TGFβ1-Dependent Contraction of Fibroblasts Is Mediated by the PDGFα Receptor

Abstract

purpose. Contraction of fibroblasts and the resultant tractional force is a contributing factor to fibrotic diseases of the eye, such as proliferative vitreoretinopathy (PVR). Transforming growth factor (TGF)-β is abundant in the eye, and is one of the growth factors thought to contribute to the development of PVR. A second is platelet-derived growth factor (PDGF). In the current study, the relationship between TGFβ1 and PDGF was investigated at the level of cellular contraction.

methods. To study cellular contraction, an in vitro type I collagen gel contraction assay was used with a panel of fibroblast lines that expressed the PDGFα receptor (αPDGFR) or PDGFβ receptor (βPDGFR) or no PDGFRs. The agents tested included rabbit vitreous, TGFβ1, and PDGF.

results. Vitreous promoted cellular contraction, and approximately 60% of this activity was eliminated by preincubation of the vitreous with neutralizing TGFβ antibodies. The αPDGFR-expressing cells responded better than cells expressing the βPDGFR or no PDGFRs. Both of the PDGFR-expressing cell lines contracted in response to PDGF, whereas the best response to TGFβ1 was observed with cells expressing theα PDGFR. Finally, TGFβ1 promoted the tyrosine phosphorylation of both of the PDGFRs, and the αPDGFR was more strongly phosphorylated than the βPDGFR.

conclusions. The results show that the vitreous promotes cellular contraction, that TGFβ is the major factor responsible, and that at least a portion of the TGFβ-dependent contraction proceeds through the αPDGFR—that is, indirectly. Therefore, the αPDGFR is responsible for mediating cellular contraction of multiple growth factors: TGFβ and members of the PDGF family.

Proliferative vitreoretinopathy (PVR) is characterized by the formation of a membrane in front of the retina that is composed of extracellular matrix (ECM) and cells. Contraction of the epiretinal membrane results in tractional retinal detachment (TRD). ^{1 2} Once the retina loses its functional contact with the underlying layer of retinal pigment epithelial (RPE) cells, it is irreversibly damaged by apoptosis of the photoreceptors. ^{3 4} PVR occurs in up to 10% of patients who undergo surgery to reattach the retina. ⁵ Anatomic correction is achieved in 60% to 80% of the patients who require additional surgery. ¹  

Contraction of the epiretinal membrane is likely to involve integrin-dependent interactions between the cells and the ECM. Evidence supporting this idea includes the findings that administration of the Arg-Gly-Asp (RGD)-containing peptide, which interferes with cellular attachment to the ECM, prevents PVR. ^{6 7 8} Furthermore, the typical PVR membrane is mainly composed of collagens I, II, and III. ⁹ Such ECM components bind to cells through integrins, such as α2β1, that are expressed by mesenchymal cells and induced by PDGF and other growth factors. ^{10 11 12} Thus, understanding cellular contraction is likely to provide insight into the pathogenesis of PVR and also to identify new approaches for treatment. 

As mentioned, growth factors appear to be important contributors to PVR. Transforming growth factor (TGF)-β and platelet-derived growth factor (PDGF) have been most strongly implicated, and interleukin (IL)-6, fibroblast growth factor, and hepatocyte growth factor may also contribute. ^{13 14 15 16 17 18 19 20} TGFβ is present in the vitreous under normal conditions and is upregulated in PVR. ¹⁶ PDGF is present in the vitreous of patients with PVR, ¹⁵ and PDGF receptors (PDGFRs) are detected in PVR membranes excised from humans. ²⁰ Furthermore, cells unable to respond to PDGF induce PVR poorly in a rabbit model of the disease, and re-expression of the PDGFα receptor (αPDGFR) markedly elevates the PVR potential of these cells. ¹³ Similarly, inhibiting the endogenous PDGFR by expressing a dominant negative PDGFR mutant suppresses the PVR potential of rabbit conjunctival fibroblasts. ¹⁷  

The receptors for PDGF and TGFβ are from fundamentally different classes of growth factor receptors. The receptor for TGFβ is a ubiquitously expressed transmembrane protein that encodes a serine-threonine kinase within the cytoplasmic domain. Binding of TGFβ to its receptor results in the phosphorylation of the Smad family of transcription factors. ²¹ As a result of phosphorylation, the Smads move from the cytoplasm into the nucleus, where they regulate gene expression. ²¹ In contrast, the receptors for PDGF are tyrosine kinase–encoding receptors and trigger cellular responses, primarily through signaling cascades that involve SH2 domain–containing proteins. ^{22 23} TGFβ is able to indirectly activate the PDGFR by promoting the synthesis and secretion of PDGF. ^{24 25 26}  

In this study we focused on the relationship between TGFβ1 and PDGF in cell contraction. We report that TGFβ1 is the major agent in the vitreous responsible for initiating cell contraction, and this response appears to proceed through the αPDGFR. 

Materials and Methods

Cells

F cells are a simian virus (SV)40–immortalized line of mouse embryo fibroblasts derived from mice nullizygous for both the α- andβ PDGFRs. They were generously provided by Michelle Tallquist and Philippe Soriano (Fred Hutchinson Cancer Research Center, Seattle, WA). The Fα and Fβ cells express only one of the PDGFRs, the α- or βPDGFR, respectively. ¹³ The FCX² cells are F cells infected with an empty expression vector. The generation, characterization, and maintenance of these cell lines have been described. ¹³ Normal growth conditions were Dulbecco’s modified Eagle’s medium (DMEM) with high glucose and 10% fetal bovine serum (FBS). The serum concentration was reduced to 1% when the cells were serum starved. 

Immunoprecipitation and Immunoblot Analysis

Cells were grown to 80% confluence, incubated in DMEM containing 1% FBS for 20 hours, and exposed at 37°C for 5 minutes to 50 ng/mL PDGF-BB or left unstimulated. After treatment, the cells were washed twice with H/S (20 mM HEPES [pH 7.4] and 150 mM NaCl) and then lysed in EB (10 mM Tris-HCl [pH 7.4], 5 mM EDTA, 50 mM NaCl, 50 mM NaF, 1% Triton X-100, 0.1% BSA, 20 μg/mL aprotinin, 2 mM Na₃VO₄, and 1 mM phenylmethylsulfonyl fluoride [PMSF]). Lysates were centrifuged for 15 minutes at 13,000g, the pellet was discarded, and the soluble fraction was used as the total cell lysate. The protein concentration was measured with a protein assay kit (Pierce, Rockford, IL) according to the manufacturer’s instructions. 

Receptors were immunoprecipitated from the soluble fraction with the 27P or 30A antibody. ^{27 28} Both are rabbit polyclonal antibodies that recognize the carboxyl terminus of the α- orβ PDGFR, respectively. They were made against a glutathione-S- transferase fusion protein encoding the entire C terminus of the human αPDGFR (amino acids 951-1089) orβ PDGFR (amino acids 958-1106). The antibodies are monospecific—that is, the PDGFR is the predominant species recognized in total cell lysates. Immune complexes were bound to formalin-fixed membranes of Staphylococcus aureus, spun through an EB sucrose gradient, and washed twice with EB and then with PAN (10 mM piperazine-N,N′-bis (2-ethanesulfonic acid)[ PIPES; pH 7.0] 100 mM NaCl, and 1% aprotinin) with 0.5% Nonidet P (NP)-40, and finally again with PAN. 

Receptor immunoprecipitates from 1.0 × 10⁶ cells were resolved in 7.5% SDS-PAGE gel under reducing conditions. Proteins were transferred onto membranes (Immobilon; Millipore, Bedford, MA), and the membranes were blocked (10 mM Tris-HCl [pH 7.5], 1.5 M Tris base, 150 mM NaCl, 10 mg/mL BSA, 10 mg/mL ovalbumin, and 0.05% Tween 20; Block) for anti-phosphotyrosine blot analysis. The membranes were blocked (10 mM Tris-HCl [pH 7.5], 1.5 M Tris base, 150 mM NaCl, 10 mg/mL nonfat dry milk, and 0.05% Tween 20; Blotto) for other antibodies. Membranes were incubated with primary antibodies for 1 hour at room temperature and washed five times (150 mM NaCl, 10 mM Tris-HCl [pH 7.5], and 1.5 mM Tris base; Western Rinse solution). Afterward, they were incubated with secondary antibody for 1 hour at room temperature, washed five times with Western Rinse, and visualized using enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech, Piscataway, NJ). 

Reagents and Antibodies

Recombinant human TGFβ1, PDGF-BB, neutralizing anti-pan-TGFβ antibody, anti-PDGF antibody, and control affinity-purified goat or rabbit IgG were purchased from R&D Systems (Minneapolis, MN). Anti-TGFβ antibody neutralizes TGFβ1, -β2, -β3, and -β5, and the anti-PDGF antibody neutralizes PDGF-AA, -AB, and -BB. 

The 27P (anti-αPDGFR), 80.8 (anti-αPDGFR), 69.3 (anti-Ras GTP activating protein; RasGAP), and 30A (anti-βPDGFR) are rabbit crude antisera and have been characterized. ^{27 28 29} 4G10 and PY20 are mouse monoclonal anti-phosphotyrosine antibodies, purchased from Upstate Biotechnology, Inc. (Lake Placid, NY) and Transduction Laboratories (Lexington, KY), respectively. For Western blot analysis the following dilutions were used for each of the primary antibodies: anti-αPDGFR, a 1:1 mixture of the 27P and 80.8 antibodies, 1:1000; anti-βPDGFR, 1:5000; anti-phosphotyrosine, 4G10:PY20 (1:1), 1:5000; and 69.3, 1:4000. Secondary antibodies were horseradish peroxidase–conjugated donkey anti-rabbit (catalog no. NA934; Amersham Pharmacia Biotech) or sheep anti-mouse (catalog no. NA931; Amersham Pharmacia Biotech) whole antibodies diluted 1:5000. 

Collagen I Contraction Assay

The contraction assay was as previously described, ³⁰ with slight modifications. Cells were suspended in 1.5 mg/mL neutralized collagen I (Cohesion Vitrogen 100; Invitrogen, Palo Alto, CA) at a density of 10⁶ cells/mL, and were transferred into a 24-well plate (Falcon, Franklin Lakes, NJ) that had been preincubated with a solution of phosphate-buffered saline (PBS) and 5 mg/mL BSA overnight. The gel was solidified by incubating at 37°C for 90 minutes, and then the well was flooded with DMEM and 5 mg/mL BSA, supplemented with buffer or the agent to be tested. The gels were incubated at 37°C with 5% CO₂. The initial gel diameter was 15 mm. The medium was replaced every 24 hours, and the gel diameter was measured after 24, 48, and 72 hours. The extent of contraction was calculated by subtracting the diameter of the well at a given time point from the initial diameter (15 mm). Each experimental condition was assayed in triplicate, and at least three independent experiments were performed. 

Rabbit Vitreous Extraction

Vitreous was collected from freshly isolated normal rabbit eyes by first removing the anterior segment (cornea, iris, and lens), and then the vitreous was squeezed out of the remaining posterior portion of the eye. The extracted vitreous was resuspended in PBS containing 5 mg/mL BSA. The samples were centrifuged at 2500g for 10 minutes at 4°C, and the resultant supernatant was aliquoted and frozen at −70°C until use. Vitreous prepared in this way could include trace amounts of retinal and/or choroidal materials. 

Statistic Analysis

An unpaired t-test was performed to detect statistically significant differences between experimental conditions in the contraction assay. In all cases, P < 0.05 was considered significant. 

Results

Effect of TGFβ in Vitreous on Fibroblast Contraction

We tested the possibility that vitreous promotes cellular contraction. The vitreous was excised from healthy rabbits and added to the culture medium of Fα cells seeded in collagen type I gels. The diameter of the gels was measured at the start of the experiment and 48 hours later. As shown in Figure 1A , vitreous promoted contraction in a dose-dependent manner. 

To identify the agents within vitreous that were responsible for this cellular response we focused on TGFβ, which is present in vitreous and stimulates contraction of collagen type I gels containing fibroblast or RPE cells. ^{31 32} When vitreous was pretreated with 100 μg/mL of neutralizing TGFβ antibody, approximately 60% of the contraction activity disappeared (Fig. 1B) . This dose of neutralizing antibody was sufficient to neutralize at least 10 ng/mL TGFβ1 (data not shown). In contrast, the same amount of a control IgG had no effect on contraction stimulated by the vitreous (Fig. 1B) . 

Response to Vitreous of Cells That Express the αPDGFR

We have observed that cells expressing the αPDGFR induced PVR in a rabbit model of the disease better than cells that express theβ PDGFR or no PDGFRs. ¹³ We related these PVR findings to in vitro contraction by testing the in vitro contraction response of the cell lines used in the PVR studies. As shown in Figure 2 , vitreous triggered contraction in cells expressing the αPDGFR (Fα) more potently than in cells expressing the βPDGFR (Fβ) or no PDGFRs (FCX²). 

Relationship between TGFβ1 and PDGF

In light of the fact that PDGF promotes contraction of fibroblasts, ³³ a likely explanation for the results shown in Figure 2 is that the vitreous contains PDGF family members that activate the αPDGFR but not the βPDGFR. However, pretreating the vitreous with a neutralizing PDGF antibody had no effect on the contractile response, even though the same dose (100 μg/mL) of this antibody completely blocked contraction induced by 10 ng/mL of PDGF-AA or -BB (data not shown). 

Because TGFβ was a major contributor to the contraction activity in vitreous, we tested whether contraction induced by purified TGFβ1 was influenced by expression of PDGFRs. Indeed, we observed an even more pronounced dependence on expression of PDGFRs when using purified TGFβ1 than with vitreous (Fig. 3D) . TGFβ1 triggered robust contraction of Fα cells, whereas the response was modest (Fβ) or undetectable (FCX²) in the other cell lines (Fig. 3D) . All three cell types responded comparably to FBS (Fig. 3B) , indicating that all cell lines had the capacity to contract under these experimental conditions. Furthermore, both of the PDGFR-positive cell lines contracted to a comparable extent after stimulation with PDGF-BB (Fig. 3C) , indicating that each of the receptors are capable of triggering this response. 

We next investigated why TGFβ1 promoted contraction of cells expressing the αPDGFR but not the βPDGFR. This phenomenon did not appear to be due to a differential ability of the two PDGFRs to trigger cell contraction, because both the α- and βPDGFRs drove this response when the receptors were directly activated with PDGF (Fig. 3C) . Consequently, we tested whether TGFβ1 could activate either of the two PDGFRs. To this end, Fα or Fβ cells were stimulated with TGFβ1 and then immunoprecipitated and subjected to an anti-phosphotyrosine Western blot analysis. TGFβ1 triggered tyrosine phosphorylation of the αPDGFR (Fig. 4A) . By comparison to the PDGF-dependent response, TGFβ1-induced tyrosine phosphorylation of the αPDGFR was much slower and less intense. This may reflect the fact that TGFβ1 does not directly activate the αPDGFR, but probably functions by promoting the synthesis and secretion of PDGF. ²⁴ In contrast to theα PDGFR, the βPDGFR was very modestly tyrosine phosphorylated in TGFβ1-treated cells at any of the time points tested (Fig. 4B) . 

Discussion

A novel finding in this study is that vitreous promoted cellular contraction and that TGFβ was the growth factor responsible for most of the activity. Furthermore, TGFβ1-induced contraction was dependent on expression of the αPDGFR. One question that arises from these experiments is which cell type(s) is the source of TGFβ in vitreous? A likely candidate is the hyalocytes, because these cells reside in the cortex of the vitreous. These cells are of the monocyte-macrophage lineage and are capable of secreting TGFβ. ^{34 35} An additional question is how is the TGFβ activated. TGFβ is typically secreted in a latent form and must undergo activation. ³⁶ We did not intentionally activate the TGFβ in the vitreous, and thus it either existed in the activated state in the vitreous of healthy rabbits or underwent activation during preparation or within the assay itself. 

A second question that is brought to light by our findings is how does TGFβ1 make use of the αPDGFR? The receptors for PDGF are not activated by ligands outside the PDGF family, ^{37 38} and hence it is highly improbable that TGFβ1 directly activates the receptor. A more likely scenario is that TGFβ1 stimulates the synthesis and secretion of a PDGF family member. Numerous groups have shown that TGFβ indeed has this capability. ^{24 25 26 39 40 41} However, we were not able to block TGFβ1-dependent contraction using neutralizing antibodies to PDGF (Ikuno and Kazlauskas, unpublished observations, 2000). A caveat of this approach is that, although the antibody used in our experiments blocked all three of the traditional PDGF isoforms (PDGF-AA, -BB, and -AB), its reactivity toward the newly discovered PDGF-CC and PDGF-DD ^{42 43 44} isoforms is unknown. Additional studies, and most probably the development of new reagents, are needed to assess the possible role of new PDGF family members in TGFβ1-dependent cellular contraction. 

We also considered the possibility of a relationship between TGFβ1 and the αPDGFR at the level of signal relay. In this scenario, TGFβ1 would activate the αPDGFR intracellularly. We generated cells expressing an αPDGFR that is missing most of the extracellular domain and thus is unable to bind ligand. These cells failed to contract when exposed to TGFβ1, and this truncated αPDGFR was not tyrosine phosphorylated in TGFβ1-treated cells (Ikuno and Kazlauskas, unpublished observations, 2000). Thus, we were not able to demonstrate intracellular cross talk between TGFβ1 and the αPDGFR. 

One of the key events in PVR is the contraction of the epiretinal membrane and consequent retinal detachment. Our findings suggest thatα PDGFR is an important contributor to this step of the disease, because this receptor is required for the contraction induced by several different growth factors. The use of cell lines that individually express the two receptor for PDGF was critical for this discovery. This is because most cell lines, including the more relevant RPE cell line, express both PDGFRs, ⁴⁵ and none of the PDGF ligands specifically activates the βPDGFR. ^{42 43 44 46 47} Consequently, it is not possible to assess the relative contribution of each of the PDGFRs in cells that have not been modified. Additional studies are needed to determine whether the αPDGFR is particularly important for PVR other animal models and in the clinical setting. 

There are a number of differences between the clinical disease and the rabbit model of PVR that we have used ^{13 17} that are relevant to the findings described herein. RPE cells, not fibroblasts, are the major cell type found in epiretinal membrane isolated from patients with PVR. ^{1 2} Our preliminary studies indicate that cultured RPE cells contract after exposure to vitreous or TGFβ1 (Ikuno and Kazlauskas, unpublished observations, 2000), and thus the RPE and fibroblasts are similar in this regard. Additional studies are needed to determine whether the αPDGFR is the primary mediator of TGFβ1-dependent contraction in the RPE cells, as it is in fibroblasts. 

An additional potentially critical difference between the rabbit model and human disease is the vascularity of the retina. Circulating platelets contain relatively high levels of TGFβ ⁴⁸ and thus may serve as a source of TGFβ in the vascular human retina. In contrast, the rabbit retina is avascular, and TGFβ therefore does not come from this source. Given the potential involvement of TGFβ in PVR, this difference in source of TGFβ may influence susceptibility and/or progression of PVR. 

The in vitro contraction assay may be a simple screen for identifying compounds that have the potential to inhibit fibrotic diseases such as PVR. This is because of the good correlation between in vitro contraction of cells (Figs. 2 3) and their in vivo PVR potential in a rabbit model of the disease. ¹³ Finally, because theα PDGFR is a mediator of cellular responses of several different growth factors, it may be a particularly relevant target for strategies to prevent PVR. 

 Supported by The Schepens Eye Research Institute Ocular Gene Therapy Program and National Institutes of Health Grant EY-012509 (AK).

 Submitted for publication February 21, 2001; revised September 28, 2001; accepted October 2, 2001.

 Commercial relationships policy: N.

 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.

 

Contraction of cells triggered by vitreous was dependent on TGFβ. (A) Contraction of the cells was dependent on the dose of vitreous added. Vitreous was obtained from normal rabbit eyes. DMEM supplemented with 5 mg/mL BSA and the indicated amount of vitreous was added to cultures of Fα cells plated in a collagen type I gel. Fα cells are mouse embryo fibroblasts that express the αPDGFR. The gel diameter was measured at the start of the experiment and after 48 hours, and the extent of contraction was calculated by subtracting the two values. The data are representative of three independent experiments. (B) TGFβ was one of the biologically active agents of the vitreous. Anti-TGFβ or control IgG (100 μg/mL) was added to DMEM supplemented with vitreous (20%) or TGFβ1 (10 ng/mL), and the collagen gel contraction assay was performed. The data are representative data of three independent experiments. Each experimental condition was assayed in triplicate. The data are the mean ± SD. **P < 0.01, compared with the control (control IgG with 20% vitreous).

Figure 2.

 

Expression of the αPDGFR potentiated vitreal-dependent contraction. The collagen gel assay was used to monitor the response of F cells devoid of PDGFRs (FCX²) or F cells expressing either theβ PDGFR (Fβ) or αPDGFR (Fα). DMEM containing 5 mg/mL BSA was supplemented with buffer (−) or vitreous (+) to a final concentration of 20% and added to the indicated cell lines. Contraction was scored at the 48-hour time point. Fα cells responded significantly better than either FCX² (P < 0.01) or Fβ cells (P < 0.01). Fβ responded slightly better than FCX², but the difference was not significant.

Figure 3.

 

TGFβ1-dependent contraction was greatly potentiated by expression of the αPDGFR. Cells expressing no PDGFRs (FCX²), theα PDGFR (Fα), or the βPDGFR (Fβ) were subjected to the collagen gel contraction assay in the presence of buffer (A), 10% FBS (B), 50 ng/mL PDGF-BB (C), or 1 ng/mL TGFβ1 (D). The gel’s diameter was measured after 24, 48, or 72 hours, and the media were replaced every day. All cells responded similarly to buffer or serum, and both of the PDGFR-expressing cells contracted when PDGF was added to the media. Cells expressing theβ PDGFR or no PDGFRs responded poorly to TGFβ1, whereas those expressing the αPDGFR contracted robustly. Each experimental condition was assayed in triplicate. The data are the mean ± SD. The data are representative of three independent experiments.* P < 0.05; **P < 0.01, compared with FCX² cells; †P < 0.05, compared with the Fβ cells.

Figure 4.

 

TGFβ1 induced tyrosine phosphorylation of the PDGFRs. Cells expressing the α- or βPDGFR were serum starved overnight and then left unstimulated (0) or stimulated with 50 ng/mL PDGF-BB or 1 ng/mL TGFβ1 for the indicated times. The cells were lysed, and the αPDGFR (A) or βPDGFR (B) was immunoprecipitated and subjected to anti-phosphotyrosine Western blot analysis (top panels). The membranes were stripped and reprobed with an anti-αPDGFR antibody (A; bottom) or an anti-βPDGFR antibody (B; bottom). The αPDGFR was tyrosine phosphorylated in response to TGFβ1 to a much greater extent than the βPDGFR. The 210- to 140-kDa region (A) and 250- to 160-kDa region (B) of the blots are shown; bar on the right indicates the position of the 205-kDa molecular mass marker.

The authors thank the members of Kazlauskas Laboratory for constructive discussions.