997. Differential matrix metalloproteinase expression in cases of multiple sclerosis and stroke. Neuropathol Appl Neurobiol 23: 406 – 415. Belien AT, Paganetti PA, Schwab ME. 1999. Membrane-type 1 matrix metalloprotease (MT1-MMP) enables invasive migration of glioma cells in central nervous system white matter. J Cell Biol 144:373– 384. Bellezza AJ, Hart RT, Burgoyne CF. 2000. The optic nerve head as a biomechanical structure: initial finite element modeling. Invest Ophthalmol Vis Sci 41:2991–3000. Borden P, Heller RA. 1997. Transcriptional control of matrix metalloproteinases and the tissue inhibitors of matrix metalloproteinases. Crit Rev Eukaryot Gene Expr 7:159 –178. Chen WT, Wang JY. 1999. Specialized surface protrusions of invasive cells, invadopodia and lamellipodia, have differential MT1-MMP, MMP-2, and TIMP-2 localization. Ann NY Acad Sci 878:361–371. Del Bigio MR, Jacque CM. 1995. Localization of proteinase expression in the developing rabbit brain. Brain Res Dev Brain Res 86:345– 347. Douglas DA, Shi YE, Sang QA. 1997. Computational sequence analysis of the tissue inhibitor of metalloproteinase family. J Protein Chem 16:237–255. Favaloro J, Treisman R, Kamen R. 1980. Transcription maps of polyoma virus-specific RNA: analysis by two-dimensional nuclease S1 gel mapping. Methods Enzymol 65:718 –749. Fini ME, Cook JR, Mohan R, Brinckerhoff CE. 1998. Regulation of matrix metalloproteinase gene expression. In: Parks WC, Mecham RP, editors. Matrix metalloproteinases. New York: Academic Press. p 299 –356. Gomez DE, Alonso DF, Yoshiji H, Thorgeirsson UP. 1997. Tissue inhibitors of metalloproteinases: structure, regulation and biological functions. Eur J Cell Biol 74:111–122. Gottschall PE, Deb S. 1996. Regulation of matrix metalloproteinase expressions in astrocytes, microglia and neurons. Neuroimmunomodulation 3:69 –75. Hernandez MR. 1992. Ultrastructural immunocytochemical analysis of elastin in the human lamina cribrosa: changes in elastic fibers in primary open-angle glaucoma. Invest Ophthalmol Vis Sci 33:2891– 2903. Hernandez MR, Pena JD. 1997. The optic nerve head in glaucomatous optic neuropathy. Arch Ophthalmol 115:389 –395. Hernandez MR. 2000. The optic nerve head in glaucoma: role of astrocytes in tissue remodeling. Prog Retin Eye Res 19:297–321. Hernandez MR, Pena JD, Selvidge JA, Salvador-Silva M, Yang P. 2000. Hydrostatic pressure stimulates synthesis of elastin in cultired optic nerve head astrocytes. Glia 32:122–136. Hibbs MS, Hoidal JR, Kang AH. 1987. Expression of a metalloproteinase that degrades native type V collagen and denatured collagens by cultured human alveolar macrophages. J Clin Invest 80: 1644 –1650. Imper V, Van Wart HE. 1998. Substrate specificity and mechanisms of substrate recognition of the matrix metalloproteinases. In: Parks WC, Mecham RP, editors. Matrix metalloproteinases. New York: Academic Press. p 219 –242. Jaworski DM. 2000. Differential regulation of tissue inhibitor of metalloproteinase mRNA expression in response to intracranial injury. Glia 30:199 –208. Kobayashi S, Vidal I, Pena JD, Hernandez MR. 1997. Expression of neural cell adhesion molecule (NCAM) characterizes a subpopulation of type 1 astrocytes in human optic nerve head. Glia 20:262–273. Leib SL, Leppert D, Clements J, Tauber MG. 2000. Matrix metalloproteinases contribute to brain damage in experimental pneumococcal meningitis. Infect Immun 68:615– 620. Matrisian LM. 1992. The matrix-degrading metalloproteinases. Bioessays 14:455– 463. Mun-Bryce S, Rosenberg GA. 1998. Gelatinase B modulates selective opening of the blood-brain barrier during inflamation. Am J Physiol 274:R1203–R1211. Nagase H. 1997. Activation mechanisms of matrix metalloproteinases. Biol Chem 378:151–160. Nagase H, Woessner JF Jr. 1999. Matrix metalloproteinases. J Biol Chem 274:21491–21494. Nakada M, Nakamura H, Ikeda E, Fujimoto N, Yamashita J, Sato H, Seiki M, Okada Y. 1999. Expression and tissue localization of membrane-type 1, 2, and 3 matrix metalloproteinases in human astrocytic tumors. Am J Pathol 154:417– 428. Neufeld AH. 1999. Microglia in the optic nerve head and the region of parapapillary chorioretinal atrophy in glaucoma. Arch Ophthalmol 117:1050 –1056.

216

AGAPOVA ET AL.

Ohuchi E, Imai K, Fujii Y, Sato H, Seiki M, Okada Y. 1997. Membrane type 1 matrix metalloproteinase digests interstitial collagens and other extracellular matrix macromolecules. J Biol Chem 272:2446 – 2451. Pagenstecher A, Stalder AK, Kincaid CL, Shapiro SD, Campbell IL. 1998. Differential expression of matrix metalloproteinase and tissue inhibitor of matrix metalloproteinase genes in the mouse central nervous system in normal and inflammatory states. Am J Pathol 152:729 –741. Pena JD, Netland PA, Vidal I, Dorr DA, Rasky A, Hernandez MR. 1998. Elastosis of the lamina cribrosa in glaucomatous optic neuropathy. Exp Eye Res 67:517–524. Pena JD, Taylor AW, Ricard CS, Vidal I, Hernandez MR. 1999. Transforming growth factor beta isoforms in human optic nerve heads. Br J Ophthalmol 83:209 –218. Pena JD, Kaufman PL, Kiland JA, Gabelt BT, Hernandez MR. 2000. Elastin expression in the lamina cribrosa of monkeys with experimental glaucoma and with optic nerve transection. Invest Ophthalmol Vis Sci 41:S949. Perry VH, Lund RD. 1990. Evidence that the lamina cribrosa prevents intraretinal myelination of retinal ganglion cell axons. J Neurocytol 19:265–272. Qin H, Moellinger JD, Wells A, Windsor LJ, Sun Y, Benveniste EN. 1998. Transcriptional suppression of matrix metalloproteinase-2 gene expression in human astroglioma cells by TNF-alpha and IFN-gamma. J Immunol 161:6664 – 6673. Quigley HA, Hohman RM, Addicks EM, Massof RW, Green WR. 1983. Morphologic changes in the lamina cribrosa correlated with neural loss in open-angle glaucoma. Am J Ophthalmol 95:673– 691. Quigley HA, Dorman-Pease ME, Brown AE. 1991. Quantitative study of collagen and elastin of the optic nerve head and sclera in human and experimental monkey glaucoma. Curr Eye Res 10:877– 888. Romanic AM, White RF, Arleth AJ, Ohlstein EH, Barone FC. 1998. Matrix metalloproteinase expression increases after cerebral focal ischemia in rats: inhibition of matrix metalloproteinase-9 reduces infarct size. Stroke 29:1020 –1030. Rooprai HK, Rucklidge GJ, Panou C, Pilkington GJ. 2000. The effects of exogenous growth factors on matrix metalloproteinase secretion by human brain tumour cells. Br J Cancer 82:52–55.

Rosenberg GA, Estrada EY, Dencoff JE. 1998. Matrix metalloproteinases and TIMPs are associated with blood-brain barrier opening after reperfusion in rat brain. Stroke 29:2189 –2195. Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA, Goldberg GI. 1995. Mechanism of cell surface activation of 72-kDa type IV collagenase. Isolation of the activated form of the membrane metalloprotease. J Biol Chem 270:5331–5338. Tyagi SC, Lewis K, Pikes D, Marcello A, Mujumdar VS, Smiley LM, Moore CK. 1998. Stretch-induced membrane type matrix metalloproteinase and tissue plasminogen activator in cardiac fibroblast cells. J Cell Physiol 176:374 –382. Varela HJ, Hernandez MR. 1997. Astrocyte responses in human optic nerve head with primary open-angle glaucoma. J Glaucoma 6:303– 313. Vos CM, Sjulson L, Nath A, McArthur JC, Pardo CA, Rothstein J, Conant K. 2000. Cytotoxicity by matrix metalloprotease-1 in organotypic spinal cord and dissociated neuronal cultures. Exp Neurol 163:324 –330. Wells GM, Catlin G, Cossins JA, Mangan M, Ward GA, Miller KM, Clements JM. 1996. Quantitation of matrix metalloproteinases in cultured rat astrocytes using the polymerase chain reaction with a multi-competitor cDNA standard. Glia 18:332–340. Werb Z. 1997. ECM and cell surface proteolysis: regulating cellular ecology. Cell 91:439 – 442. Woessner JF Jr. 1991. Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEB J 5:2145–2154. Yan X, Tezel G, Wax MB, Edward DP. 2000. Matrix metalloproteinases and tumor necrosis factor alpha in glaucomatous optic nerve head. Arch Ophthalmol 118:666 – 673. Yu AE, Murphy AN, Stetler-Stevenson WG. 1998. 72-kDa gelatinase (gelatinase A): structure, activation,regulation, and substrate specificity. In: Parks WC, Mecham RP, editors. Matrix metalloproteinases. New York: Academic Press. p 85–113. Yu Q, Stamenkovic I. 2000. Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor invasion and angiogenesis. Genes Dev 14:163–176. Yuan L, Neufeld AH. 2000. Tumor necrosis factor-alpha: a potentially neurodestructive cytokine produced by glia in the human glaucomatous optic nerve head. Glia 32:42–50.