Bone Morphogenetic Protein-7 Is an Antagonist of Transforming Growth Factor-β2 in Human Trabecular Meshwork Cells

Abstract

purpose. The increase in intraocular pressure in primary open-angle glaucoma (POAG) may involve transforming growth factor (TGF)-β2 signaling, as TGF-β2 is found in higher amounts than normal in the aqueous humor of patients with POAG. In vitro, TGF-β2 causes an accumulation of extracellular matrix (ECM) in the trabecular meshwork (TM) and an increase in TM outflow resistance. The present study was undertaken to determine whether bone morphogenetic protein (BMP)-7 signaling antagonizes the effects of TGF-β2 on TM cells.

methods. Cultured TM cells from nine human donors were treated with BMP-7, TGF-β2, or a combination of both for 24 or 72 hours. The expression of connective tissue growth factor (CTGF); thrombospondin (TSP)-1; fibronectin; collagen types I, III, and IV; plasminogen activator inhibitor (PAI)-1; and matrix metalloproteinase (MMP)-2 were analyzed by immunohistochemistry, real time RT-PCR, Western and Northern blot analysis, and zymography (MMP-2).

results. Treatment with TGF-β2 induced the expression of CTGF, TSP-1, fibronectin, collagen types IV and VI, and PAI-1. All these effects were inhibited when TGF-β2 was added in combination with BMP-7, whereas BMP-7 alone had no effects. Treatment with TGF-β2, BMP-7, or the combination of both had no effect on the expression of collagen types I and III.

conclusions. BMP-7 strongly antagonizes in vitro the TGF-β-induced expression of a broad panel of molecules, which would result in an accumulation of TM ECM in situ. As BMP-7 is expressed in the adult human TM in situ, it seems reasonable to assume that it similarly modulates and antagonizes the effects of TGF-β2 signaling on the tissues of the outflow pathways in vivo. The pharmacological modulation of BMP-7 signaling in the TM might be a promising strategy to treat POAG.

An increase in intraocular pressure that leads to damage of the optic nerve head is the critical risk factor for primary open-angle glaucoma (POAG), ^{1 2 3} a major cause of blindness worldwide. ⁴ Intraocular pressure is increased because of an abnormally high aqueous humor outflow resistance in the trabecular meshwork (TM). ⁵ The mechanisms that are responsible for the increase in TM outflow resistance in POAG are unclear, but there is some evidence that changes in the amount and quality of the TM extracellular matrix (ECM) are involved. Eyes with POAG show a significant increase in extracellular “plaque material” in the TM. ^{6 7} The molecular nature of plaque material is unclear, but there is some evidence that collagen type VI is associated with it. ⁸ Other components of the ECM that have been found in higher amounts in the TM of eyes with POAG are collagen type IV, laminin, and fibronectin. ^{9 10}  

There is limited information on the nature of factors that modulate ECM turnover in the normal TM and its increase in POAG. Transforming growth factor (TGF)-β2 may be involved, as several independent studies reported on a higher than normal concentration of TGF-β2 in the aqueous humor of patients with POAG. ^{11 12 13 14} In a variety of disorders, TGF-β signaling mediates fibrosis and a pathologic increase in ECM deposition. ^{15 16 17} In vitro studies suggest a similar role of TGF-β2 for the increase in TM ECM deposition in POAG. Treatment of cultured human TM cells with TGF-β2 causes an increase in fibronectin synthesis and a substantial and irreversible cross-linking of fibronectin by the action of tissue transglutaminase. ^{18 19} In addition, we have shown that treatment of human TM cells in vitro decreases the activity of matrix metalloproteinases (MMPs), which normally degrade ECM compounds. ²⁰ In anterior segment perfusion cultures, perfusion with TGF-β2 promotes a focal accumulation of fine fibrillar extracellular material in the TM, an effect that correlates with a reduction in outflow facility. ²¹ TGF-β2 signaling in POAG may also contribute to structural changes in the posterior eye, as affected patients show an increase in fibrillar ECM in the optic nerve head, ²² and treatment of cultured optic nerve astrocytes with TGF-β2 increases the expression of ECM molecules. ²³  

It is reasonable to assume that the fibrogenic action of TGF-β2 is subject to multiple control mechanisms in situ. Most of the TGF-β2 in the aqueous humor of normal eyes and in POAG eyes is found in a latent, inactive form that is unable to interact with cellular receptors. The physiological mechanisms of TGF-β activation are not well understood. A very potent activator of latent TGF-β in vivo and in vitro is the matricellular protein thrombospondin-1, which is expressed in the TM in situ. ²⁴ Recent studies suggest another mechanism to modulate the action of TGF-βs, which involves counteracting TGF-β action through direct antagonism involving Smad signaling pathways. ²⁵ There is considerable evidence that bone morphogenetic protein (BMP)-7, a 35-kDa homodimeric protein and member of the TGF-β superfamily of cysteine knot cytokines, is involved in such counteracting activity and serves as an endogenous antagonist of TGF-β. In a corneal alkali injury model, BMP-7 has been shown to suppress TGF-β-induced effects on corneal scarring. ²⁶ In the kidney, BMP-7 counteracts an epithelial-to-mesenchymal transition that is induced by TGF-β1 and reverses chronic renal injury. ²⁷ In addition, BMP-7 antagonizes the TGF-β-dependent fibrogenesis in mesangial cells. ^{28 29} In embryonic life, BMP-7 plays a critical role during renal and eye development. The targeted inactivation of BMP-7 leads to severe eye defects that show variable penetrance and depend on the genetic background. ^{30 31} In the adult organism, the expression of BMP-7 is retained in a few tissues, including the eye, in which the expression of BMP-7 and its receptors has been shown in cornea, trabecular meshwork, and optic nerve head. ^{32 33}  

In the present study, we investigated whether BMP-7 has the potential to reduce the accumulation of ECM in the human outflow pathways during POAG and antagonizes the fibrogenic action of TGF-β on TM cells. 

Materials and Methods

Cell Cultures

Cultures of human trabecular meshwork (HTM) cells were established from the eyes of nine human donors according to protocols published previously. ^{20 34} HTM cells of the third passage were seeded in 35-mm culture wells (4.0 × 10⁵ cells per well) and grown to a confluent monolayer. After 7 days of confluence, wells were incubated in serum-free medium F-10 (Invitrogen, Karlsruhe, Germany) for 24 hours followed by incubation in fresh serum-free medium supplemented with 300 pM BMP-7, 300 pM TGF-β2 (both R&D Systems GmbH, Wiesbaden, Germany)_, or a combination of both for 72 hours. Treated cells were compared with those from control dishes that were incubated under identical conditions for 72 hours but without supplementation with TGF-β2 or BMP-7 or with those that had been treated with a combination of TGF-β2 and BMP-7 in the presence of neutralizing antibodies against BMP-7 (5 μg/mL murine monoclonal anti-human BMP-7 IgG; R&D Systems). To investigate the effects of different doses and treatment times, we treated HTM cells with TGF-β2 at a concentration of either 100 or 300 pM and at the same time with BMP-7 at concentrations of 50 to 300 pM for 24 and 72 hours, respectively. 

Methods for securing human tissue were humane, included proper consent and approval, and complied with the Declaration of Helsinki. 

RNA Isolation and Polymerase Chain Reaction

HTM cells were harvested from culture wells, and total RNA was extracted with an RNA isolation kit according to the manufacturer’s protocol (Stratagene, Heidelberg, Germany). Structural integrity of the RNA samples was confirmed by electrophoresis in 1% Tris acetate EDTA (TAE) agarose gels and visualization of ribosomal RNA by staining with ethidium bromide. Yield and purity of RNA samples were determined on a spectrophotometer. For generation of gene-specific antisense Northern blot probes, first strand complementary DNA (cDNA) for PCR was prepared from total RNA by using Moloney murine leukemia virus reverse transcriptase (M-MLV-RT) and oligo (dT)-17 primers. RT-PCR amplification of gene-specific probes was performed with a temperature profile as follows: 36 cycles of 1-minute denaturation at 94°C; 1 minute annealing (for specific temperatures, see Table 1 ) and 2 minutes of extension at 72°C, followed by an end extension step of 10 minutes at 72°C after the last cycle. All PCR primers were purchased from Metabion (Planegg-Martinsried, Germany) and crossed exon–intron boundaries. Primer sequences, positions, annealing temperatures, and PCR product sizes of primer pairs are shown in Table 1 . Sizes of PCR products were estimated by comparing with the migration of a DNA size marker run in parallel (100 bp DNA ladder; Promega, Madison, WI) on a 1% TAE-agarose gel. Non–reverse transcribed RNA served as the negative control for RT-PCR and showed no amplification products (data not shown). PCR products were purified with a PCR purification kit (Qiagen, Hilden, Germany), cloned into the vector pTOPO/TA (Invitrogen) and send to Seqlab (Regensburg, Germany) for sequence analysis. 

Northern Blot Analysis

Total RNA (2 μg) were size fractionated by gel electrophoresis in 1% agarose gels containing 2.2 M formaldehyde, blotted onto a positive charged nylon membrane (Roche, Basel, Switzerland), and cross-linked at 1600 μJ in a Stratalinker (Stratagene). To assess the amount and quality of the RNA, membranes were stained with methylene blue, and images were obtained on an imaging workstation (LAS3000; Raytest, Straubenhardt, Germany). Prehybridization was performed at 68°C for 1 hour (Dig Easy Hyb; Roche, Mannheim, Germany). Hybridizations were performed at 68°C overnight in prehybridization solution containing 50 ng/mL of specific antisense probe. Antisense RNA probes for Northern blot analysis were generated by linearization of pTOPO/TA vectors containing appropriate PCR products (described earlier) and labeling with DIG-11-UTP with T7-polymerase (Roche). After hybridization, the membranes were washed twice with 2× SSC/0.1% SDS at room temperature for 1 hour, followed by two washes in 0.1% SDS at room temperature for 1 hour or 15 minutes at 68°C, respectively. Afterward, the membranes were washed for 5 minutes in washing buffer (100 mM maleic acid, 150 mM NaCl, and 0.3% Tween-20, pH 7.5) and incubated for 1 hour in blocking solution (100 mM maleic acid, 150 mM NaCl, and 1% blocking reagent; pH 7.5 [Roche]). Anti-digoxygenin-alkaline-phosphatase (Roche) was diluted 1:10,000 in blocking solution and added to the membranes for 30 minutes. After incubation, the membranes were washed four times for 15 minutes each in washing buffer and equilibrated in detection buffer (100 mM Tris-HCl, 100 mM NaCl; pH 9.5) for 10 minutes. For detection by chemiluminescence, CDP-star (Roche) was diluted 1:100 in detection buffer and the membranes were incubated for 5 minutes at room temperature. After air-drying, semidry membranes were sealed in plastic bags and chemiluminescence was detected on the imaging workstation. Exposure times ranged between 2 minutes and 30 minutes, and quantification was then performed (for all quantifications, AIDA Biopackage software [Raytest] was used). 

Western Blot Analysis

The culture medium of treated HTM cells was collected, and aliquots (500 μL) were concentrated 50-fold by centrifugation through a tube (10 kDa cutoff; Vivaspin500; Vivascience, Hannover, Germany), according to the manufacturer’s instructions. Protein contents of the probes were determined by a Bradford protein assay (Bio-Rad, Munich, Germany). To obtain protein extracts of cells grown on tissue culture dishes, cells were directly lysed in RIPA lysis buffer (150 mM NaCl, 1% NP-40, 0.5% DOC, 0.1% SDS, 50 mM Tris [pH 8]) and protein content was measured using the BCA protein assay reagent (Pierce, Rockford, IL). All probes were supplemented with SDS-loading buffer and denatured by boiling for 5 minutes. Two micrograms of each sample were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and transferred onto a polyvinyl difluoride membrane (PVDF; Roche) either by semidry blotting (CTGF) or by tank blot (collagens I, III, IV, and VI; fibronectin; and thrombospondin-1) at 70 V for 45 minutes in 1× transfer buffer (10 mM CAPS [3-[cyclohexylamino]-1 propane sulfonic acid], 20% methanol, 0.1% SDS, [pH 11]). Membranes were blocked in PBST/BSA (phosphate-buffered saline, 0.1% Tween-20, 5% bovine serum albumin, [pH 7.2]) for 1 hour. Primary antibodies were added in PBST (dilutions see Table 2 ) and allowed to react overnight at 4°C. After washing with PBST, secondary antibodies were added in PBST at the appropriate dilution (see Table 2 ) for 30 minutes at room temperature. For detection, CDP-star was diluted 1:100 in detection buffer, and the membranes were incubated for 5 minutes at room temperature. Chemiluminescence signals were analyzed on the imaging workstation (exposure times ranged between 1 and 5 minutes), followed by quantification. 

Gelatin Zymography

Aliquots (10 μL) of conditioned medium collected from treated and untreated HTM cells were immediately analyzed for gelatinolytic activity by electrophoresis at 4°C in 10% SDS polyacrylamide cross-linked gels, containing 0.1% gelatin or β-casein (Sigma-Aldrich, Steinheim, Germany). After electrophoresis, gels were washed with 2.5% Triton X-100 and incubated in Tris-HCl, 0.5 mM CaCl₂, 10⁻⁶ M ZnCl₂ (pH 8.0), at 37°C for 16 hours. Subsequently, Coomassie brilliant blue staining was performed, and gels were washed in 5% acetic acid and 10% methanol in water. The presence of MMPs appeared as unstained bands in gelatin zymography. Quantification was then performed. 

Immunohistochemistry

HTM cells were grown on microscopic slides and treated as described earlier. After incubation, cells were fixed with 4% paraformaldehyde for 15 minutes and subsequently washed twice with PBS containing 0.1% Triton X-100. Primary antibodies were added in appropriate dilutions in PBS/BSA (5%; Table 2 ) and allowed to bind for 4 hours at room temperature. After the cells were washed three times with PBS, fluorescein-conjugated secondary antibodies were added (Table 2)for 1 hour at room temperature. 4′,6-Diamidino-2-phenylindole (DAPI) was used to counterstain nuclear DNA. Slides were mounted in medium containing DAPI (Vectashield; Vector Laboratories, Burlington, CA), and analyzed under a fluorescence microscope (Zeiss Axio Imager; Carl Zeiss AG, Oberkochen, Germany). To control for nonspecific binding of secondary antibodies, negative control experiments were performed, which were handled similarly, but the cells were incubated in PBS/BSA without primary antibody. 

Real-Time PCR Analysis

Semiquantification by real-time PCR was performed on a real-time thermal analyzer (Rotor-Gene 2000; Corbett Life Science, Mortlake, VIC, Australia). The S9 gene served as an endogenous control to normalize the differences in the amount of cDNA in each sample. Taq polymerase (Hot Star; Qiagen) was used for PCR reaction according to the manufacturer’s protocol. PCR reaction was performed in a volume of 25 μL, consisting of 2.5 μL of 10× PCR buffer, 2.0 to 2.5 μL of MgCl₂ (25 mM), 0.5 μL of dNTPs (10 mM each; Promega), 0.5 μL of Taq (5 U/μL), 0.5 μL of primer mix (20 μM each), and 2.5 μL of 1× SYBR green I solution (Sigma-Aldrich). 

All samples that had to be compared for expression differences were run in the same assay as duplicates. After completion of PCR amplification, data were analyzed (Rotor-Gene Software, ver. 6.0; Corbett Life Science). Data were initially expressed as a threshold cycle and are expressed as increases (x-fold) in gene expression in untreated HTM cells compared with the expression of the different treatments investigated. For each experiment, the mean value of untreated cells was set at 100%. Each individual experiment was repeated four times and mean values and standard deviations were calculated. After amplification was complete, the PCR products were analyzed by agarose gel electrophoresis. Sequences of primers and PCR product sizes of primer pairs are given in Table 1 . 

Number of Experiments and Statistical Analysis

To assess the effects of a 72-hour treatment with 300 pM BMP-7 and/or 300 pM TGF-β2, each Northern and Western blot experiment was repeated at least six times with RNA or protein extract from primary HTM cell lines of six different donors, respectively. Each real-time RT-PCR analysis was performed in duplicate and repeated at least four times with RNA from HTM cell lines of four different donors. All experiments to study the effects of lower doses and shorter treatment times were performed at least three times using RNA and protein extract from primary HTM cell lines from three different donors, respectively. Student’s t-test was used for statistical analysis of the RNA data. 

Results

Connective Tissue Growth Factor and Thrombospondin-1

Connective tissue growth factor (CTGF) and thrombospondin (TSP)-1 are prominent target genes of TGF-β1 and -2. ^{23 24} CTGF mediates the effects of TGF-βs on ECM turnover, ³⁵ whereas TSP-1 is a major endogenous activator of latent TGF-βs. ³⁶ To investigate the expression of CTGF and TSP-1 in HTM cells and the influences of TGF-β2 and BMP-7, Western and Northern blot analyses were performed. Western blot analysis with antibodies specific for CTGF showed two major bands migrating at approximately 36 and 38 kDa in HTM cell culture medium (Fig. 1A) , corresponding to the molecular weight of CTGF and its modified form. ^{37 38} After treatment with 300 pM TGF-β2 for 72 hours, a marked increase in secreted CTGF was observed (Fig. 1A) . Densitometric analysis of multiple experiments showed 3.6 ± 0.4-fold more CTGF expression in TGF-β2-treated cells compared with the control. In contrast, when HTM cells were treated with a combination of TGF-β2 and BMP-7 in like amounts (300 pM) and for like periods (72 hours), the amount of CTGF that was detected in cell culture medium was markedly lower (1.2 ± 0.2) than after treatment with TGF-β2 alone (Fig. 1A) . Treatment with 300 pM BMP-7 alone appeared to have no obvious effects on CTGF production in HTM cells, as the amount of CTGF, which was detected in HTM cell culture medium was comparable to that of control cultures (Fig. 1A) . Corresponding results were observed by Northern blot analysis. Using a probe specific for CTGF mRNA, a hybridization signal corresponding to the expected transcript of 2.1 kb was observed in HTM cells (Fig. 1B) . After treatment with TGF-β2, a substantial increase in intensity of the CTGF-specific hybridization signal was observed (Fig. 1B) . Densitometric analysis of multiple experiments showed 4.1 ± 0.8-fold more CTGF mRNA in TGF-β2-treated cells compared with the control. In contrast, after combined treatment with TGF-β2 and BMP-7, the increase in CTGF mRNA was significantly lower (1.4 ± 0.3, P < 0.01; Fig. 1B ). No increase in CTGF mRNA was detected after treatment with BMP-7 alone (Fig. 1B) . To analyze whether the observed effects of BMP-7 on TGF-β2 signaling were due to specific BMP-7 activity, we added neutralizing antibodies against BMP-7 when HTM cells were treated with a combination of TGF-β2 and BMP-7. Under these conditions, no antagonizing effects of BMP-7 on the TGF-β2-induced increase in CTGF expression were observed (Fig. 1C) , strongly indicating a specific action of BMP-7. 

When cell culture medium of HTM cells was assessed for the presence of TSP-1, a specific signal at approximately 180 kDa was observed by Western blot analysis (Fig. 1D) , corroborating data from previous studies. ²⁴ After treatment with TGF-β2, higher amounts of TSP-1 were detected in the culture medium of HTM cells compared with that of untreated cells (2.1 ± 0.7, Fig. 1D ). In contrast, no increase in TSP-1 was observed after combined treatment with TGF-β2 and BMP-7, or treatment with BMP-7 alone (Fig. 1D) . Northern blot analysis confirmed the effects of TGF-β2 and BMP-7 on TSP-1 expression in HTM cells. After TGF-β2 treatment, the signal for TSP-1 mRNA, which migrated at the expected size of 3.2 kb, was more intense than in the control cells (Fig. 1E) . Densitometric analysis of multiple experiments showed 2.1 ± 0.7-fold more TSP-1 mRNA in TGF-β2–treated cells. No increase in TSP-1 mRNA expression was observed after combined treatment with TGF-β2 and BMP-7 or treatment with BMP-7 alone (Fig. 1E) . The difference between combined TGF-β2 and BMP-7 or treatment with TGF-β2 alone was statistically significant (P < 0.01). 

In a subsequent step, effects of TGF-β2 and BMP-7 on the expression of CTGF and TSP-1 in HTM cells were assessed by immunohistochemistry. In control cultures, most HTM cells showed no detectable immunoreactivity for CTGF (Fig. 2A) . In those few cells that showed positive immunoreactivity, CTGF staining was observed in cellular vesicles of variable sizes (inset in Fig. 2A ). In contrast, after 72 hours of treatment with TGF-β2, CTGF immunoreactivity was found to be evenly distributed throughout the culture dish, with a positive signal in most of the cells (Fig. 2A) . Staining for CTGF was considerable weaker in cells treated for 72 hours with a combination of BMP-7 and TGF-β2 or with BMP-7 alone. The results obtained for TSP-1 were essentially comparable to that of CTGF. By immunohistochemistry with antibodies specific for TSP-1, weak cytoplasmic staining was observed in some HTM cells of control cultures, or HTM cells treated for 72 hours with a combination of BMP-7 and TGF-β2 or with BMP-7 alone. In contrast, after 72 hours of treatment with TGF-β2 alone, the immunoreactivity for TSP-1 in HTM cells was considerably stronger. No immunoreactivity for CTGF or TSP-1 was observed in control experiments, in which the respective specific antibodies were not added (data not shown). 

To analyze the effects of different doses of TGF-β2 and/or BMP-7 on TGF-β2-induced CTGF expression, we analyzed the expression of CTGF after 72 hours of treatment with 100 or 300 pM TGF-β2 in combination with 50, 100, 150, or 300 pM of BMP-7. Real-time RT-PCR analysis showed a comparable threefold induction of CTGF mRNA after 72 hours of treatment of HTM cells with 100 or 300 pM TGF-β2, respectively (Figs. 3A 3C) . In contrast, after combined treatment with TGF-β2 and BMP-7, there was a dose-dependent decrease in CTGF mRNA induction, which was almost at baseline when BMP-7 was added at equimolar concentrations (Figs. 3A 3C) . The differences between 72 hours combined BMP-7 (100, 150, or 300 pM) and TGF-β2 (100 pM) treatment, or treatment with 100 pM TGF-β2 alone were statistically significant (P < 0.01). The same was true for the difference between 72 hours of combined BMP-7 and TGF-β2 (300 pM each) treatment or treatment with 300 pM TGF-β2 alone. Corresponding results were observed by Western blot analysis (Figs. 3B 3D) . To investigate the effects of shorter treatment times, we analyzed the expression of CTGF after 24 hours of treatment with 100 or 300 pM TGF-β2 in combination with 50, 100, 150, or 300 pM BMP-7. Again, real-time RT-PCR analysis showed a threefold induction of CTGF mRNA after treatment with 100 or 300 pM TGF-β2, respectively (Figs. 3E 3G) . In contrast, CTGF mRNA remained near baseline when BMP-7 was added at equimolar concentrations (Figs. 3E 3G) . When HTM cells were treated for 24 hours with 100 pM TGF-β2 in combination with 300 pM BMP-7, the amount of CTGF mRNA was markedly reduced when compared with control levels (Fig. 3E) . The differences between 24 hours of combined treatment with BMP-7 (100, 150, or 300 pM) and TGF-β2 (100 pM), or treatment with 100 pM TGF-β2 alone were statistically significant (P < 0.01). The same was true of the differences between 24 hours of combined treatment with BMP-7 (150 or 300 pM) and TGF-β2 (300 pM), or treatment with 300 pM TGF-β2 alone. Again, corresponding results were observed in Western blot analysis (Figs. 3F 3H) . Overall, the amounts of CTGF that were detected in cell culture medium after 24 hours of treatment were lower than that after 72 hours of treatment, consisting with a time-dependent increase of CTGF after continuous stimulation with TGF-β2. 

Fibronectin

Because our results indicated that BMP-7 specifically blocks the inducing effects of TGF-β2 on the expression of CTGF, a critical downstream mediator of fibrogenic TGF-β signaling, ³⁷ we wanted to determine whether BMP-7 has the potential to block a TGF-β2-induced accumulation of ECM compounds in HTM cells. In previous work, we had shown that the expression of fibronectin, a major component of the ECM in the juxtacanalicular region of HTM in situ, ^{9 10 39} is increased after treatment of HTM cells with TGF-β2. ¹⁸ Comparable results were obtained in the present study. When immunohistochemistry was performed to visualize fibronectin after treatment with TGF-β2, a dense network of fibrillar fibronectin was observed in the ECM surrounding individual HTM cells in culture (Fig. 2C) . In contrast, immunoreactivity for fibronectin was considerably less intense in cultures treated with a combination of BMP-7 and TGF-β2, in which the amount of fibronectin did not markedly differ from that of control cultures (Fig. 2C) . Similar results were obtained after treatment with BMP-7 alone (Fig. 2C) . The results obtained for fibrillar fibronectin correlated with those obtained for soluble fibronectin in cell culture medium. In the culture medium of HTM cells, the polymerized fibronectin complex that migrates at 240 kDa was increased after treatment with TGF-β2 (Fig. 4A) . Densitometry of multiple experiments showed a 3.2 ± 0.4-fold increase compared with control cultures. In contrast, no increase was observed when HTM cells were treated with a combination of BMP-7 and TGF-β2, or with BMP-7 alone (Fig. 4A) . The data obtained by Northern blot analysis correlated with those obtained by immunohistochemistry and Western blot analysis. In HTM cells treated with TGF-β2, a 3.5 ± 0.6 increase in fibronectin mRNA was observed compared with control cultures (Fig. 4B) . In contrast, the amount of the fibronectin transcript, which migrated at the expected size of 7.7 kb, was not changed after treatment with a combination of BMP-7 and TGF-β2, or with BMP-7 alone (Fig. 4B) . The difference between combined TGF-β2 and BMP-7 treatment or treatment with TGF-β2 alone was statistically significant (P < 0.01). 

Collagen Type IV

Collagen type IV, a critical component of basement membranes throughout the body has been found in the basement membrane of the trabecular beams and in filamentous material in the juxtacanalicular layer of the HTM. ^{9 39 40} Collagen type IV is composed of different subunits with α1(IV) and α2(IV) being those that are ubiquitously expressed. Both subunits are encoded by different genes (COL41A and COL4A2). In our experiments, real-time RT-PCR analysis showed a 4.6 ± 0.3-fold increase of COL41A mRNA and a 5.0 ± 0.6-fold increase of COL4A2 mRNA after 72 hours of treatment of HTM cells with TGF-β2 (Fig. 4D) . In contrast, the effects on COL4A1 and COL4A2 were reduced by almost 50% when BMP-7 was added in combination with TGF-β2 (Fig. 4D) . The differences between combined TGF-β2 and BMP-7 treatment or treatment with TGF-β2 alone were statistically significant (P < 0.01). In parallel Western blot experiments, collagen type IV was detected in the culture medium of HTM cells at the expected molecular weight of 230 kDa (Fig. 4C) . After treatment with TGF-β2, a 3.6 ± 0.8-fold increase in collagen type IV was observed in the cell culture medium (Fig. 4C) . Combined treatment with TGF-β2 and BMP-7 showed a markedly smaller 1.7 ± 04-fold increase in the culture medium of HTM cells when compared with that of untreated HTM cells (Fig. 4C) . No differences were observed between control cells and those treated with BMP-7 alone (Fig. 4C) . 

Collagen Type VI

Similar to fibronectin, collagen type VI is a major component of the fibrillar ECM in the juxtacanalicular region of the HTM. ^{8 41} We therefore wanted to know whether treatment with TGF-β2 has an influence on the HTM expression of collagen type VI and whether any TGF-β2-induced effects can be blocked by the addition of BMP-7. Collagen type VI consists of three subunits, each of which are encoded by different genes (COL6A1, COL6A2, and COL6A3). The regulation of COL6A3 expression has been reported to be critical for the control of collagen type VI synthesis and the deposition of the heterotrimeric molecule. ⁴² We performed real-time RT-PCR on RNA obtained from HTM cells treated with TGF-β2, TGF-β2, and BMP-7 or BMP-7 alone. Control cultures expressed mRNA for each of the subunits (Fig. 5B) . Treatment with TGF-β2 increased mRNA for COL6A2 (1.9 ± 0.3-fold) and COL6A3 (2.3 ± 0.5-fold), only mRNA for COL6A1 was not induced by TGF-β2 treatment (Fig. 5B) . Combined treatment with BMP-7 and TGF-β2 showed a much lower increase in COL6A2 (1.2 ± 0.4-fold) and COL6A3 mRNAs (1.1 ± 0.3-fold), whereas no changes in the amounts of COL6A1, COL6A2, and COL6A3 mRNAs were observed after treatment with BMP-7 alone (Fig. 5B) . The differences between combined TGF-β2 and BMP-7 treatment or treatment with TGF-β2 alone were statistically significant (P < 0.01 for COL6A3, and P < 0.05 for COL6A2). Corresponding data were obtained by Western blot analysis. In the culture medium of TGF-β2-treated HTM cells, the 220-kDA band corresponding to collagen type VI was upregulated (2.1 ± 0.3-fold) in comparison to untreated control cells (Fig. 5A) . No changes in the amount of collagen type VI were observed after combined treatment with TGF-β2 and BMP-7 or treatment with BMP-7 alone (Fig. 5A) . 

Collagen Types I and III

Collagen types I and III in the HTM are primarily localized to the striated collagen fibrils of the trabecular beams. ⁴¹ Real-time RT-PCR of RNA obtained from HTM cells treated with BMP-7, TGF-β2, or the combined treatment with both growth factors for 72 hours showed no changes in the amount of COL1A1 mRNA when compared to controls (Fig. 6B) . In Western blot analyses, a 140-kDa band migrating at the molecular weight of Col1α1 was detected in the medium of control cultures, as well in that of HTM cells treated with TGF-β2, BMP-7, and TGF-β2, or BMP-7 alone, while no changes in signal intensity of the 140-kDa band were observed (Fig. 6A) . Comparable results were observed for the expression of COL3A1. Real-time RT-PCR showed no changes in the amount of mRNA between control cultures and those treated with TGF-β2, BMP-7, and TGF-β2 or BMP-7 alone (Fig. 6B) . Similarly, the analysis of the Col3α1 content in the culture medium of HTM cells showed the same signal intensity of the specific Col3α1 band of 139 kDa, irrespective of the individual treatment protocol (Fig. 6A) . 

PAI-1 and MMP-2

PAI-1, the inhibitor of the plasminogen/plasmin system, has been shown to reduce the enzymatic activation of MMPs. Recently, we showed that PAI-1 is a target gene of TGF-β2 in HTM. ²⁰ These results were confirmed in the present study, in which PAI-1 mRNA was found to be upregulated by treatment with TGF-β2. PAI-1-specific bands were detected by Northern blot analysis at 2.2 and 3.2 kb. Densitometry of multiple experiments showed a 4.3 ± 0.8-fold higher expression as in untreated control cultures (Fig. 7B) . In contrast, a much smaller increase was found after treatment with the combination of BMP-7 and TGF-β2 (1.3 ± 0.4-fold; Fig. 7B ). The difference between combined TGF-β2 and BMP-7 treatment, or treatment with TGF-β2 alone was statistically significant (P < 0.01). No increase in mRNA expression was seen after treatment with BMP-7 (0.9 ± 0.2-fold; Fig. 7B ). Western blot analysis of medium obtained from HTM cell cultures visualized a specific band at 50 kDa migrating at the molecular weight of PAI-1. No differences in the intensity of the PAI-1 band were observed between untreated cells and cells treated with BMP-7 or the combination of TGF-β2 and BMP-7, whereas TGF-β2 increased the amounts of PAI-1 expression by 3.4 ± 0.6 (Fig. 7A) . 

The expression of matrix-metalloproteinase 2 (MMP-2), which is commonly and constitutively expressed in the HTM, ⁴³ was investigated by Northern blot analysis. By using a specific probe for MMP-2, a hybridization signal was visualized in RNA from cultured HTM cells, which migrated at the expected size of 3.1 kb (Fig. 7D) . No changes in intensity of this signal were observed after treatment with BMP-7 (Fig. 7D) . In contrast, treatment with TGF-β2 caused a prominent, 3.4 ± 0.7-fold increase of MMP-2 mRNA (Fig. 7D) . The increase was much smaller, when TGF-β2 was added in the presence of BMP-7 (1.5 ± 0.3-fold; Fig. 7D ). The difference between combined TGF-β2 and BMP-7 treatment or treatment with TGF-β2 alone was statistically significant (P < 0.01). In parallel experiments, the activity of MMP-2 was examined by zymography. A prominent band was visualized at 72 kDa, corresponding to the molecular weight of the proform of MMP-2 (Fig. 7C) . In addition, a faint band migrating at 68-kDa was observed indicating the presence of active MMP-2. Only in the culture medium of TGF-β2–treated HTM cells, a considerable increase of proMMP-2 (3.1 ± 0.4-fold) and active MMP-2 was observed (Fig. 7C) . Again, this increase was markedly smaller when TGF-β2 was added in the presence of BMP-7 (1.5 ± 0.4-fold, Fig. 7C ). 

Discussion

We conclude that BMP-7 reduces the TGF-β2-induced accumulation of fibrillar ECM compounds in the human TM by antagonizing the fibrogenic effects of TGF-β signaling. This conclusion is based on (1) the TGF-β2-induced increased expression of a broad panel of molecules that contribute to the deposition of a fibrillar ECM and (2) the observation that the combined treatment of BMP-7 and TGF-β2 largely prevents the TGF-β2-induced expression of fibrillar ECM molecules. 

The fibrogenic effects of TGF-β2 signaling as well as the antagonizing effects of BMP-7 on them were not observed for all the molecules that were analyzed in the present study. Although the expression of fibronectin and of collagens IV and VI was modulated by TGF-β2 or combined TGF-β2/BMP-7 treatment, the expression of collagens I and III was not. Collagens I and III are the major interstitial collagens of the TM, which are predominantly found in the core of corneoscleral and uveal TM lamellae. ⁴¹ In contrast, collagens IV and VI and fibronectin are not typical components of the interstitial ECM, but are localized more closely to cell surfaces, where they are involved in the formation and stabilization of cell-ECM contacts. All three molecules can attach to cells by binding to cell surface integrins. ^{44 45} In addition, collagen VI may bind to fibronectin and/or collagen IV thereby anchoring basement membrane molecules to interstitial collagens. ⁴⁶ Integrins are ECM ligands that serve as transmembrane mechanical links from extracellular contacts to the cytoskeleton inside cells. ⁴⁷ For almost all integrins, the linkage is to the actin-based microfilament system. ⁴⁷ It is of interest to note that TGF-β not only modulates the expression of extracellular integrin ligands, but similarly acts on the actin-based microfilament system in TM cells, as treatment of cultured TM cells with TGF-β causes a substantial increase in stress fibers composed of smooth muscle actin. ³⁴ The physiological role of TGF-β2 and BMP-7 signaling in the TM may be directly related to the modulation of the number and quality of contacts between TM cells and their surrounding ECM. Several agents that increase TM outflow facility in vivo and in perfused anterior segment organ cultures, including the serine-threonine kinase inhibitor H-7, ⁴⁸ the Rho-kinase inhibitor Y27632, ⁴⁹ and the recombinant Hep II domain of fibronectin ⁵⁰ affect the integrity of TM actin stress fibers and cell-ECM contacts. A modulating role of TGF-β2 and BMP-7 signaling on this functional system may serve a critical role for TM biology. 

The antagonizing effects of BMP-7 on TGF-β2 induced gene expression were not restricted to the expression of ECM molecules but were also observed for genes that are involved in the degradation process of those molecules. In a previous study, we identified PAI-1, the inhibitor of the plasminogen/plasmin system, as a TGF-β2-regulated key factor for the enzymatic activation of TM MMPs. ²⁰ The TGF-β2-induced increase in PAI-1 expression, which should lead to a decreased activity of MMPs and should augment the accumulation of TM ECM compounds was not seen after combined treatment with TGF-β2 and BMP-7, a finding that markedly correlated with our data on the effects of TGF-β2 and BMP-7 on ECM expression. In addition, TGF-β2 induced the expression of the proform of MMP2, an effect that was again effectively antagonized by the combined treatment of TGF-β2 and BMP-7. In a previous study, we showed that the presence of TGF-β2-induced PAI-1 largely inhibits the activation of pro-MMP2 to active MMP-2. ²⁰  

To obtain some information on putative elements of the signaling cascade that may be used by BMP-7 to antagonize TGF-β2 signaling, we investigated the expression of CTGF and TSP-1. Both molecules are prominent target genes of TGF-β1 and -2. ^{23 24} In addition, CTGF largely mediates the effects of TGF-βs on ECM turnover, ³⁵ whereas TSP-1 is a major endogenous activator of latent TGF-βs. ³⁶ Our results show an antagonizing effect of the combined TGF-β2/BMP-7 treatment on the TGF-β2-induced expression of both CTGF and TSP-1 and strongly indicate that both molecules contribute to the effects of the combined treatment on TM ECM deposition and expression. The signal transduction pathways inside TM cells that are used by BMP-7 to antagonize TGF-β2 signaling are largely unclear. In general, BMP signals are mediated by type I and II BMP receptors and their downstream molecules Smad1, -5, and -8. Phosphorylated Smad1, -5, and -8 proteins form a complex with Smad4 and are translocated into the nucleus to interact with other transcription factors and to initiate transcription of target genes. ⁵¹ In contrast, activation of TGF-β receptors by their ligands causes the activation and phosphorylation of Smad2 and -3, followed again by heterodimerization with Smad4 and translocation into the nucleus. ⁵² Data obtained in mesangial cells from the kidney, a system in which BMP-7 antagonizes TGF-β signaling as effective as in the TM, indicate that the BMP-7-induced inhibition of the fibrogenic activities of TGF-β requires Smad5. ²⁸  

We hypothesize that BMP-7 plays a comparable role in the living eye, as under the cell culture conditions that were used to obtain the data of the present study. The expression of BMP-7 has been observed in TM samples from the adult human eye, ³² whereas TGF-β2 is found in large amounts in the aqueous humor that passes through the TM. ^{12 53} Both BMP-7 and TGF-β2 may contribute to a functional system that modulates and balances the expression and deposition of a distinct set of TM ECM molecules. This balance may be disturbed in those patients with POAG that show increasing amounts of TGF-β2 in their aqueous humor. It is tempting to speculate that an abnormal low BMP-7 expression should lead to comparable changes in the aqueous humor outflow pathways, as an increase in TGF-β2, but so far, there are no data available on the relative amounts of BMP-7 in TM and aqueous humor of normal patients and those with POAG. In a variety of biological systems, signaling of members of the bone morphogenetic protein family including BMP-7 has been shown to be regulated by proteins, which bind directly to BMPs, and prevent the ligand from interacting with the receptor complex. ⁵¹ The expression of mRNA for some of these BMP-associated molecules that act as BMP antagonists like follistatin, gremlin, and chordin has been observed in RNA from human TM cells. ³² In this context, it would certainly be highly interesting to obtain data on the expression of BMP antagonists in the eyes of patients with POAG. 

An intriguing aspect supported by the data of the present work is that the pharmacologic modulation of TGF-β2 signaling by BMP-7 may be a useful approach to the development of novel therapeutic strategies to prevent or to reverse the structural changes that occur in the TM of eyes with POAG. 

 Presented at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2005.

 Supported by Deutsche Forschungsgemeinschaft Grant TA 115/15-1.

 Submitted for publication March 1, 2006; revised August 8 and September 24, 2006; accepted December 6, 2006.

 Disclosure: R. Fuchshofer, None; A.H.L. Yu, None; U. Welge-Lüssen, None; E.R. Tamm, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Primers Used for PCR Amplification in the Present Study

Target	Sequence	Position	Annealing Temperature (°C)	Size (bp)
Col1α1	5′-GAACGCGTGTCATCCCTTGT-3′	5222	60.0	88
	5′-GAACGAGGTAGTCTTTCAGCAAC-3′	−5310
Col3α1	5′-TGGTCCCCAAGGTGTCAAAG-3′	2562	60.0	117
	5′-GGGGGTCCTGGGTTACCATTA-3′	−2678
Col4α1	5′-CTCGCTGTGGATCGGCTAC-3′	4809	60.0	128
	5′-CGTGACACTCGATGAATGGC-3′	−4936
Col4α2	5′-GAAGTTTGATGTGCCGTGTGG-3′	396	60.0	155
	5′-CTTTACGTCCCTGCAGCCC-3′	−550
Col6α1	5′-GCGACGCACTCAAAAGCAG-3′	426	60.0	68
	5′-AGCGCAGTCGGTGTAGGTG-3′	−493
Col6α2	5′-CACCATCAACCGCATCATCA-3′	760	60.0	57
	5′-TTGTAGCACTCTCCGTAGGCTTC-3	−816
Col6α3	5′-CCACCACTAAGCCCATGGTTA-3′	9203	60.0	147
	5′-GGGACTGATCATGGGCTGAG-3′	−9349
CTGF	5′-CACAAGGGCCTCTTCTGTGA-3′	342	59.6	517
	5′-TCTCTTCCAGGTCAGCTTCG-3′	−858
FN	5′-GAAGCTCTCTCTCAGACAACCA-3′	6514	55.9	669
	5′-AGGTCTGCGGCAGTTGTCAC-3′	−7182
S9	5′-GGGATGTTCACCACCTG-3′	6	60.0	165
	5′-GCAAGATGAAGCTGGATTAC-3′	−231
PAI-1	5′-AGGACCGCAACGTGGTTTTCTC-3′	227–732	58.9	506
	5′-AGTGCTGCCGTCTGATTTGTG-3′
TSP-1	5′-TGCCTGATGACAAGTTCCAA-3′	388	57.4	520
	5′-TCAAGGGTGAGGAGGACAC-3′	−907

Table 2.

 

Antibodies Used for Western Blot and Immunohistochemistry in the Present Study

Antibodies	Abbreviation	Dilution	Companies
Polyclonal goat anti-human CTGF	G anti hCTGF	1:500 (WB) 1:200 (IHC)	Santa Cruz Biotechnology, Santa Cruz, CA
Monoclonal mouse anti-human TSP-1 (Clone A6.1)	M anti hTSP.1	1:250	Biocarta Europe GmbH, Hamburg, Germany
Polyclonal goat anti-human PAI-1	G anti hPAI-1	1:250	Santa Cruz Biotechnology
Alkaline phosphatase-conjugated swine α mouse IgG	Sw anti M-AP	1:5,000	Dianova, Hamburg, Germany
Alkaline phosphatase-conjugated swine α goat IgG	Sw anti G-AP	1:10,000	Dianova
Fluorescein swine anti-rabbit IgG	Sw anti R-Cy3	1:100	Dako, Glostrup, Denmark
Polyclonal rabbit anti-human FN	R anti hFN	1:750	Sigma-Aldrich, Seelze, Germany
Polyclonal rabbit anti-human Col-IVα2	R anti Col4α2	1:250	RDI, Flanders, NJ,
Polyclonal rabbit anti-human Col-I	Anti Col1α1	1:250	Rockland, Gilbertsville, PA,
Polyclonal rabbit anti-human Col-III	Anti Col3α1	1:250	Rockland
Polyclonal rabbit anti-human Col-VI	Anti Col6α1	1:250	Rockland
Alkaline phosphatase-conjugated goat α rabbit IgG	G anti R-AP	1:10,000	Promega, Madison, WI

 WB, Western blot; IHC, Immunohistochemistry.

Figure 1.

 

Western (A, C, D) and Northern blot (B, E) analysis of connective tissue growth factor (CTGF; A, C) and thrombospondin-1 (TSP-1; D), and their respective mRNAs (CTGF mRNA in B, and TSP-1 mRNA in E) in culture medium and RNA from cultured HTM cells after treatment with 300 pM BMP-7, 300 pM TGF-β2, a combination of both (TGF-β2+BMP7), or TGF-β2 and BMP7 in the presence of neutralizing antibodies for BMP-7 (TGF-β2+BMP-7+anti BMP-7). For Western blot analysis, membranes were stained with Coomassie blue to confirm equal loading of proteins. For Northern blot analyses, integrity of RNA and equal loading were controlled by staining ribosomal RNA with methylene blue. Co, control.

Figure 2.

 

Immunoreactivity for CTGF (A), TSP-1 (B), and fibronectin (FN; C) in cultured HTM cells after treatment with 300 pM BMP-7, 300 pM TGF-β2, or a combination of both. Treatment with 300 pM BMP-7 caused no changes in immunoreactivity. In contrast, after treatment with 300 pM TGF-β2, the intensity of staining for CTGF, TSP-1, and FN was considerably enhanced, an effect that was markedly reduced after treatment with a combination of BMP-7 and TGF-β2. (A, inset) Immunoreactivity for CTGF in cellular vesicles (arrowhead). Scale bar: 100 μm.

Figure 3.

 

Real-time PCR (A, C, E, G) and Western blot (B, D, F, H) analysis for CTGF and its mRNA in RNA and culture medium from cultured HTM cells after treatment with 100 or 300 pM TGF-β2 in combination with 50, 100, 150, or 300 pM BMP-7 for (A–D) 72 or (E–H) 24 hours. For real-time PCR analysis, means ± SD of four independent experiments run in duplicate are shown. The mean value obtained with RNA from untreated cells was set at 100%. *Statistically significant differences between combined TGF-β2/BMP-7 treatment and treatment with TGF-β2 alone (P < 0.01). For Western blot analysis, membranes were stained with Coomassie blue to confirm equal loading of proteins. Co, control.

Figure 4.

 

Western blot (A, C), Northern blot (B), and real-time PCR analysis (D) of fibronectin (FN; A) and collagen type IV (Col IV; C) and their respective mRNAs (FN mRNA in B, and COL4A1 and COL4A2 mRNA in D) in cell culture medium and RNA from cultured HTM cells after treatment with 300 pM BMP-7, 300 pM TGF-β2, or a combination of both. For Western blot analysis, membranes were stained with Coomassie blue to confirm equal loading of proteins. Integrity of RNA and equal loading in (C) were controlled by staining ribosomal RNA with methylene blue. For real-time PCR analysis, means ± SD of four independent experiments run in duplicate are shown. The mean value obtained with RNA from untreated cells was set at 100%. *Statistically significant differences between combined TGF-β2/BMP-7 treatment and treatment with TGF-β2 alone (P < 0.01). Co, control.

Figure 5.

 

Western blot for collagen type VI (A) and real-time PCR analysis for mRNA of the three collagen type VI subunits, COL6A1, COL6A2, and COL6A3 (B), in culture medium and RNA from cultured HTM cells after treatment with 300 pM BMP-7, 300 pM TGF-β2, or a combination of both. For Western blot analysis, membranes were stained with Coomassie blue to confirm equal loading of proteins. For real-time PCR analysis, means ± SD of four independent experiments run in duplicate are shown. The mean value obtained with RNA from untreated cells was set at 100%. *Statistically significant differences between combined TGF-β2/BMP-7 treatment and treatment with TGF-β2 alone (*P < 0.01; **P < 0.05). Co, control.

Figure 6.

 

Western blot for collagen types I and III (Col I and Col III, A) and real-time PCR analysis for mRNA of their collagen subunits, COL1A1 and COL3A2 (B) in culture medium and RNA from cultured HTM cells after treatment with 300 pM BMP-7, 300 pM TGF-β2, or a combination of both. For Western blot analysis, membranes were stained with Coomassie blue to confirm equal loading of proteins. For real-time PCR analysis, means ± SD of four independent experiments run in duplicate are shown. The mean value obtained with RNA from untreated cells was set at 100%. Co, control.

Figure 7.

 

Western blot analysis for PAI-1 (A), zymography for activity of MMP-2 (C), and Northern blot analysis for PAI-1 (B) or MMP-2 (D) mRNA in culture medium and RNA from cultured HTM cells after treatment with 300 pM BMP-7, 300 pM TGF-β2, or a combination of both. For Northern blot analyses, integrity of RNA and equal loading were controlled by staining of ribosomal RNA with methylene blue. Co, Control.

The authors thank Angelika Pach for excellent technical help.