Systemic sclerosis (SSc) is a severe connective tissue disease of unknown etiology, characterized by fibrosis of the skin and multiple internal organs. Recent findings suggested that the disease is driven by stimulatory autoantibodies to platelet-derived growth factor receptor (PDGFR), which stimulate the production of reactive oxygen species (ROS) and collagen by fibroblasts. These results opened novel avenues of research into the diagnosis and treatment of SSc. The present study was undertaken to confirm the presence of anti-PDGFR antibodies in patients with SSc.
Immunoglobulins from 37 patients with SSc were purified by protein A/G chromatography. PDGFR activation was tested using 4 different sensitive bioassays, i.e., cell proliferation, ROS production, signal transduction, and receptor phosphorylation; the latter was also tested in a separate population of 7 patients with SSc from a different research center.
Purified IgG samples from patients with SSc were positive when tested for antinuclear autoantibodies, but did not specifically activate PDGFRα or PDGFRβ in any of the tests. Cell stimulation with PDGF itself consistently produced a strong signal.
The present results raise questions regarding the existence of agonistic autoantibodies to PDGFR in SSc.
Systemic sclerosis (SSc; scleroderma) is a connective tissue disease characterized by autoimmunity, inflammation, blood vessel damage, and interstitial fibrosis of the skin, lungs, and other organs (1). Two distinct subsets of the disease are commonly distinguished based on skin involvement: diffuse cutaneous SSc and limited cutaneous SSc (2). Serious complications, such as pulmonary arterial hypertension and lung fibrosis, remain major treatment challenges (3). Scleroderma invariably involves fibroblast activation and excessive extracellular matrix production. Several cytokines and their receptors are expressed in SSc lesions and may contribute to fibroblast activation, including transforming growth factor β (TGFβ), connective tissue growth factor, endothelin 1, and platelet-derived growth factor (PDGF) (1, 4). In particular, analysis of mice with experimental scleroderma-like disease indicates that TGFβ has a key role in the development of fibrosis (1, 5). However, these mouse models do not share all of the features of human scleroderma, and results from a phase I/II clinical trial of recombinant anti-TGFβ1 antibody treatment in patients with scleroderma were discouraging (6).
Scleroderma is an autoimmune disorder, as illustrated by the presence of autoantibodies against nuclei (ANAs), centromere (ACAs), topoisomerase I (Scl-70), endothelial cells, and many other self antigens (7, 8). Some of these autoantibodies are useful diagnostic and prognostic markers, but whether they play a role in the pathogenesis of the disease is a matter of debate (7, 8).
A recent report by Baroni et al suggested that fibroblast activation in scleroderma may be caused by stimulatory autoantibodies to the PDGF receptor (PDGFR) α and β subunits, which are members of the type III receptor tyrosine kinase family (9). These autoantibodies have also been found in extensive chronic graft-versus-host disease, which is characterized by fibrotic lesions similar to those observed in scleroderma (10). In contrast to most serum biomarkers of autoimmune disorders, these autoantibodies were described as being fully specific, since they were detected in all patients with scleroderma or extensive chronic graft-versus-host disease, but in none of the study patients with other autoimmune disorders and none of the healthy controls. PDGFR antibodies are thought to induce PDGFR α chain and/or β chain dimerization, mimicking the effect of the natural dimeric ligands. IgG isolated from SSc patients has been shown to induce PDGFR tyrosine phosphorylation, to increase the production of reactive oxygen species (ROS), and to stimulate collagen production by fibroblasts in vitro. Based on these data and on the demonstration that PDGF and TGFβ are important mediators of fibrosis in different models, imatinib mesylate has been suggested as a rational treatment of SSc (11–13). This potent tyrosine kinase inhibitor is selective for PDGFR and Abl, which may be involved in TGFβ signaling and in fibrosis (14).
The above-described findings suggested that autoantibodies to PDGFR might have important implications with regard to the understanding of SSc as well as for its diagnosis and for design of a rational treatment. The present study was conducted to confirm this.
PATIENTS AND METHODS
A total of 44 patients with SSc were studied: 37 from Saint-Luc University Hospital and 7 from Uppsala University Hospital (Table 1). All patients fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for the classification of SSc (15), and their disease was further classified as diffuse cutaneous SSc or limited cutaneous SSc according to the criteria of LeRoy et al (2, 16). The control group comprised 9 patients with systemic lupus erythematosus (SLE) and 17 healthy individuals from Saint-Luc University Hospital, and 5 healthy individuals from Uppsala University Hospital. The study was approved by the local ethics committees (Université Catholique de Louvain and Uppsala University), and informed consent was obtained from all study subjects.
Table 1. Characteristics of the systemic sclerosis patients*
Uppsala University (Sweden) (n = 7)
Université Catholique de Louvain (Belgium) (n = 37)
Except where indicated otherwise, values are the number of patients. PAH = pulmonary arterial hypertension; NSAIDs = nonsteroidal antiinflammatory drugs.
Age, median (range) years
Disease duration, median (range) years
Interstitial lung fibrosis
Immunoglobulins were purified from serum using Ultralink immobilized protein A/G, according to the instructions of the manufacturer (Pierce, Rockford, IL). After extensive washing, elution was performed in the presence of 50 mM glycine (pH 2.8). The samples were quickly neutralized, diluted with phosphate buffered saline (PBS), and concentrated in a Centricon Plus 20 (Millipore, Bedford, MA).
Alternatively, immunoglobulin enrichment was performed by precipitation in the presence of ammonium sulfate, as follows. One volume of saturated ammonium sulfate was added to diluted serum, followed by incubation overnight at 4°C. Precipitates were dissolved in PBS and dialyzed extensively against PBS. IgG samples from Swedish patients were purified using the Melon Gel IgG Spin Purification Kit (Pierce).
Concentrations of all samples were calculated based on absorbance at 280 nm, and IgG purity was confirmed using sodium dodecyl sulfate–polyacrylamide gels. ANAs were detected using HEp-2 cells (Bio-Rad, Richmond, CA). Anti–Scl-70 and ACAs were quantified by EliA (Sweden Diagnostics, Freiburg, Germany). Levels of human PDGF-AB were measured by enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN).
Measurements of PDGFR activity.
After culture of 32D cells in medium supplemented with 10% fetal calf serum (FCS) and interleukin-3 (IL-3) as previously described (17), 32Dα/β cells were generated by sequentially electroporating PDGFRα complementary DNA (cDNA) cloned in the pEF-MYC-Cyto vector, and PDGFRβ cDNA inserted into pEF-BOS-Puro (18). Cells were selected in the presence of Geneticin and Puromycin as described (17). Homogeneous cell populations expressing both receptors were sorted by flow cytometry using specific antibodies against PDGFRα (R&D Systems) and PDGFRβ (19). Parental 32D cells were used as a control. For proliferation assays, cells were washed 3 times with medium and seeded in a 96-well plate at 104 cells per well in medium containing 10% FCS in the presence of PDGF (PeproTech, Rocky Hill, NJ), IL-3 (positive control), or purified IgG. After 20 hours, 3H-thymidine (0.5 μCi) was added to each well. Four hours later, cells were harvested and radioactivity incorporated into DNA was counted using a TopCount platform (PerkinElmer, Zaventem, Belgium) as previously described (20). Results were expressed as a percentage, i.e., ([I – C]/[P – C]) × 100, where I, P, and C represent the average thymidine incorporation in cells stimulated with IgG, PDGF, and control medium, respectively.
For ROS detection, cells (106/ml) were washed 3 times, loaded with dichlorofluorescein diacetate (10 μM; Invitrogen, San Diego, CA) in serum-free medium for 30 minutes at 37°C, and then stimulated with PDGF (25 ng/ml), imatinib mesylate (100 nM; Novartis, Basel, Switzerland), or IgG (0.2–1 mg/ml) for 15 minutes in duplicate. Cells were washed twice with PBS and analyzed by flow cytometry (Becton Dickinson, Mountain View, CA). BJ human fibroblasts immortalized with telomerase (kindly provided by Dr. F. d'Adda di Fagagna, IFOM Foundation, Milan, Italy) were treated similarly after starvation for 24 hours. Cells were trypsinized before analysis by flow cytometry.
Fα mouse embryonic fibroblasts transfected with human PDGFRα (a kind gift from Dr. Andrius Kazlauskas, Harvard Medical School, Boston, MA) were cultured as previously described (9, 21). To remove cells that did not express the receptor, cells were sorted by flow cytometry after staining with anti-PDGFRα antibodies. Porcine aortic endothelial (PAE) cells stably transfected with PDGFRα or PDGFRβ were cultured as described (19, 22). Fα and PAE cells were starved overnight in medium containing 0.1% FCS and incubated for 15 minutes in the presence of PDGF-BB or purified patient IgG (0.3 mg/ml). Cells were lysed in ice-cold radioimmunoprecipitation buffer (1% Triton X-100, 5 mM EDTA, 140 mM NaCl, 50 mM Tris [pH 8], 0.1% sodium deoxycholate, and 10% glycerol) supplemented with 1% Trasylol, 1 mM Pefabloc (Roche, Basel, Switzerland), and 1 mM sodium orthovanadate. After centrifugation, ERK phosphorylation in cell lysates was analyzed by immunoblotting with anti–phospho–T202/Y204–ERK-1/2 antibodies (Cell Signaling Technology, Beverly, MA) and anti–ERK-1/2 antibodies (19). Receptors were immunoprecipitated from lysate with antibodies recognizing PDGFRα or PDGFRβ and immunoblotted against phosphotyrosine (PY99; Santa Cruz Biotechnology, Santa Cruz, CA) or the receptor, as previously described (19, 22).
Features of SSc in the patients included in this study varied in terms of disease type, treatment, complications, and autoantibodies (Table 1). In the majority of patients, the disease was in an active stage.
IgG was purified by standard protein A/G affinity chromatography, which had been used previously to isolate anti-PDGFR autoantibodies (9, 10). As a control for antibody integrity, we compared levels of ANAs, ACAs, and anti–Scl-70 in the serum and in the purified immunoglobulins of patients with scleroderma (8). Results matched perfectly for most patients, indicating that the purification process did not significantly affect the autoantibody profile (data available online at www.icp.be/mexp/pdgf/ssc).
A sensitive bioassay for detecting PDGFR activation was developed using the 32D mouse cell line, which proliferates in the presence of IL-3. Cells were transfected with the human PDGFR α and β subunits and selected to obtain 32Dα/β cells. Cell surface expression of each receptor was tested by flow cytometry (Figure 1A). In accordance with the results of previous studies (23), we observed that 32Dα/β cells proliferated in the presence of PDGF-AA or -BB, as shown in Figure 1B. PDGF-AA binds to PDGFR α-chain only, while PDGF-BB binds to both receptor subunits. A significant signal was consistently observed with PDGF-BB at concentrations of <0.5 ng/ml. Nontranfected 32D cells were used as control. We used this sensitive assay for PDGFR activation to analyze the presence of stimulatory autoantibodies in IgG fractions isolated from the serum of SSc patients from Saint-Luc University Hospital. At a concentration of 0.2 mg/ml, used in the reports that described anti-PDGFR antibodies (9, 10), IgG samples stimulated 32Dα/β cell growth very weakly (Figure 1C). At a higher IgG concentration (0.6 mg/ml), proliferation of 32Dα/β cells was increased, but SSc IgG had no specific effect (Figure 1D). We obtained similar results using antibodies purified by ammonium sulfate precipitation (data available online at www.icp.be/mexp/pdgf/ssc). Overall, no difference between SSc and control samples was observed under any of the experimental conditions tested.
Autoantibodies to PDGFR were initially detected by monitoring the production of cellular ROS, measured after loading mouse fibroblasts with dichlorofluorescein diacetate, which is oxidized into highly fluorescent dichlorofluorescein in the presence of ROS (9). We first observed that PDGF-BB stimulated ROS production in 32Dα/β cells (Figure 2A). This effect could be blocked by imatinib and was absent in nontransfected 32D cells, demonstrating that it was mediated by PDGFR. Incubation of the cells with IgG from SSc patients did not increase ROS production (Figure 2B). We obtained the same results with IgG concentrations of up to 2 mg/ml and cell stimulation for up to 1 hour (Figures 2C and D). We also tested ROS production in human BJ fibroblasts, which express endogenous PDGFR. PDGF weakly but reproducibly increased the ROS content of these cells; again, no effect was observed with incubation of the cells with patient IgG (Figure 2B).
ROS production has been linked to the Ras/MAP kinase pathway, which, according to Baroni et al, is activated by SSc autoantibodies (9). To investigate this, we used Fα mouse embryonic fibroblasts, which express human PDGFRα (data available online at www.icp.be/mexp/pdgf/ssc) and have been shown to respond to SSc IgG (9). We observed a highly variable increase in ERK phosphorylation upon stimulation of Fα cells with any type of purified IgG. To measure average ERK phosphorylation in the presence of immunoglobulins, we incubated fibroblasts with pooled IgG from groups of 4 SSc patients, 4 SLE patients, and 4 healthy controls. As shown by the immunoblotting results presented in Figure 3A, IgG from all sources (SSc, SLE, and controls) had a similar weak stimulatory effect, compared with the effect observed with PDGF.
We next tested whether IgG activated PDGFRα in these cells, by performing antiphosphotyrosine immunoblot analysis on immunoprecipitated PDGFR. Again, a weak but significant increase in PDGFR phosphorylation was observed in cells incubated with IgG. As in the ERK phosphorylation studies, there was no significant difference in stimulatory effect between SSc and control IgG (Figure 3B), suggesting that the effect on ERK and PDGFR phosphorylation was mediated by factors contaminating the IgG preparations, rather than by autoantibodies. PDGF-AB is the most abundant PDGF isoform in serum. Using a specific ELISA, we were able to detect traces of PDGF-AB in our IgG samples, which may partially explain the nonspecific effect of antibody preparations at high concentrations.
Sera were collected from a separate population of 7 SSc patients at Uppsala University Hospital and analyzed at an independent laboratory (Uppsala branch of the Ludwig Institute for Cancer Research). Two sera from the Belgian cohort were also included in these analyses. IgG was purified using Melon Gel (see Patients and Methods) and tested using PAE cells transfected with PDGFR. In this well-characterized model for analysis of PDGFR activation, PDGF induces strong receptor phosphorylation, followed by the activation of numerous signal transduction pathways, chemotaxis, and cell division (19, 22, 24). Addition of SSc IgG (20–600 μg/ml) to these cells for various amounts of time did not induce any detectable increase in PDGFRα or PDGFRβ phosphorylation, as shown by Western blotting (Figure 4). These observations confirmed that purified IgG from SSc patients does not stimulate PDGFRα or PDGFRβ.
Using 4 different methods for assessing PDGFR activation, i.e., receptor phosphorylation, MAP kinase signaling, ROS production, and cell proliferation, we were unable to detect any specific stimulatory activity in purified IgG from patients with scleroderma. In each experiment, PDGF, used as a positive control, produced a strong signal. We used 32D and PAE cells that were highly sensitive to PDGF, due to high levels of expression of transfected human PDGFRα and/or PDGFRβ. Flow cytometry experiments using mouse anti-PDGFR antibodies showed that the receptors were accessible to antibodies at the cell surface (Figure 1A). However, it is conceivable that anti-PDGFR activity is sensitive to cell type. For instance, a cell-specific posttranslational modification of the receptor could affect antibody binding. For this reason, we also tested human fibroblasts expressing endogenous receptors, as well as Fα mouse fibroblasts, which had initially been used to detect activating antibodies to PDGFR (9, 10). We conclude that our IgG preparations did not contain detectable amounts of antibodies able to activate human PDGFR.
The study patients had either diffuse or limited cutaneous SSc. In many of the patients SSc was at an early progressive stage, when pathologic autoantibodies are more likely to be detected; however, patients with SSc at later stages, presenting with complications, were also included. Patients were from different regions, including Western, Northern, and Southern Europe. At the time samples were obtained for this study, many patients had not yet been treated, whereas others had received corticosteroids, cytotoxic drugs, or angiotensin-converting enzyme inhibitors. It is not likely that differences in patient cohorts explain the discrepancy between our findings and previously published results regarding the role of PDGFR in SSc (9).
We studied immunoglobulins that were purified from serum, using the well-established protein A/G affinity purification method, which was also used by Baroni and colleagues (9). By assessing the level of ANAs in the samples produced by this method compared with the corresponding sera, we were able to confirm that IgG activity was not affected by the purification procedure. In addition, we obtained similar negative results with immunoglobulins produced by ammonium sulfate precipitation, which contained more serum contaminants but included IgM. An independent set of sera from 7 SSc patients was purified using Melon Gel technology, which is based on removal of most abundant serum proteins by affinity chromatography. One advantage of this method is that it does not involve an acidic elution step, which may sometimes alter antibody activity. It is unlikely that specific human IgG could not be isolated by at least one of these purification protocols, unless it displays very peculiar biochemical properties.
One-step purification methods do not produce pure antibodies. In Fα and 32Dα/β cells, we observed a nonspecific effect of IgG preparations produced by protein A/G chromatography. This is likely due to contamination of the antibodies with PDGF, which is released in serum in massive amounts by activated platelets, and with other molecules that are able to indirectly activate PDGFRs. This process, known as receptor transactivation, can be triggered by angiotensin II, inflammatory cytokines, drugs, and many other factors (25–28). Such contaminants are difficult to trace because they are highly variable in terms of chemical structure and are active at low concentrations. Of note, levels of PDGF and angiotensin II have been found to be elevated in patients with SSc (29–31).
Recently, Balada et al described antibodies that bind to the cytoplasmic domain of PDGFRα, produced as a tagged recombinant protein (32). These antibodies were found in the serum of healthy individuals. Because these antibodies are not expected to bind to intact cells, they would not be detectable with the assays used in the present study and in the study by Baroni et al (9).
In conclusion, using well-established methods, we were unable to find evidence of stimulatory anti-PDGFR autoantibodies in 2 groups of patients with scleroderma. PDGF ligands and receptors, which are highly expressed in SSc lesions, may play a role in the disease independently of the presence of autoantibodies. However, determining which of the multiple cytokines and receptors expressed in SSc is the best target for therapy will require further careful studies.
Dr. Demoulin had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Henrohn, F. Rorsman, Lennartsson, Wikström, C. Rorsman, Kämpe, Demoulin.
Acquisition of data. Classen, Henrohn, F. Rorsman, Lennartsson, Wikström, C. Rorsman, Lenglez, Franck-Larsson, Tomasi, Demoulin.
Analysis and interpretation of data. Classen, Henrohn, Lennartsson, Wikström, C. Rorsman, Lenglez, Franck-Larsson, Tomasi, Kämpe, Vanthuyne, Houssiau, Demoulin.
Sample collection. Henrohn, F. Rorsman, Lauwerys, Franck-Larsson.
We are very grateful to Dr. Carl-Henrik Heldin (Uppsala University) and Dr. Pedro Buc-Calderon (Université Catholique de Louvain) for helpful discussions, and to Mrs. Hayat Bardani (Saint-Luc University Hospital) for technical assistance. We thank Dr. Andrius Kazlauskas (Harvard Medical School, Boston, MA), Dr. Anabelle Decottignies (de Duve Institute), and Dr. F. d'Adda di Fagagna (IFOM Foundation, Milan, Italy) for generous donation of reagents, and Dr. Stefan Constantinescu (de Duve Institute) for critical reading of the manuscript.