Activation of the BMP Canonical Signaling Pathway in Human Optic Nerve Head Tissue and Isolated Optic Nerve Head Astrocytes and Lamina Cribrosa Cells

Abstract

purpose. Bone morphogenetic proteins (BMPs) are members of the transforming growth factor (TGF)-β superfamily that controls multiple functions in a variety of cells. We have previously shown that human optic nerve head (ONH) astrocytes and lamina cribrosa (LC) cells express BMP and BMP receptor mRNA and proteins. The purpose of the present study was to determine whether human ONH tissues express the canonical BMP signaling pathway and whether ONH cells secrete BMP-4 and respond to exogenous BMP-4 through this pathway.

methods. Well-characterized human ONH astrocytes (N = 2) and LC cells (N = 3) were treated with exogenous BMP-4 (20 ng/mL) for various times. Western immunoblot analysis was used to detect secreted BMP-4 in serum-free conditioned media of ONH cells and in human ONH tissues (N = 4) and Smad proteins in total cell lysate of ONH cells and tissues. Intracellular colocalization of p-R-Smad1 with Co-Smad4 and localization of inhibitory Smads (e.g., I-Smad6 and I-Smad7) were studied through immunocytochemistry. In addition, coimmunoprecipitation was used to verify the interaction of p-R-Smad1 with Co-Smad4.

results. ONH astrocytes and LC cells secrete BMP-4 and synthesize R-Smad1, R-Smad5, I-Smad6, I-Smad7, and Co-Smad4 proteins. Exposure to BMP-4 for either 10 or 60 minutes resulted in increased p-R-Smad1 and p-R-Smad1/5/8 protein levels that declined after 12 hours of treatment. Immunocytochemistry and coimmunoprecipitation studies revealed that p-R-Smad1/5/8 and Co-Smad4 interact and colocalize in the nucleus. BMP-4 treatment resulted in increased coprecipitation of p-R-Smad1/5/ 8 and Co-Smad4. I-Smad6 and I-Smad7 are localized in the nucleus and cytoplasm of ONH astrocytes and LC cells. Proteins for BMP-4, p-R-Smad1/5/8, R-Smad1, R-Smad5, R-Smad8, and Co-Smad4 are present in human ONH tissues. In addition, phosphorylated Smad1 and Smad5 colocalize with Smad4 in the nuclei of ONH tissues.

conclusions. These results indicate that BMP-4 and Smad signaling proteins are present in human ONH tissues, isolated ONH astrocytes, and LC cells. In addition, exogenous BMP-4 treatment of ONH astrocytes and LC cells results in downstream signaling through the canonical Smad pathway. Thus, cells within the human ONH may respond to locally released BMP through paracrine or autocrine mechanisms.

Glaucoma is a heterogeneous group of optic neuropathies affecting more than 67 million individuals worldwide and is a major cause of visual impairment and irreversible blindness. ¹ Elevated intraocular pressure (IOP) is the major risk factor in the development of glaucomatous optic neuropathy. ^{2 3 4} Elevated IOP has been shown to affect the optic nerve head (ONH) by causing cupping and excavation that is associated with collapse and remodeling of the lamina cribrosa (LC). ^{5 6 7 8} Optic nerve head changes are also associated with inhibited retrograde transport of neurotrophic factors in retinal ganglion cells (RGCs) ^{9 10} and death of RGCs through apoptosis. ^{11 12}  

The LC region is the main structural component of the ONH through which RGC axons exit the eye. ^{11 13} ONH astrocytes and LC cells are two major cell types that can be isolated from this region. ^{8 14} Glaucomatous ONH changes are associated with activation of astrocytes, altered growth factor synthesis, and changes in extracellular matrix (ECM) synthesis or degradation. ^{6 7 11 15 16} Growth factors play an important role in maintaining normal homeostasis in ocular tissues, including the ONH. Maintenance of an intricate balance between growth factors is essential for normal functioning. ¹⁷ Thus, alterations in growth factor secretion or signaling may play a role in the pathogenesis of glaucoma. 

Bone morphogenetic proteins (BMPs) were originally identified as osteoinductive cytokines that promote bone and cartilage formation, but they are now known to control multiple functions in a variety of cells. ^{18 19} BMPs initiate signaling by binding to cell surface type I and type II serine/threonine kinase receptors. ^{19 20 21 22} On ligand binding, the type II BMP receptor transphosphorylates the type I BMP receptor. Downstream BMP signaling involves Smad signaling proteins. Receptor-regulated Smads (R-Smad1, R-Smad5, and R-Smad8) transiently associate with the type I BMP receptor and undergo direct phosphorylation. Subsequently, the phosphorylated R-Smad associates with common Smad4 (Co-Smad4), and the heteromeric complex translocates to the nucleus to regulate target genes. ^{18 19 20 21 22 23 24 25}  

Mice with a heterozygous deficiency of Bmp-4 result in developmental abnormalities of the optic nerve and anterior segment dysgenesis and elevated intraocular pressure, ²⁶ indicating an important role of BMP-4 in the eye. In addition, recent reports have indicated that BMPs can inhibit TGF-β signaling. For example, in mesangial cells of the kidney, BMP-7 inhibits TGF-β signaling by reducing nuclear accumulation of R-Smad3. ²⁷ Interestingly, we have demonstrated that BMP-4 can counteract the action of TGF-β2–induced fibronectin synthesis in human trabecular meshwork (TM) cells. ²⁸  

We have previously reported that isolated ONH astrocytes and LC cells express mRNA and proteins for various BMPs and BMP receptors. ¹⁷ However, for BMPs to be able to signal through autocrine or paracrine mechanisms in the human ONH, it is important to determine whether the canonical BMP/Smad signaling pathway is active. Thus, the objectives of this present study were to determine whether (1) BMP-4 and Smad proteins are present in human ONH cells and tissue, (2) cultured human ONH astrocytes and LC cells secrete BMP-4 and respond to exogenous BMP-4 treatment through the canonical Smad signaling pathway, and (3) human ONH tissues express the canonical BMP/Smad canonical signaling pathway. 

Materials and Methods

Optic Nerve Head Dissection and Cell Culture

Human ONH cells were generated from dissected ONHs and characterized according to previous reports. ^{29 30} Briefly, human donor eyes from regional eye banks were obtained within 24 hours of death, and the LC region of the ONH was dissected from the remaining ocular tissue. Lamina cribrosa tissues were cut into three to four explants and placed in culture plates containing Dulbecco modified Eagle medium (DMEM; Hyclone Laboratories, Logan, UT) containing l-glutamine (0.292 mg/mL; Gibco BRL Life Technologies, Grand Island, NY), penicillin (100 U/mL)/streptomycin (0.1 mg/mL; Gibco BRL Life Technologies), and amphotericin B (4 μg/mL; Gibco BRL Life Technologies). Cells that did not express glial fibrillary acidic protein (GFAP) were characterized as LC cells. Cells that expressed neural cell adhesion molecule [NCAM] and GFAF were characterized as ONH astrocytes. Confluent cells were passaged using 0.25% trypsin (Sigma-Aldrich, St. Louis, MO) and were maintained in 5%/CO₂/95% O₂ at 37°C. 

Treatment with Exogenous BMP-4

ONH astrocytes and LC cells were grown in DMEM plus 10% FBS until they were 70% to 80% confluent. The cells were washed twice with PBS, cultured in serum-free DMEM for 24 hours, and treated with 20 ng/mL recombinant BMP-4 protein (R&D Systems, Minneapolis, MN) for 10 minutes, 30 minutes, 60 minutes, 12 hours, or 24 hours. 

Protein Extraction and Western Blot Analysis

Cell Lysate.

Total cellular protein was extracted from cultured ONH astrocytes and LC cells (Mammalian Protein Extraction Buffer; 78501; Pierce Biotech, Rockford, IL) with protease inhibitor cocktail (78415; Pierce Biotech). Protein concentration was determined using a protein assay system (Bio-Rad Dc; Bio-Rad Laboratories, Richmond, CA). Cellular protein was separated on denaturing polyacrylamide gels and then transferred to polyvinylidene difluoride (PVDF) membranes by electrophoresis. ²⁹ The blots were blocked (SuperBlock Blocking Buffer, Prod 37537; Pierce Biotech) for 2 hours. The blots were then incubated overnight with the following primary antibodies: R-Smad1 (9512; Cell Signaling, Beverly, MA), R-Smad 5 (51–3700; Zymed, San Francisco, CA), p-R-Smad1 (566411; Calbiochem, La Jolla, CA), p-R-Smad1/5/8 (AB3848; Chemicon International, Temecula, CA), I-Smad6 (Zymed), I-Smad7 (Santa Cruz Biotechnology, Santa Cruz, CA), and Co-Smad4 (MAB 1132; Chemicon International). The membranes were washed with Tris-buffered saline Tween buffer (TBST) and processed with corresponding horseradish peroxidase-conjugated secondary antibody (donkey anti-rabbit [SC-2077] or goat anti-mouse [SC-2005]; Santa Cruz Biotechnology). The proteins were then visualized in an imager (Fluor Chem 8900; Alpha Innotech, San Leandro, CA) using enhanced chemiluminescence detection reagents (SuperSignal West Femto Maximum Sensitivity Substrate; Pierce Biotechnology). To ensure equal protein loading, the same blot—developed with a horseradish peroxidase-conjugated secondary antibody (goat anti-mouse [SC-2005]; Santa Cruz Biotechnology)—was incubated again with a β-actin monoclonal antibody (Chemicon International). 

Secreted BMP-4.

To detect secreted BMP in conditioned medium, cells were maintained in serum-free DMEM for either 24 hours or 48 hours, and conditioned medium was concentrated 50-fold using a resin (StrataClean Resin, catalog no. 400714–61; Stratagene, La Jolla, CA). To ensure equal loading of protein, equal volumes of conditioned medium were loaded. Western blot analysis was conducted as described, and the blots were incubated overnight with a BMP-4 antibody (mAb757; R&D Systems). The secondary antibody consisted of goat anti-mouse (SC-2005; Santa Cruz Biotechnology). 

Tissue Lysate.

Four normal human eyes (donor age, 53–90 years) were obtained from regional eye banks within 6 hours of death. ONH was isolated and collected in buffer (Mammalian Protein Extraction Buffer, 78501; Pierce Biotech, Rockford, IL) with protease inhibitor cocktail (78415; Pierce Biotech). Samples were sonicated three times for 5 seconds each and kept on ice for 15 minutes. Solubilized proteins were centrifuged, and supernatants were used for Western blot analysis of BMP-4 and Smad signaling proteins. 

Immunostaining

ONH astrocytes and LC cells were grown on glass coverslips, fixed with 3.5% (vol/vol) formaldehyde in PBS, and treated with 0.02% (vol/vol) Triton X-100 in PBS (Fisher Scientific, Pittsburgh, PA). Nonspecific binding was blocked by 2-hour incubation with 10% (vol/vol) normal serum in PBS (Gibco BRL Life Technologies). For localization of I-Smad6 and I-Smad7, coverslips were incubated overnight with I-Smad6 antibody (Zymed) and I-Smad7 antibody (Santa Cruz Biotechnology). For colocalization of p-R-Smad1 with Co-Smad4, coverslips were incubated overnight with p-R-Smad1 antibody (566411; Calbiochem) and co-Smad4 antibody (mAb 1132; Chemicon International) diluted in 1.5% (vol/vol) normal serum, followed by a 2-hour incubation in appropriate Alexa Fluor 488 donkey anti-rabbit and 633-labeled donkey anti-mouse secondary antibodies (1:200 in 1.5% [vol/vol] normal serum; Molecular Probes, Inc., Eugene, OR). The IgG controls used rabbit and mouse IgG in place of the primary antibody. Coverslips incubated in 1.5% (vol/vol) normal serum in place of primary antibody served as negative controls. To visualize nuclei, sections were treated with 4′,6′-diamino-2-phenylindole (DAPI) nuclear stain for 30 minutes. Coverslips were then mounted on clean glass slides using mountant (Aquamount; Lerner Laboratories Inc., Pittsburgh, PA). Images were captured using a confocal imaging system (Zeiss 410; Carl Zeiss, Thornwood, NY). 

For immunostaining of BMP-4 in ONH tissues, three normal human donor eyes were obtained from regional eye banks within 6 hours of death and fixed in 10% formalin. Fixed tissues were dehydrated and embedded in paraffin, and 8-μm sections were obtained. Sections were deparaffinized, rehydrated, and placed in 0.02% Triton X-100, followed by 20 mM glycine for 15 minutes each. Sections were blocked in 10% normal serum and incubated with anti–BMP-4 antibody (mAb1049; Chemicon International). After three washes, sections were incubated with appropriate secondary antibody (Alexa Fluor donkey anti-mouse, 1:200 in 1.5% [vol/vol] normal serum; Molecular Probes, Inc.) for 45 minutes. Sections were treated with DAPI for 30 minutes, washed, and mounted. Images were captured using a confocal imaging system (Zeiss 410; Carl Zeiss, Thornwood, NY). 

For colocalization of p-R-Smad1 with Co-Smad4 in ONH tissues, coverslips were incubated overnight with p-R-Smad1 antibody and Co-Smad4 antibody diluted 1:100 in 1.5% (vol/vol) normal serum, followed by 2-hour incubation in appropriate Alexa Fluor488 donkey anti-rabbit and 633-labeled donkey anti-mouse secondary antibodies. Sections were treated with DAPI for 30 minutes, washed, and mounted. Images were captured using a confocal imaging system (Zeiss 410; Carl Zeiss). 

Coimmunoprecipitation

ONH astrocyte and LC cell lysates were solubilized in immunoprecipitation buffer (M-PER Mammalian Protein Extraction Buffer; Pierce Biotechnology) with an excess amount of immunoprecipitating p-R-Smad1/5/8 antibody (3 μg). The total volume was adjusted to 0.5 mL with immunoprecipitation buffer. Samples were then incubated overnight at 4°C. An appropriate amount of immobilized protein G (Sigma-Aldrich) was added, and samples were incubated overnight with gentle mixing at 4°C. The samples were then washed six times with the immunoprecipitating buffer and were eluted using SDS-PAGE electrophoresis buffer, which was subsequently immunoblotted for interacting Co-Smad-4 protein. 

Results

ONH Astrocytes and LC Cells Secrete BMP-4

We have previously shown that cells isolated from human ONH express mRNA and proteins for BMPs and BMP receptors. ¹⁷ In this study, we first sought to determine whether cells isolated from the human ONH secrete BMPs. As demonstrated in Figure 1 , ONH astrocytes and LC cells secrete BMP-4. A recombinant protein for BMP-4 was used as a positive control for antibody specificity. Two bands at 22 kDa and 25 kDa, corresponding to the glycosylated and nonglycosylated forms of BMP-4, were detected. Secretion of BMP-4 appeared to increase in most cell lines from 24 to 48 hours. There was no apparent difference in secretion of BMP-4 between ONH astrocytes and LC cells. Although we previously demonstrated that ONH astrocytes and LC cells make proteins for BMP-4, BMP-5, and BMP-2, in Western blot analysis of conditioned medium, we found that ONH astrocytes and LC cells secrete higher amounts of BMP-4 (Fig. 1)than BMP-5 (data not shown). BMP-2 secretion was not detected in ONH cells. 

Presence of Smad Signaling Proteins in ONH Astrocytes and LC Cells

The canonical downstream signaling pathway for BMPs uses intracellular Smad proteins. ^{18 20 21 25} Thus, we sought to determine whether R-Smads (p-R-Smad1, R-Smad1, and R-Smad5) and Co-Smad4 were present in ONH astrocytes and LC cells. Figure 2demonstrates Western blot analysis for several Smad signaling proteins in each of two ONH astrocyte and LC cell lines. Proteins for R-Smad1, R-Smad5, and Co-Smad4 were detected in the cell lysates of ONH astrocytes and LC cells. In addition, Western blot analysis detected phosphorylated R-Smad1 (p-R-Smad1) protein in cell lysates of ONH astrocytes and LC cells in the absence of exogenous BMP, indicating autocrine/paracrine BMP signaling was occurring. Antibodies specific for phosphorylated R-Smad5 were unavailable. There were no apparent differences in protein levels of p-R-Smad1, R-Smad1, R-Smad5, and Co-Smad4 between ONH astrocytes and LC cells. 

Coimmunolocalization of Phosphorylated R-Smad1 (p-R-Smad1) and Co-Smad4 in ONH Astrocytes and LC Cells

BMP ligands bind to cell-surface BMP receptors and activate downstream signaling through phosphorylation of R-Smads. ^{20 21 22} Phosphorylated R-Smads form complexes with Co-Smad4 in the cytoplasm and then together translocate to the nucleus. We therefore examined colocalization of p-R-Smad1 and Co-Smad4 in ONH astrocytes and LC cells. Representative images of colocalization of Co-Smad4 and p-R-Smad1 in ONH astrocytes and LC cells are shown in Figure 3 . Nuclear and cytoplasmic colocalization of p-R-Smad1 with Co-Smad4 were observed in untreated ONH astrocytes and LC cells, again indicating endogenous autocrine/paracrine Smad signaling was occurring. No staining was observed with nonimmune serum (data not shown) or when primary antibody was omitted. 

Exogenous BMP-4 Increases Phosphorylated Smad Proteins in ONH Astrocytes and LC Cells

Having shown that ONH astrocytes and LC cells secrete BMP-4 and synthesize R-Smad and Co-Smad4 signaling proteins, we next sought to determine whether exogenous BMP-4 causes activation of downstream R-Smad signaling proteins. As shown in Figure 4 , we examined protein levels of p-R-Smad1, p-R-Smad1/5/8, R-Smad1, R-Smad5, and Co-Smad4 in BMP-4–treated ONH astrocytes (Fig. 4A)and LC cells (Fig. 4C) . Increased protein levels of phosphorylated R-Smad1 and R-Smad1/5/8 were detected in ONH astrocytes and LC cells when treated with exogenous BMP-4. Densitometry analysis demonstrated that phosphorylation of R-Smad1 and R-Smad1/5/8 increased after 10 minutes of BMP-4 treatment and washighest at 60 minutes of BMP-4 treatment (Figs. 4B 4D) . There was a fourfold increase in the phosphorylation of R-Smad1/5/8 over R-Smad1 in ONH astrocytes and LC cells when treated with BMP-4 for 60 minutes. Phosphorylation appeared to decrease after 12 hours. Levels of nonphosphorylated R-Smad1 and R-Smad5 did not change in ONH astrocytes and LC cells when treated with exogenous BMP-4. Thus, our data indicate that ONH astrocytes and LC cells are capable of responding to exogenous BMP-4 through the canonical Smad signaling proteins (e.g., R-Smads and Co-Smad4). 

BMP-4 Treatment Increases Coimmunoprecipitation of Co-Smad4 with p-R-Smad1/5/8 in ONH Astrocytes and LC Cells

To verify our coimmunolocalization study, we coimmunoprecipitated p-R-Smad1/5/8 with Co-Smad4 to determine whether BMP-4 treatment would increase the interaction between Co-Smad4 and p-R-Smad1. Co-Smad4 was detected in the Western blot when the p-R-Smad1/5/8 antibody was used to immunoprecipitate the cell lysate, followed by immunoblotting for Co-Smad4 (Fig. 5) . The coimmunoprecipitation appeared to increase after 60 minutes of BMP-4 treatment and declined after 2 hours of BMP-4 treatment. 

Presence of Inhibitory Smads (I-Smad6 and I-Smad7) in ONH Astrocytes and LC Cells

BMP and TGF-β signaling are regulated by a variety of extracellular and intracellular mechanisms. ^{18 31 32 33} Intracellular control of BMP and TGF-β signaling is mediated by the inhibitory Smads (e.g., I-Smad6 and I-Smad7). I-Smads are located predominantly in the nucleus but can be transported rapidly into the cytoplasm after ligand stimulation. In the cytoplasm, they bind to R-Smads and prevent interaction with Co-Smad4 and subsequent translocation to the nucleus. ^{18 20 21 32 34} We used immunostaining and Western blot analysis to determine whether ONH astrocytes and LC cells express I-Smads. Positive immunostaining for I-Smad6 and I-Smad7 was observed in LC (Figs. 6A 6B)and ONH astrocyte (Figs. 6C 6D)cell lines. Nuclear and cytoplasmic localization of I-Smad6 and I-Smad7 were observed. No staining was observed with nonimmune serum or when primary antibody was omitted. To support the immunolocalization data, we performed Western blot analysis for I-Smad6 and I-Smad7 in total cell lysate of ONH astrocytes and LC cell lines. I-Smad6 and I-Smad-7 were present in astrocytes and LC cells (Fig. 6E) . With respect to the Western blot analysis, I-Smad7 protein appeared to be present in greater amounts in ONH astrocyte and LC cell lines, suggesting that intracellular control of BMP and TGF-β signaling in ONH astrocytes and LC cells may occur via I-Smads. 

BMP-4 Is Localized in Human ONH Tissues

Having shown secretion of BMP-4 and the presence of endogenous Smad signaling in cultured ONH astrocytes and LC cells, we sought to determine whether BMP-4 is localized to ONH tissues obtained from postmortem human donors. We performed immunostaining on paraffin-embedded human ONH sections. H&E staining revealed the orientation of retina and lamina cribrosa region (Fig. 7A) . Immunostaining of ONH tissues demonstrated that BMP-4 is localized in ONH (Fig. 7B)and is more prominent in the LC region (Fig. 7C) . 

Presence of BMP-4 and Smad Signaling Proteins in Human Optic Nerve Head Tissues

We performed Western blot analysis of human ONH tissues to determine whether BMP-4 and Smad signaling proteins are present in vivo. BMP-4 was present in ONH tissues (Fig. 8) . Western blot analysis also demonstrated the presence of phosphorylated R-Smad1, total R-Smad1, R-Smad5, and Co-Smad4 at their expected relative sizes. The presence of BMP-4 and phosphorylated Smads in human ONH tissues indicates that Smad signaling can occur in the human ONH. 

Coimmunolocalization of R-Smad5, Phosphorylated R-Smad1 (p-R-Smad1) with Co-Smad4 in Human ONH Tissue

Having shown the presence of BMP-4 and Smad signaling proteins in human ONH tissues, we next sought to determine whether phosphorylated R-Smads colocalize in the nucleus in situ. We performed colocalization of p-R-Smad1 and R-Smad5 with Smad4 in ONH tissues (Fig. 9) . Nuclear and cytoplasmic colocalization of p-R-Smad1 and Smad5 with Co-Smad4 was observed in ONH tissues, indicating the presence of endogenous Smad signaling in the human ONH. No staining was observed with IgG treatment or when primary antibody was omitted. Interestingly, increased colocalization of R-Smad5 with Smad4 was observed compared with R-Smad1, indicating that R-Smad5 may play a major role in endogenous Smad signaling in the human ONH. 

Discussion

Glaucomatous damage to the ONH consists of cupping and excavation of the optic disc, collapse and remodeling of the LC, and activation of optic nerve head astrocytes. ^{7 8 11 15 35 36} Many of these changes are induced by elevated IOP because similar ONH changes are seen in animal models of ocular hypertension-induced glaucoma. ^{13 16 37} However, glaucomatous damage to the ONH also occurs at low IOP in some glaucoma patients. Growth factors appear to play an important role in maintaining normal homeostasis in the ONH, and alterations in growth factors or growth factor receptors may be involved in glaucoma pathogenesis of the ONH. ²⁸ We have reported that cells isolated from the human ONH express growth factors, including neurotrophins and their receptors. ^{29 30} Members of the TGF-β superfamily of growth factors have been implicated in glaucomatous changes in the ONH. ^{35 38 39} Elevated levels of TGF-β2 in the ONH are associated with glaucomatous changes in the ECM, and TGF-β2 and connective tissue growth factor alter ECM metabolism in cultured ONH cells. ³⁸ TGF-β1 expression is increased in stretched human LC cells, simulating the backward bowing of the LC tissue in glaucoma. ^{40 41}  

Other members of the TGF-β family of growth factors include the BMPs. ¹⁹ We have previously demonstrated that human ONH astrocytes and LC cells express mRNA for BMPs and BMP receptors and synthesize proteins for BMP-2, BMP-5, BMP-4, BMP-7, and BMP receptors BMPRIA, BMPRIB and BMPRII. ¹⁷ Although BMPs were initially identified as osteoinductive factors that promote bone and cartilage formation, they also regulate a number of cellular functions in other tissues, including development, morphogenesis, cell proliferation, and apoptosis. ¹⁹ A good example of the role of BMPs in the eye is the finding that mice with a heterozygous deficiency of Bmp-4 have developmental abnormalities of the optic nerve, anterior segment dysgenesis, and elevated intraocular pressure. ²⁶  

BMPs also appear to modulate TGF-β2 signaling and fibrosis. BMP-7 inhibits TGFβ-induced kidney fibrosis, ²⁷ and we recently showed that BMP-4 inhibits TGF-β2 induction of fibronectin in cultured human trabecular meshwork cells. ³⁰ It is interesting to speculate that BMPs may act similarly in the ONH given the apparent involvement of TGF-β in damage and remodeling of the glaucomatous ONH and the finding that ONH cells and tissue express BMPs. As a first step to evaluate the potential role of BMPs in the ONH, we determined whether isolated ONH astrocytes and LC cells secrete BMP-4, express Smad proteins of the canonical BMP signaling pathway, and respond to exogenous BMP-4. We also sought to determine whether BMP-4 and proteins for the canonical BMP signaling pathway were present in human ONH tissues. 

Our data demonstrate that ONH astrocytes and LC cells constitutively secrete BMP-4 and express R-Smad1, R-Smad5, and Co-Smad4 proteins. Phosphorylated R-Smad1 was present in both cell types, and p-R-Smad1 and Co-Samd4 were found within cell nuclei. Co-Smad4 requires association with activated R-Smads to enter the nucleus. ²⁵ The presence of Co-Smad4 in the nuclei of ONH astrocytes and LC cells in conjunction with p-R-Smad1 indicates that translocation of the signaling complex occurs in the absence of exogenous BMP-4. Together these data indicate that autocrine BMP signaling occurs in cultured ONH astrocytes and LC cells. We previously reported that BMP protein and BMP receptor proteins are present in ONH tissues, ¹⁷ suggesting in vivo BMP autocrine signaling. 

In addition to Smad signaling in ONH cells, we demonstrate that BMP-4 is localized in the ONH and is more prominent in the LC region. The presence of BMP-4, R-Smad1, R-Smad5, Co-Smad4, and phosphorylated R-Smad1/5/8 in the human ONH and the finding that pSmad1 and Smad5 colocalize with Smad4 in the nucleus indicates activation of the canonical Smad signaling pathway in the ONH tissues. To our knowledge, this is the first study to demonstrate that Smad signaling occurs in human ONH cells and tissues. 

In addition to our evidence for BMP autocrine signaling, human ONH astrocytes and LC cells also to respond to exogenous BMP-4 with increased activation of the Smad signaling pathway The addition of exogenous BMP-4 increased the phosphorylation of p-R-Smad1 and p-R-Smad1, p-R-Smad5, and p-R-Smad8 (specific antibodies to the phosphorylated forms of R-Smad5 and R-Smad8 are unavailable). We also found increased coprecipitation of Smad4 with pSmad1/5/8. The increased protein levels recognized by the p-R-Smad1/5/ 8 antibody was fivefold greater than the increase in protein levels of p-R-Smad1, indicating that R-Smad5 or R-Smad8 may be activated to a greater degree by BMP-4 in ONH astrocytes and LC cells. The inhibitory effect of BMP on TGF-β2 signaling in mesangial cells is mediated mainly through R-Smad5, ²⁷ supporting the preferential involvement of select R-Smads in some BMP signaling pathways. Taken together, these results indicate that ONH astrocytes and LC cells respond to exogenous BMP-4 through the phosphorylation of R-Smads and increased interaction with Smad4, and they suggest that R-Smad5 and R-Smad8 are the major R-Smads used by ONH astrocytes and LC cells in BMP signaling. 

The BMP signaling pathway is tightly controlled in virtually all cells. ¹⁸ Extracellular and intracellular control mechanisms have been reported. ²⁵ The most widely studied intracellular control mechanism involves inhibitory Smads (I-Smads). ^{20 32 34} I-Smad6 and I-Smad7 compete with the receptor Smads (e.g., R-Smad1, R-Smad5, R-Smad8) for binding to the activated BMPRI receptor. Inhibitory Smads function to block transmission of signals from the membrane to the nucleus and are located predominantly in the nucleus but are transported rapidly to the cytoplasm after ligand stimulation. ^{20 21} Inhibitory Smad6 can inhibit BMP signaling by competing with Co-Smad4 for complex formation with R-Smads. Inhibitory Smad6 is induced by various signals such as mechanical stress, TGF-β1, and BMPs. Smad6 preferentially inhibits BMP signaling, whereas Smad7 inhibits TGF-β and BMP signaling. ^{19 20 21 31} In our study, we also focused on the presence of I-Smad proteins in cells isolated from the human ONH. Our results indicate that I-Smad6 and I-Smad7 were present in the nucleus and cytoplasm of the ONH astrocytes and LC cells in the absence of BMP, suggesting that both I-Smads are constitutively expressed in these cell lines. The presence of both I-Smads indicates that the inhibition of BMP signaling may be occurring. 

The function of BMP secretion and signaling within the human ONH may be multifaceted. BMP proteins could use unique autocrine/paracrine signaling mechanisms to regulate homeostasis and the microenvironment of the normal ONH. BMP signaling within the human ONH may regulate the local cellular activity of TGF-β2. Within a given tissue, the action of most growth factors is often counterbalanced by the action of other growth factors. ²⁸ This results in only small changes in tissue structure and function. TGF-β2 is capable of inducing the expression of ECM and basement membrane components in cultured ONH astrocytes. ³⁸ Thus autocrine/paracrine BMP signaling may be involved in the regulation of TGF-β2 effects on ECM components in the ONH. Elevated levels of TGF-β2 in glaucoma may disrupt the intricate balance between BMP and TGF-β2, leading to increased deposition of ECM proteins within the ONH. 

In conclusion, we show that ONH astrocytes and LC cells secrete BMP-4. ONH astrocytes and LC cells make Smad signaling proteins that undergo phosphorylation, suggesting BMP endogenous signaling. Exposure to exogenous BMP-4 caused increased phosphorylation of R-Smad and increased interaction R-Smad with Smad4, suggesting that ONH astrocytes and LC cells respond to exogenous BMP. Our results demonstrate that autocrine/paracrine BMP signaling occurs in ONH astrocytes and LC cells. Cytoplasmic localization of inhibitory Smads in ONH astrocytes and LC cells suggest that these factors play roles in regulating autocrine/paracrine BMP signaling. We also show that BMP-4 is localized in human ONH tissues. The presence of phosphorylated Smad1, total Smad5, and Smad4 and their colocalization in the nucleus indicates the presence of autocrine/paracrine Smad signaling in ONH tissues. These studies demonstrate that ONH astrocytes and LC cells are capable of responding to exogenous BMP-4 through receptor Smads. Thus, cells of the ONH may be targets for and may respond to locally released BMP. 

 Supported by Alcon Research Ltd. and by National Institutes of Health Grant R01 EY12783 (RJW).

 Submitted for publication February 1, 2007; revised May 21 and July 2, 2007; accepted August 28, 2007.

 Disclosure: G.S. Zode, None; A.F. Clark, Alcon (F, I, E); R.J. Wordinger, 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.

 

Figure 1.

 

Chemiluminescence detection of BMP-4 secreted by human ONH astrocytes and lamina cribrosa. Two ONH astrocyte and two LC cell lines were grown in serum-free DMEM for 24 to 48 hours. Conditioned medium was collected and concentrated 50 times before it was loaded onto a gel. A recombinant protein for BMP-4 (50 ng/lane) was loaded as a positive control.

Figure 2.

 

Chemiluminescence detection of Smad proteins in ONH astrocytes and LC cells. Total cellular protein was collected from two ONH astrocyte and two LC cell lines and electrophoresed in SDS-PAGE gels; this was followed by Western immunoblotting. Both cell types expressed intermediate Smad signaling proteins, including R-Smad1, R-Smad5, and Co-Smad4 and phosphorylated R-Smad1. β-actin was used as internal loading control.

Figure 3.

 

Colocalization of Smad4 and pSmad1 in ONH astrocytes and LC cells. ONH astrocytes and LC cells were kept in serum-free medium for 24 hours. Cells were then fixed and stained with antibodies for Co-Smad4 (monoclonal antibody) and p-R-Smad1 (rabbit antibody), followed by incubation with two secondary antibodies, donkey anti-mouse Alexa Fluor 633 (red) and donkey anti-rabbit Alexa Fluor 488 (green). (A, C) Colocalization of Smad4 (red) with pSmad1 (green) in cultured LC cells and ONH astrocytes, respectively. DAPI stain was used to counterstain the nucleus blue. (B, D) Cells incubated without primary antibody or with IgG (not shown) were used as negative controls. Colocalization of Co-Smad4 and p-R-Smad1 in the nucleus in untreated cells indicates the presence of endogenous BMP signaling.

Figure 4.

 

Chemiluminescence detection of Smad proteins in BMP-4–treated ONH astrocytes and LC cells. Human ONH astrocytes and LC cells were treated with exogenous BMP-4 (20 ng/mL) for various times (10 minutes, 30 minutes, 60 minutes, 12 hours, 24 hours) and compared with vehicle control (0 minutes). Phosphorylated R-Smad1, phosphorylated R-Smad1/5/8, R-Smad5, R-Smad1, and co-Smad4 were measured by Western blot and analyzed by densitometry. β-actin was used as loading control. (A, B) Western immunoblot and corresponding densitometric analysis of relative R-pSmad1 and R-pSmad1/5/8 protein levels normalized to β-actin in LC cells. (C, D) Western immunoblot and corresponding densitometric analysis of relative R-pSmad1 and R-pSmad1/5/8 protein levels normalized to β-actin in ONH astrocytes.

Figure 5.

 

Coimmunoprecipitation study of Co-Smad4 with phosphorylated R-Smad1/5/8 in ONH astrocytes and LC cells. ONH astrocytes and LC cells were treated with exogenous BMP-4 for various times (0 minutes, 20 minutes, 60 minutes, 2 hours). Equal amount of cell lysate was incubated with immunoprecipitating antibody p-R-Smad1/5/8 and immunoblotted for co-Smad4. Co-Smad4 immunoprecipitated with p-R-Smad1/5/8 in ONH astrocytes and LC cells.

Figure 6.

 

Localization of Smad6 and Smad7 in human ONH astrocytes and LC cells. Human ONH astrocytes and LC cells were fixed and stained with antibodies for I-Smad6 and I-Smad7. (A, B) Immunostaining for I-Smad6, I-Smad7, IgG, and no primary antibody control in ONH astrocytes. (C, D) Immunostaining for I-Smad6, I-Smad7, IgG, no primary antibody control in LC cells. Positive staining of these inhibitory Smads in the nucleus and cytoplasm indicates an active role of inhibitory I-Smad6 and I-Smad7 in ONH astrocytes and LC cells. Cells incubated without primary antibody or with IgG were used as negative control. (E) Western immunoblot analysis of I-Smad6 and I-Smad7 in two ONH astrocytes and two LC cell lines. Western blot demonstrated the presence of Smad6 and Smad7 proteins of 66 kDa and 60 kDa, respectively.

Figure 7.

 

Immunohistochemical localization of BMP-4 in human ONH tissue. Representative immunohistochemical localization of BMP-4 in human ONH tissue shown in Figure 7. Four normal human eyes were fixed, sectioned, and stained with anti BMP-4 antibody. Slides were incubated in 1.5% PBS-BSA without primary antibody and with IgG as negative control. (A) H&E staining of ONH tissue (100×) and (B) BMP-4 localization (100×) were used for orientation of staining in the LC region. (C) BMP-4 localization in the human ONH at higher magnification (400×). (D) No primary control in the human ONH (400×).

Figure 8.

 

Western blot analysis of BMP-4 and Smad signaling proteins by human ONH tissues. Western blot analysis of BMP-4 (58 kDa), p-R-Smad1/5/8 (66 kDa), R-Smad1 (60 kDa), R-Smad5 (60 kDa), and co-Smad4 (66 kDa) in four human ONH tissues. Approximately 15 μg total protein lysate was obtained from four human eye donors (ages 53, 80, 85, 88 years) and loaded on SDS-PAGE gel. β-actin was used as loading control. The Western blot demonstrates the presence of BMP-4 and Smad signaling proteins in human ONH tissues.

Figure 9.

 

Colocalization of Smad4 with R-pSmad1 and Smad5 in ONH tissue. (A) Colocalization of pSmad1 (green) with Smad4 (red) in human ONH tissue (1200×). (B) Colocalization of Smad5 (green) with Smad4 (red) in human ONH tissue (1200×). (C) Colocalization of Smad5 with Smad4 at higher magnification (2800×) of the sections used in (B). (D) IgG and no primary antibody control with a DAPI (400×). Colocalization of pSmad1 and Smad5 with smad4 in the nucleus and cytoplasm indicates the presence of endogenous Smad signaling in human ONH tissues.

The authors thank Paula Billman (Alcon Research, Ltd.) and the Central Florida Lions Eye and Tissue Bank for ocular tissue and Anne-Marie Brun for her technical assistance.