Transforming Growth Factor-β2 Modulated Extracellular Matrix Component Expression in Cultured Human Optic Nerve Head Astrocytes

Abstract

purpose. To study whether glaucomatous extracellular matrix (ECM) modifications in the lamina cribrosa might be induced by TGF-β2, the effect of TGF-β2 on the expression of collagen types I (Col1α1), III (Col3α1), and IV (Col4α2); fibronectin (FN); tissue transglutaminase (TGM2); connective tissue growth factor (CTGF); and thrombospondin (TSP-1) in cultured human optic nerve head (ONH) astrocytes was investigated.

methods. Astrocytes were isolated from eyes of five human donors, and cultured monolayers were treated with 1.0 ng/mL TGF-β2 for 24 and 48 hours. Expression of Col1α1, Col3α1, Col4α2, FN, TGM2, CTGF, and TSP-1 was examined by semiquantitative RT-PCR and Northern and Western blot analyses. The effect of CTGF silencing on the TGF-β2–modulated expression of these genes was investigated by transfection of CTGF small interfering (si)RNA before TGF-β2 treatment.

results. TGF-β2 treatment upregulated the expression of Col1α1, Col4α2, FN, CTGF, TGM2, and TSP-1 mRNA and protein in cultured astrocytes. Inductions ranged between 1.5- and 4-fold. Expression of Col3α1 remained unaffected. Transfection of 10 nM CTGF siRNA inhibited the TGF-β2–induced upregulation of CTGF, Col4α2, Col1α1, TGM2, and FN, whereas TSP-1 expression was not reduced.

conclusions. TGF-β2 is capable of inducing the expression of ECM and basement membrane components in cultured ONH astrocytes via CTGF and upregulated TSP-1, a protein naturally involved in the activation of latent TGF-β. Therefore, TGF-β2 could be a factor in the initiation of the modification of ECM in the glaucomatous ONH. In addition, TSP-1 induction may be a mechanism by which TGF-β2 amplifies its own activation.

In most patients with primary open-angle glaucoma (POAG), the intraocular pressure (IOP) in the eyes is increased, and the optic nerve head (ONH) shows characteristic cupping correlated with visual field defects. This increase in IOP is due to enhanced resistance in the aqueous humor (AH) outflow pathways. The underlying molecular mechanisms and the responsible pathogenic factors leading to these changes in outflow resistance remain unknown. There are explicit indications, however, that transforming growth factor (TGF)-β2 may be involved in the pathogenesis of POAG. Statistical data have shown that the amount of this factor is significantly increased in the AH of approximately 50% of patients with POAG. ^{1 2} In tissue culture experiments in trabecular meshwork (TM) cells, TGF-β2 augmented the synthesis of extracellular matrix (ECM) and its cross-linking by transglutaminases. ³ Moreover, perfusion of human anterior eye segments with TGF-β2 in concentrations comparable to those measured in the AH of patients with POAG resulted in an increase of IOP and an increase of ECM in the outflow tissue. ⁴ In a previous study, our group has shown that there is a significant correlation between the amount of ECM underneath the inner wall of Schlemm’s canal (SC) and the degree of axon loss in the ONH of glaucomatous eyes. In contrast, a correlation between ECM in the TM and increased IOP was not found, indicating that the elevation of IOP is not due to increased sheath-derived plaques alone. ⁵ Because the ECM in the TM cannot directly influence the optic nerve and ECM modifications are also visible in the ONH region of glaucomatous eyes, one might assume that both glaucomatous changes are caused by common, transacting factors. Based on our previous findings, we investigated in the present study whether the same TGF-β2–triggered mechanism described in the TM might function also in the ONH. Support for a possible involvement of TGF-β2 in ONH modification was provided by the finding that the TGF-β2 levels in the ONH of glaucomatous eyes are actually elevated and that astrocytes of the ONH, the major glial cell population in that region that participate in ECM synthesis, predominantly express this isoform of the TGF-β-family. ⁶  

In this study we analyzed the effect of TGF-β2 on cultured ONH astrocytes with respect to ECM synthesis; on the expression of connective tissue growth factor (CTGF), an upstream regulator of ECM synthesis; and on thrombospondin (TSP)-1, an activator of TGF-β2. 

Material and Methods

Cell Culture

Explant cultures of human lamina cribrosa astrocytes were obtained from our collaborators at the Department of Ophthalmology, Maximilians-University, Munich, Germany. Monolayer cultures were established from eyes of five human donors (52, 54, 67, 69, and 82 years old, obtained 4 to 8 hours after death) without any history of eye disease. The eyes were prepared and tissue cultures grown as previously described by Kobayashi et al. ⁷ In detail, eyes were cut equatorially behind the ora serrata and the ONH was isolated from the neighboring tissues. The ONH was sagittally dissected under a microscope, and the lamina cribrosa was identified. Discs of lamina cribrosa were prepared by dissection from the pre- and postlaminar regions and were subsequently chopped and placed in Petri dishes with 2 mL Dulbecco’s modified Eagle’s medium (DMEM)/F-12 supplemented with 10% fetal bovine serum (FBS; Invitrogen-Gibco-Life Science Technology, Karlsruhe, Germany), 5 ng/mL human basic pituitary fibroblast growth factor (bFGF; Sigma-Aldrich, Darmstadt, Germany), 5 ng/mL human platelet-derived growth factor-A chain (PDGFAA; Sigma-Aldrich), and 50 U/mL penicillin and 50 μg/mL streptomycin (Invitrogen-Gibco-Life Science Technology). The medium was changed every second day, and cells were passaged in a ratio of 1:2, using 0.1% trypsin and 0.02% EDTA in phosphate-buffered 0.15 M NaCl (pH 7.2; Invitrogen-Gibco-Life Science Technology). Cells were shipped from Munich to Erlangen at passage 2 and treated the same way for one to three additional passages. In accordance with Hernandez et al., ⁸ we classified astrocytes by immunohistochemical staining with an antibody against glial fibrillary acidic protein (GFAP, Table 1 ), which is expressed by ONH astrocytes in contrast to microglia and other cell types of the lamina cribrosa region. Only GFAP-positive cultures were used for further experiments (Fig. 1) . Methods of securing human tissue were humane, included proper consent and approval, and complied with the declaration of Helsinki. 

To investigate the effects of TGF-β2, 5 × 10⁵ astrocytes of passages 3 to 5 were plated in 35-mm Petri dishes and grown to confluence in DMEM/F-12 supplemented with 10% FBS at 37°C in 5% CO₂. At confluence, the medium was changed to serum-free DMEM/F-12, and, after 24 hours of incubation, the medium was replaced by fresh serum-free DMEM/F-12 supplemented with active TGF-β2 (Roche, Basel, Switzerland), to a final concentration of 1.0 ng/mL. This concentration was chosen as it resembles the mean average of TGF-β2 levels measured in the aqueous humor of patients with POAG. Under these conditions, cells were incubated 24 and 48 hours, respectively. In control cultures the medium was changed at the same time points but no TGF-β2 was added. Cell viability was tested before and at the end of treatment and did not reveal any signs of increased cell death in TGF-β2–treated cells. 

RNA Isolation and Polymerase Chain Reaction

Astrocytes were harvested from 35-mm Petri dishes at the indicated time points, and total RNA was extracted with an RNA isolation kit (Stratagene, Heidelberg, Germany). Structural integrity of the RNA samples was confirmed by electrophoresis in 1% Tris-acetate-EDTA (TAE)-agarose gels. Yield and purity were determined photometrically. For preparation of gene-specific antisense Northern blot probes, first-strand cDNA for PCR was prepared from total RNA by using Moloney murine leukemia virus reverse transcriptase (M-MLV-RT) and oligo(dT)-17 primer. PCR amplification of gene-specific probes was performed with a temperature profile as follows: 36 cycles of a 1-minute denaturation at 94°C, a 1-minute annealing period (for specific temperatures, see Table 2 ), and a 2-minute extension at 72°C, followed by an extension step of 10 minutes at 72°C after the last cycle. All PCR primers were purchased from Invitrogen (Darmstadt, Germany) and spanned exon–intron boundaries. Primer sequences, positions, annealing temperatures, and PCR product sizes are shown in Table 2 . Sizes of the PCR products were estimated from the migration of a DNA size marker run concurrently (100-bp DNA Ladder; Promega, Madison, WI) on a 1% TAE-agarose gel. No-reverse-transcriptase RNA served as the negative control for PCR and showed no amplification products (data not shown). PCR products were purified with a kit (Qiagen, Hilden, Germany), cloned into the vector pTOPO/TA (Invitrogen), and sent to Seqlab (Regensburg, Germany) for sequence analysis. 

Semiquantitative RT-PCR

The cDNA of treated astrocytes was produced as just described. Ratios of the different cDNAs were initially determined by GAPDH PCR. The linear range of the gene-specific PCRs was tested before the experiments. PCRs were performed with a temperature profile as follows: 30 seconds at 94°C, 30 seconds of annealing, and 45 seconds at 72°C. Specific annealing temperatures and cycle numbers are shown in Table 2 . PCR products were size fractionated in 1% TAE-agarose gels and ethidium bromide stained, and signal intensities were quantified on computer (BioDocAnalyze software; Whatman Biometra Biomedizinische Analytik GmbH, Göttingen, Germany). 

Northern Blot Analysis

A portion (15 μg) of total RNA was size-fractionated by gel electrophoresis in 1% agarose gels containing 2.2 M formaldehyde, subsequently transferred onto a nylon membrane (Roche) by vacuum blot and cross-linked at 1600 μJ (Stratalinker; Stratagene, La Jolla, CA). To assess the amount and quality of the RNA, membranes were stained with methylene blue, and images were taken (Lumi-Imager; Roche, Mannheim, Germany). Prehybridization was performed at 68°C for 1 hour (Dig Easy Hyb; Roche). Hybridizations were performed at 68°C overnight in prehybridization solution containing 50 ng/mL of a specific antisense probe. Riboprobe synthesis has been described previously. ⁹ After hybridization, the membranes were washed twice with 2× SSC/0.1% SDS at room temperature, followed by two washes in 0.1% SDS at room temperature or for 15 minutes at 68°C. Afterward, the membranes were washed for 5 minutes in washing buffer (100 mM maleic acid, 150 mM NaCl, [pH 7.5] and 0.3% Tween-20) and incubated for 1 hour in blocking solution (100 mM maleic acid, 150 mM NaCl [pH 7.5], and 1% blocking reagent; 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 and 100 mM NaCl [pH 9.5]) for 10 minutes. For chemiluminescence detection, the detection reagent (CDP-star; Roche) was diluted 1:100 in detection buffer, and the membranes were incubated for 5 minutes at room temperature. After they were air dried, the semidry membranes were sealed in plastic bags, and chemiluminescence was detected (Lumi-Imager workstation; Roche). Exposure times ranged between 2 minutes and 1 hour. Quantification was performed with the accompanying software (Lumi-Analyst; Roche). 

Western Blot Analysis

Media of the treated cells were collected, and aliquots (500 μL) were concentrated 50-fold by centrifugation through a membrane (10-kDa cutoff; Amicon; Millipore, Bedford, MA), 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% desoxycholic acid, 0.1% SDS, and 50 mM Tris [pH 8]) and protein content was measured with the bicinchoninic acid (BCA) protein assay (Pierce, Rockford, IL). All probes were supplemented with SDS loading buffer and denatured by boiling for 5 minutes, and 2 μg of each sample was separated by SDS-polyacrylamide gel electrophoresis (PAGE) and transferred onto a polyvinylidene difluoride membrane (PVDF; Roche), either by semidry blotting (CTGF, TGM2, TSP-1) or by tank blot (Col1α1, Col3α1, Col4α2, FN) at 70 V for 45 minutes in 1× transfer buffer (10 mM cyclohexylaminopropan sulfonic acid [pH 11], and 20% methanol, 0.1% SDS). Membranes were blocked in PBST/BSA (phosphate-buffered saline, 0.1% Tween-20, and 5% bovine serum albumin [pH 7.2]) for 1 hour. Primary antibodies were added in PBST (for dilutions, see Table 1 ) and allowed to react overnight at 4°C. After the membranes were washed with PBST, secondary antibodies were added in PBST at the appropriate dilution (see Table 1 ) for 30 minutes at room temperature. For detection, the chemiluminescent reagent (CDP-Star; Roche) was diluted 1:100 in detection buffer, the membranes were incubated for 5 minutes at room temperature, and chemiluminescence signals were analyzed and quantified (Lumi-Imager workstation, running Lumi-Analyst software; Roche). Exposure times ranged between 1 and 5 minutes. 

Generation and Transfection of siRNA

The target sequences for the human CTGF small interfering (si)RNA were designed with Web-based criteria and generated with an siRNA construction kit (Silencer; Ambion, Austin, TX). Different CTGF siRNAs were tested in initial transfection and subsequent Northern blot experiments (data not shown). Best results were obtained by transfecting 10 nM of the CTGF-859 siRNA (named after the nucleotide start site in the CTGF sequence), with the transfection reagent, according to the manufacturer’s instructions (Invitrogen). The primers used to generate this CTGF siRNA were CTGF-859 5′-AATGTTCT-CTTCCAGGTCAG-CCCTGTCTC-3′ (sense) and 5′-AAGCTGACCTGGAAGAGAACAC-CTGTCTC-3′ (antisense). Maximum silencing was reached 3 hours after transfection at the latest, and the effect lasted up to 72 hours after transfection, at least (data not shown). To assess the effect on the TGF-β2–mediated changes in expression of CTGF, FN, Col1α1, Col4α2, TGM2, and TSP-1, cells were seeded as previously described, transfected with 10 nM CTGF siRNA, and supplemented with medium containing TGF-β2 to a final concentration of 1.0 ng/mL after 4 hours. Cells were incubated in this manner for 48 hours before they were harvested for RNA isolation. 

Immunohistochemistry

Cultured astrocytes were grown on microscope slides and treated with 1 ng/mL TGF-β2, as just described. Control populations were prepared the same way, but were not exposed to TGF-β2. After incubation, cells were fixed with 4% paraformaldehyde (PFA) for 15 minutes and subsequently washed twice with PBS containing 0.1% Triton X-100. Primary antibodies were added in appropriate dilutions in PBS and 5% BSA (Table 1)and allowed to bind for 4 hours at room temperature. After three wash steps with PBS, fluorescein-conjugated secondary antibodies were added (Table 1)for 1 hour at room temperature. A marker of nuclei, 4′,6-diamidino-2-phenylindole (DAPI), was used to counterstain DNA. After immunohistochemical labeling, cells were mounted on slides with mounting medium containing DAPI (Vectashield; Vector Laboratories, Burlingame, CA). Slides were analyzed under a fluorescence microscope (Aristoplan; Leitz, Wetzlar, Germany). Corresponding negative control cultures to estimate unspecific binding of secondary antibodies were handled similarly, but were incubated in PBS/BSA without primary antibody. 

Results

Characterization of ONH Astrocytes

Astrocyte explant cultures of passages 3 to 5 showed an intense GFAP staining throughout the cytoplasm (Fig. 1B) . Corresponding negative controls, to assure specificity of the used antibodies, showed no staining (Fig. 1A) . 

FN and Col4α2

Treatment of ONH astrocytes with 1.0 ng/mL TGF-β2 caused an enhanced signal intensity of the FN-specific 7.7-kb RNA band in Northern blot analysis compared with untreated cells (Fig. 2A , left). Quantification by comparison of the signal intensities revealed a 2.9 ± 1.1 (SD)-fold induction after 24 hours and a 3.5 ± 1.5-fold induction after 48 hours (Table 3) . Hybridization with an antisense Col4α2 RNA probe showed a 2.0 ± 0.2-fold upregulation of the Col4α2-specific 6.3-kb band after a 24-hour treatment and a 2.9 ± 1.1-fold increase after 48 hours, compared with untreated astrocytes (Fig. 2A , right, Table 3 ). Signals of the 28S and 18S rRNAs served as controls for equal loading and were considered for quantification (Fig. 2B) . The same upregulation of the FN and Col4α2 proteins was also detected in Western blot analysis (Fig. 2C) . Because both proteins are secreted, we analyzed concentrated aliquots of the media in which the astrocytes were incubated. In the medium of untreated astrocytes, only the polymerized FN complex of 240 kDa was detectable, whereas after 48 hours of TGF-β2 treatment, faster migrating protein bands, presumably resembling FN complexes of lower aggregation, were detectable, and the 240-kDa band showed a stronger signal intensity (Fig. 2C , left). Quantification gave a 2.6 ± 0.9-fold upregulation of the assembled 240-kDa FN complex after 48 hours of TGF-β2 treatment compared with untreated astrocytes (Table 3) . In the medium of TGF-β2–treated astrocytes, the 230-kDa band corresponding to Col4α2 showed greater intensity in Western blot analysis than did the signal in the medium of untreated astrocytes (Fig. 2C , right). Quantitative analysis revealed a 3.6 ± 0.9-fold upregulation after 48 hours of treatment with TGF-β2 (Table 3) . 

Col1α1 and Col3α1

Semiquantitative PCR (sqPCR) analysis of cDNAs obtained from astrocytes treated with TGF-β2 for 48 hours and from untreated control astrocytes showed an increase in the Col1α1 mRNA level by a factor of 2.7 ± 0.6 (SD) in response to treatment, whereas the Col3α1 level did not increase (1.3 ± 0.4; Figs. 3A 3B , Table 3 ). Application of equal cDNA amounts in the PCR was controlled by GAPDH PCR (Figs. 3A 3B) . Experiments were conducted on cDNAs obtained from three independent astrocyte cultures. Figure 3Ashows a representative single experiment. To confirm this regulation on the protein level as well, we conducted Western blot analyses analogous to those we performed for FN and Col4α2. A 140-kDa band corresponding to Col1α1 was significantly amplified in the medium of TGF-β2–treated astrocytes (Fig. 3C , left) whereas the Col3α1-specific band of 139 kDa was not altered (Fig. 3C , right). Quantitative analysis determined a 3.6 ± 0.6-fold upregulation of Col1α1, whereas the value for Col3α1 was 1.2 ± 0.3-fold (Table 3) . 

TGM2 and CTGF

TGM2 is involved in cross-linking of ECM components and is TGF-β2 inducible in human TM cells. ³ TGF-β2 treatment of ONH astrocytes resulted in a 2.3 ± 0.4-fold upregulation of the TGM2-specific, 3.2-kb RNA band after 24 hours that slightly decreased to 2.1 ± 0.3-fold after 48 hours (Fig. 4A , left; Table 3 ). CTGF is a major regulator of ECM component expression and is a very likely direct target gene of TGF-β2. Treatment with 1.0 ng/mL TGF-β2 significantly increased CTGF mRNA levels in astrocytes in Northern blot analysis (Fig. 4A , right). By comparison to untreated astrocytes, an induction of 3.5 ± 0.6-fold of the 2.1-kb CTGF mRNA was reached after 24 hours of exposure to TGF-β2. Prolonged treatment for 48 hours did not result in a stronger induction of CTGF (Table 3) . 28S and 18S rRNA band intensities served as loading controls and were included for quantification (Fig. 4B) . 

Western blot analysis also confirmed elevated amounts of TGM2 and CTGF protein in response to TGF-β2 treatment on the protein level. The intensity of the TGM2 signal at 77 kDa was increased 2.2 ± 0.5-fold after 48 hours (Fig. 4C , left; Table 3 ). The CTGF corresponding band of 36 kDa was detectable at all time points analyzed, but intensity increased uniformly with the length of treatment (Fig. 4C , right). After 24 hours of exposure to TGF-β2, the CTGF protein level was augmented by a factor of 2.6 ± 0.5 compared with the protein amount of untreated astrocytes and increased by a factor of 3.5 ± 0.2 after prolonged treatment (48 hours; Table 3 ). To allow quantification, protein contents were determined biochemically before electrophoresis, and equal protein amounts were loaded on the gels. 

The presence of the CTGF protein in the cultures was also demonstrated by immunohistochemical staining with an anti-CTGF antibody (Fig. 5A) . Treatment with 1.0 ng/mL TGF-β2 enhanced the staining intensity (Fig. 5B) . Negative control cultures, which were not incubated with the primary antibody, showed no staining (data not shown). 

CTGF Silencing

To test the possibility of direct blockage of TGF-β2–induced CTGF, FN, Col1α1, Col4α2, and TGM2 upregulation, we conducted a series of silencing experiments with a CTGF siRNA. TGF-β2–treated astrocyte cultures showed the same upregulation of CTGF, FN, Col4α2, Col1α1, and TGM2 mRNA after 48 hours of exposure as did untreated cells, as described earlier (Figs. 6A 7Aleft, 7Ctop, lanes 1 and 2; Table 3 ). In contrast, CTGF upregulation was almost abolished when astrocytes were transfected with 10 nM CTGF siRNA before exposure to TGF-β2 (Fig. 6Aleft, lane 3). The expression analysis of TGM2, Col1α1, Col4α2, and FN gave a similar result, but with a lesser reduction (lane 3 in Figs. 6Amiddle and right, 7A left, 7C top). Quantification of the data revealed reductions of the TGF-β2–induced expression of 91% for CTGF, 82% for TGM2, 76% for Col1α1, 84% for Col4α2, and 68% for FN, on the mRNA level (Table 3) . The staining intensities of the 28S and 18S rRNA served as controls for equal loading and were included in the quantitative analysis (Figs. 6B 7C , bottom). For the data of Col1α1, GAPDH expression was considered for quantification. The reduction in TGF-β2–induced expression was also demonstrated on the protein level by Western blot analysis (Figs. 6C 7D) . The CTGF protein level in siRNA-transfected, TGF-β2–treated astrocytes was reduced to approximately 12% of the amount detected in untransfected, TGF-β2–treated astrocytes (Fig. 6C , left, lane 3; Table 3 ), whereas the level of FN was reduced to approximately 19%, and Col4α2 was reduced to approximately 5% (Figs. 6C , middle and right, lane 3; Table 3 ). For the polymerized FN complex of 240 kDa, the reduction seemed even stronger, but for quantification, all detectable bands were included. Quantification of TGM2 and Col1α1 expression in siCTGF transfected cells showed reductions to 17% and 8%, respectively (Fig. 7D , lane 3, Table 3 ). 

Thrombospondin-1

To investigate whether TGF-β2 treatment of astrocytes has an effect on TGF-β2–activation pathways, we examined the changes of TSP-1 expression after exposure to TGF-β2. 

In treated ONH astrocytes, increased levels of TSP-1 mRNA were detected in Northern blot analysis compared with that in untreated astrocytes (Fig. 8A , top). A 24-hour treatment lead to a 4.2 ± 2.4-fold upregulation of the TSP-1–specific 3.2-kb band, and a 5.2 ± 2.3-fold induction was measurable after 48 hours (Table 3) . Equal loading was assured by analysis of the 28S and 18S rRNA signals, which were also included in the quantification (Fig. 8A , bottom). Such an increased expression of TSP-1 after TGF-β2 treatment was also detectable on the protein level in the cytosolic fraction and in the medium, as demonstrated by Western blot analysis (Fig. 8B) . The TSP-1–corresponding band of 180 kDa was upregulated by a factor of 3.5 ± 0.4 after 24 hours of treatment and by a factor of 3.3 ± 0.6 after 48 hours in the cytosolic fraction (Table 3) . Silencing of CTGF by transfection of CTGF siRNA did not affect the TGF-β2–induced upregulation of TSP-1, as shown by Northern blot analysis (Fig. 8C , top). Equal loading was confirmed by staining of the 28S and 18S RNA (Fig. 8C , bottom). 

Discussion

In eyes with POAG, Hernandez et al. ¹⁰ described differences in ECM components of the lamina cribrosa and optic nerve head region compared with those of age-matched control eyes. Among these differences, an increase in density and area occupied by basement membrane (BM) material was the prominent finding. In addition, increased collagen type IV mRNA expression was found in human glaucomatous ONH. ¹¹ From an ongoing morphologic study, we have clear indications that the amount of thick, cross-banded collagen fibers is increased in the ONH of POAG eyes (Gottanka J, unpublished data, 2004). The technical setup of that study did not allow precise classification of the collagen species, but from immunohistochemical studies, it is known that collagen types I and III are present in the septae of the optic nerve. ¹² In early and moderate forms of glaucoma, an increase in astrocytes has been described in the prelaminar and laminar regions of the optic nerve ¹³ ; therefore, Minckler and Spaeth assumed that the proliferating glial cells might be the source of the newly synthesized BM and ECM material. ¹⁴ Moreover, it is known that cultured astrocytes of the central nervous system (CNS) express collagen types I and III. ¹⁵  

In this study, TGF-β2 treatment of cultured human ONH astrocytes leads to elevated expression of Col4α1 and Col1α1. Quantification of Northern blot data showed that the Col4α2 expression was upregulated by a factor of 3 in TGF-β2–treated astrocytes compared with untreated cells, and the same upregulation was found for Col1α1. These findings were confirmed on the protein level by Western blot analysis of both collagens. From that, we conclude that astrocytes of the lamina cribrosa could be the cell population responsible for the augmented production of BM material in the ONH of glaucomatous eyes. Moreover, our data indicate that TGF-β2 could be one of the initiating factors of these glaucomatous changes by induction of collagen expression in astrocytes, but our data also show that not all collagens are induced by TGF-β2, as the expression of Col3α1 was not modulated. 

There are reports that collagen assembly and correct organization and deposition as fibers in BM is dependent on another ECM component, FN, which contains a collagen-binding site. ^{16 17 18 19} Expression of FN has not yet been investigated in the lamina cribrosa, but there are correlations between recruitment of astrocytes and local FN expression in wound healing of brain lesions. ²⁰ Studies of the human TM by our group showed that increased FN expression is a hallmark of TGF-β2–mediated ECM modification in POAG. ³ The data we present herein show that astrocytes of the ONH also induced FN expression in response to TGF-β2 treatment. Together with our finding that TGM2 was also upregulated by TGF-β2, it becomes obvious that TGF-β2 induces not only prominent BM components but also architectural enzymes that contribute to the assembly of a stable BM. TGM2 is essential for the structural integrity of BMs by cross-linking of proteins, such as fibronectin, ²¹ vitronectin, ²² laminin–nidogen ^{23 24} complexes, or collagen type III, ²⁵ and disregulation of TGM2 has been described in numerous diseases such as pulmonary fibroses ²⁶ and arteriosclerosis. ^{27 28 29}  

In the current work, CTGF was significantly upregulated on the mRNA and protein levels by TGF-β2. The promoter region of the CTGF gene contains a TGF-β response element and TGF-β–dependent upregulation of CTGF expression has been described in human skin fibroblasts. ³⁰ Hints of a putative connection between TGF-β and CTGF expression in astrocytes resulted from the analysis of neurologic diseases. In patients with amyotrophic lateral sclerosis of the spinal cord, different groups found elevated levels of CTGF ³¹ as well as of TGF-β. ³² The same augmented expression of both growth factors was described in reactive astrocytes in human cerebral infarction. ^{33 34} Our data show that the same TGF-β2–dependent upregulation applied also to human ONH astrocytes. CTGF itself is a positive regulator of FN expression in rat kidney fibroblasts, ³⁵ and previous studies have shown that TGF-β induces FN synthesis in cerebral astrocytes. ^{36 37} In the current study, FN mRNA induction was abolished when astrocytes were transfected with a CTGF-specific siRNA before TGF-β2 treatment, indicating that FN is rather an indirect target of TGF-β2, regulated via CTGF, also in human ONH astrocytes. The same regulation pathway seems to hold true for Col1α1, Col4α2, and TGM2, as CTGF silencing also repressed TGF-β2–triggered upregulation of the genes. These findings indicate that CTGF may be one of the key targets of TGF-β2. 

A question that remains open is how TGF-β2 becomes activated in the ONH. Active TGF-β2 dimers are derived from dimeric 55-kDa precursor polypeptides that are proteolytically processed to yield the mature growth factor and N-terminal propeptides. These propeptides are linked by disulfide bonds to form the latency-associated peptide (LAP), which binds noncovalently to the mature growth factor to retain latent TGF-β2. This small latent TGF-β (SL-TGF-β) complex is intracellularly associated with the latent TGF-β–binding protein (LTBP) to form the large latent TGF-β (LL-TGF-β), which is secreted from the cells and recruited via LTBP to the ECM. TSP-1 is a known natural trigger of TGF-β2 activation by binding to LAP and inducing conformational changes that lead to TGF-β2 release. ³⁸ Our data show that TGF-β2 treatment of ONH astrocytes induces the expression of TSP-1 on both the mRNA and the protein levels. This implies a putative positive feedback mechanism, meaning that once active TGF-β2 is released in the ONH, it can amplify its own activation in an autocrine manner by upregulation of TSP-1. 

Taken together, our data support the idea that TGF-β2 becomes a major proglaucomatous factor by misregulating ECM synthesis and cross-linking of ECM and BM components in the ONH, as well as in the TM. ^{3 4 39}  

 Supported by Grant SFB 539 (Glaukome) of the Deutsche Forschungsgemeinschaft.

 Submitted for publication June 4, 2004; revised October 22, 2004; accepted November 1, 2004.

 Disclosure: R. Fuchshofer, None; M. Birke, None; U. Welge-Lussen, None; D. Kook, None; E. Lütjen-Drecoll, 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.

 

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)	Santa Cruz Biotechnology, Santa Cruz, CA
		1:200 (IHC)
Monoclonal mouse anti-human TSP-1 (clone A6.1)	M anti-hTSP.1	1:250	Biocarta Europe GmbH, Hamburg, Germany
Polyclonal rabbit anti-human GFAP	R anti-hGFAP	1:100	Sigma Aldrich, Darmstadt, Germany
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
Monoclonal mouse anti-human TGM2 (Cub7402)	anti-TGM2	1:100	Quartett, Berlin, Germany
Polyclonal rabbit anti-human Col-I	anti-Col1α1	1:250	Rockland, Gilbertsville, PA
Polyclonal rabbit anti-human Col-III	anti-Col3α1	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.

 

Immunofluorescence staining for GFAP. (A) Negative control, treated with secondary antibody only, showed no staining. Nuclei are DAPI stained (blue). (B) Monolayers of explanted cells showed intense staining for GFAP. Staining was seen throughout the cytoplasm (red), whereas the nuclei remain unstained. Magnification, ×400.

Table 2.

 

Primers Used for PCR Amplification in the Present Study

Target	Sequence	Position	Annealing Temperature (°C)	Size (bp)	Cycles (sqPCR)
Col1α1	5′-GATGGACTCAACGGTCTCC-3′	3576	54.0	458	35
	5′-CCTTGGGGTTCTTGCTGATG-3′	−4034
Col3α1	5′-GACCTGAAATTCTGCCATCC-3′	2996	54.0	461	35
	5′-CAACCATCCTCCAGAACTGTG-3′	−3456
Col4α2	5′-AACCAGGTTTTCGTGGGGCT-3′	4568	61.0	403	ND
	5′-TTCCGGCTGGCATAGTAGCA-3′	−4970
CTGF	5′-CACAAGGGCCTCTTCTGTGA-3′	342	59.6	517	ND
	5′-TCTCTTCCAGGTCAGCTTCG-3′	−858
FN	5′-GAAGCTCTCTCTCAGACAACCA-3′	6514	55.9	669	ND
	5′-AGGTCTGCGGCAGTTGTCAC-3′	−7182
GAPDH	5′-GAAGGTGAAGGTCGGAGTC-3′	6	60.0	225	22
	5′-GAAGATGGTGATGGGATTTC-3′	−231
TGM2	5′-GGTCAACTGCAACGATGACC-3′	781	53.0	618	ND
	5′-TCGGCCCACGCTCTTAGTGC-3′	−1399
TSP-1	5′-TGCCTGATGACAAGTTCCAA-3′	388	57.4	520	ND
	5′-TCAAGGGTGAGGAGGACAC-3′	−907

Figure 2.

 

TGF-β2 induced upregulation of FN and Col4α2. (A) Northern blot analysis of FN (left) and Col4α2 (right) mRNA levels in TGF-β2–treated astrocytes (lanes 2 and 4) and untreated controls (Co, lanes 1 and 3). (B) Methylene blue–stained corresponding 28S and 18S RNA bands. (C) Western blot analysis of FN (left) and Col4α2 (right) expression.

Table 3.

 

Quantitative Measurements for Expression of CTGF, FN, Col1α1, Col3α1, Col4α2, TGM2 and TSP-1 after TGF-β2 Treatment

Gene	mRNA Protein	Control	24 Hours TGF-β2	48 Hours TGF-β2	48 Hours TGF-β2 siRNA CTGF
Col1α1	mRNA	1.0	ND	2.7 ± 0.6	1.4 ± 0.4
	Protein	1.0	ND	3.6 ± 0.6	1.2 ± 0.2
Col3α1	mRNA	1.0	ND	1.3 ± 0.4	ND
	Protein	1.0	ND	1.2 ± 0.3	ND
Col4α2	mRNA	1.0	2.0 ± 0.2	2.9 ± 1.1	1.3 ± 0.5
	Protein	1.0	ND	3.6 ± 0.9	1.1 ± 0.3
CTGF	mRNA	1.0	3.5 ± 0.6	3.3 ± 0.6	1.2 ± 0.7
	Protein	1.0	2.6 ± 0.5	3.5 ± 0.2	1.3 ± 0.2
FN	mRNA	1.0	2.9 ± 1.1	3.5 ± 1.5	1.8 ± 0.4
	Protein	1.0	ND	2.6 ± 0.9	1.3 ± 0.4
TGM2	mRNA	1.0	2.3 ± 0.4	2.1 ± 0.3	1.2 ± 0.3
	Protein	1.0	ND	2.2 ± 0.5	1.2 ± 0.2
TSP-1	mRNA	1.0	4.2 ± 2.4	5.2 ± 2.3	3.8 ± 0.9
	Protein	1.0	3.5 ± 0.4	3.3 ± 0.6	ND

Figure 3.

 

Col1α1 and Col3α1 expression in TGF-β2–treated ONH astrocytes. (A) Ethidium bromide stained 1%-TAE agarose gel of (left) Col1α1-, (middle) Col3α1-, and (right) GAPDH-specific PCR products on cDNA obtained from 48-hour TGF-β2–treated astrocytes (lanes 2) and untreated controls (lanes 1). (B) Statistical analysis of the band intensities in (A). Mean ± SD; n = 3. Control values = 1. (C) Western blot analysis of Col1α1 (left) and Col3α1 (right) expression.

Figure 4.

 

TGF-β2–induced upregulation of TGM2 and CTGF. (A) Northern blot analysis of TGM2 (left) and CTGF (right) mRNA levels in TGF-β2–treated astrocytes (lanes 2 and 4) and untreated controls (Co, lanes 1 and 3). (B) Methylene blue staining of the corresponding 28S and 18S RNA bands. (C) Western blot analysis of TGM2 (left) and CTGF (right) expression.

Figure 5.

 

Immunofluorescence staining for CTGF. (A) Untreated monolayers of cultured astrocytes showed faint staining for CTGF (red). (B) After 24 hours of TGF-β2 treatment CTGF staining intensity was significantly increased (red). Nuclei were DAPI stained (blue). Magnification, ×400.

Figure 6.

 

CTGF siRNA blocked the TGF-β2–induced upregulation of CTGF, FN, and Col4α2. (A) Northern blot analysis of CTGF (left), FN (middle), and Col4α2 (right) mRNA levels in 48-hour TGF-β2–treated astrocytes either transfected with 10 nM CTGF siRNA (lanes 3) or untransfected (lanes 2) before treatment. Untreated cells served as the control (Co, lanes 1). (B) Methylene blue staining of the corresponding 28S and 18S RNA bands. (C) Western blot analysis of CTGF (left), FN (middle), and Col4α2 (right) expression.

Figure 7.

 

CTGF siRNA blocks the TGF-β2–induced upregulation of Col1α1 and TGM2. (A) Ethidium bromide–stained 1%-TAE agarose gel of (left) Col1α1 and (right) GAPDH-specific PCR products on cDNA obtained from 48-hour TGF-β2–treated astrocytes transfected with 10 nM CTGF siRNA (lanes 3), untransfected TGF-β2–treated astrocytes (lanes 2), and untreated controls (lanes 1). (B) Statistical analysis of the band intensities of (A). Mean ± SD; n = 3. Control values = 1. (C) Northern blot analysis of TGM2 mRNA levels in 48-hour TGF-β2–treated astrocytes either transfected with 10 nM CTGF siRNA (top, lane 3) or untransfected (top, lane 2) before treatment, untreated control cells (Co, top, lane 1), and methylene blue–stained corresponding 28S and 18S RNA bands (bottom). (D) Western blot analysis of Col1α1 (left) and TGM2 (right) expression.

Figure 8.

 

TGF-β2 induced upregulation of TSP-1. (A) Northern blot analysis of TSP-1 mRNA levels in TGF-β2–treated astrocytes (top, lanes 2 and 3), untreated controls (Co, top, lane 1), and methylene blue–stained corresponding 28S and 18S RNA bands (bottom). (B) Western blot analysis of TSP-1 expression in untreated astrocytes (Co, lane 1) and after 24 hours (lane 2) and 48 hours TGF-β2 treatment respectively (lane 3). (C) Northern blot analysis of TSP-1 mRNA levels in 48-hour TGF-β2–treated astrocytes transfected with 10 nM CTGF siRNA (top, lane 3), untransfected TGF-β2–treated astrocytes (top, lane 2), untreated controls (Co, top, lane 1), and methylene blue–stained corresponding 28S and 18S RNA bands (bottom).

The authors thank Angelika Pach, Julia Mausolf, and Heide Wiederschein for expert technical assistance.