Drs. Loizos and Kussie and Ms LaRiccia own stock or stock options in ImClone Systems Corp.
Systemic Sclerosis Basic Science Studies
Lack of detection of agonist activity by antibodies to platelet-derived growth factor receptor α in a subset of normal and systemic sclerosis patient sera
Version of Record online: 30 MAR 2009
Copyright © 2009 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 60, Issue 4, pages 1145–1151, April 2009
How to Cite
Loizos, N., LaRiccia, L., Weiner, J., Griffith, H., Boin, F., Hummers, L., Wigley, F. and Kussie, P. (2009), Lack of detection of agonist activity by antibodies to platelet-derived growth factor receptor α in a subset of normal and systemic sclerosis patient sera. Arthritis & Rheumatism, 60: 1145–1151. doi: 10.1002/art.24365
- Issue online: 30 MAR 2009
- Version of Record online: 30 MAR 2009
- Manuscript Accepted: 8 DEC 2008
- Manuscript Received: 20 DEC 2007
To investigate whether agonist anti–platelet-derived growth factor receptor α (anti-PDGFRα) antibodies are present in the serum of patients with systemic sclerosis (SSc; scleroderma).
Sera were obtained from healthy subjects and scleroderma patients. An electrochemiluminescence binding assay was performed for detection of serum autoantibodies to PDGFRα, PDGFRβ, epidermal growth factor receptor (EGFR), and colony-stimulating factor receptor 1 (CSFR1). Serum immunoglobulin was purified by protein A/G chromatography. To assess Ig agonist activity, PDGFRα-expressing cells were incubated with pure Ig and the level of receptor phosphorylation determined in an enzyme-linked immunoassay, as well as by Western blotting. Ig agonist activity was also assessed in a mitogenic assay and by MAP kinase activation in a PDGFRα-expressing cell line.
Sera from 34.3% of the healthy subjects and 32.7% of the SSc patients contained detectable autoantibodies to PDGFRα and PDGFRβ, but not EGFR or CSFR1. Purified Ig from these sera was shown to retain PDGFR binding activity and, at 200-1,000 μg/ml, exhibited no agonist activity in a cell-based PDGFRα phosphorylation assay and did not stimulate a mitogenic response or MAP kinase activation in a PDGFRα-expressing cell line. Two purified Ig samples that were unable to bind PDGFRα did exhibit binding activity to a nonglycosylated form of PDGFRα.
Although approximately one-third of sera from scleroderma patients contained detectable autoantibodies to PDGFR, these antibodies were not specific to scleroderma, since they were also detected in a similar percentage of samples from normal subjects. PDGFRα agonist activity was not demonstrated when purified Ig from these sera was tested in cell-based assays.
Systemic sclerosis (SSc; scleroderma) is a rheumatic disease characterized pathologically by the presence of autoimmunity, a unique vasculopathy, and progressive tissue fibrosis (1). The clinical expression of scleroderma is highly variable, ranging from mild skin sclerosis with minimal internal organ involvement to widespread severe skin sclerosis associated with tissue fibrosis and dysfunction of multiple internal organs (2). Fibrosis in scleroderma, as well as in other fibrotic diseases, involves the proliferation of mesenchymal cells and excessive deposition of collagen and other extracellular matrix proteins by these cells, causing scarring and loss of organ function (3). The fibrotic process contributes significantly to the morbidity and mortality of scleroderma.
Fibroblasts from the affected skin of scleroderma patients show increased synthesis of collagen when grown in culture, compared with skin fibroblasts from healthy subjects (4). The overproduction of collagen by scleroderma fibroblasts is associated with an increase in collagen messenger RNA levels (5). Fibroblast-specific autoantibodies have been detected in the serum of scleroderma patients, and in vitro data suggest that these antibodies can induce fibroblast activation, as measured by increases in expression of proinflammatory cytokines (6).
Autoantibodies specific to platelet-derived growth factor receptors (PDGFRs) have been detected in the serum of patients with SSc (7). These autoantibodies displayed agonist activity, as demonstrated by induction of PDGFR phosphorylation in cultured fibroblasts (7). PDGFRs are expressed on cells of mesenchymal origin, and the up-regulation of PDGFRα by transforming growth factor β1 is thought to have an important role in promoting hyperplastic growth of SSc fibroblasts (3). The present study was undertaken to attempt to confirm the presence of PDGFRα-specific autoantibodies in scleroderma and to evaluate their potential agonist activity using a cell-based receptor phosphorylation assay.
PATIENTS AND METHODS
Serum samples and immunoglobulin purification.
Normal human serum was purchased from Bioreclamation (Hicksville, NY). Scleroderma serum samples were obtained from randomly selected banked serum samples from patients attending the Johns Hopkins Scleroderma Center. Patients were considered to have scleroderma if they met the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) criteria (8) or had at least 3 of 5 features of the CREST syndrome (calcinosis, Raynaud's phenomenon, esophageal dysmotility, sclerodactyly, telangiectasias) variant of limited cutaneous SSc (lcSSc). All patients had provided informed consent for participation in the serum bank. Antibodies from serum samples were purified by protein A/G chromatography. IMC-3G3, an anti-PDGFRα antibody capable of blocking PDGF binding to the receptor, was developed at ImClone Systems Corp. Normal human polyclonal IgG was purchased from Meridian Life Sciences (Memphis, TN).
Receptor binding assays.
Assays were developed to detect autoantibodies specific for human PDGFRα, PDGFRβ, colony-stimulating factor receptor 1 (CSFR1), epidermal growth factor receptor (EGFR), and a PDGFRα nonglycosylated form. Detection was dependent on the bivalent antibodies binding both biotin- and ruthenium-conjugated receptors. These bridging assays theoretically detect all classes of immunoglobulins as long as they are functionally multivalent. Each receptor was conjugated with a ruthenium complex (Meso Scale Discovery [MSD], Gaithersburg, MD) that produces light (electrochemiluminescence [ECL]) on application of an electric potential. A separate sample of each receptor was also conjugated with biotin using the EZ-link Sulfo-NHS biotinylation kit (Pierce, Rockford, IL). With the biotin and ruthenium labeling ratios used, immunochemical reactivity for each receptor was retained, as determined by ECL using receptor-specific antibodies.
Biotinylated receptors (0.5 pmoles/well) were added to plates equipped with electrodes and coated with streptavidin (MSD). Wells were blocked with MSD Buffer A in phosphate buffered saline with 0.1% Tween 20. Serum samples were diluted 2-fold in 2% MSD Buffer A with 0.1% Tween 20. Purified Ig samples were diluted in 1% MSD Buffer A with 0.05% Tween 20. Diluted samples (25 μl) were then dispensed into the wells and incubated for 1.5 hours. Ruthenylated receptors were diluted in 1% MSD Buffer A with 0.05% Tween 20 to a concentration of 2 μg/ml, and 25 μl dispensed into the wells. After a 1-hour incubation, the plates were washed, 150 μl of diluted MSD Read Buffer T was added to each well, and the ECL signals were measured on an MSD Sector PR analyzer. IMC-3G3, 2C5, CS1, and C225 (antibodies specific for PDGFRα, PDGFRβ, CSFR1, and EGFR, respectively; ImClone Systems Corp.) served as positive controls for each binding assay. Unlabeled PDGFRα competed with ruthenylated PDGFRα for binding to IMC-3G3 in the PDGFRα binding assay (data not shown).
All of the receptors used in this study were expressed as extracellular domain proteins. The human PDGFRα extracellular domain was expressed in the mouse myeloid NS-0 cell line. The PDGFRα extracellular domain was constructed to contain a C-terminal His tag for purification by affinity chromatography. The PDGFRα extracellular domain prepared in this manner was determined to be properly folded based on the following: first, the PDGFRα extracellular domain was shown to be monodispersed (not aggregated), by size-exclusion chromatography with multi-angle light scattering. Second, the PDGFRα extracellular domain was shown to bind PDGF-AA, by enzyme-linked immunosorbent assay (ELISA) and affinity chromatography. Finally, 2 commercially available antibodies (MAB1264, a mouse monoclonal anti-human PDGFRα antibody [R&D Systems, Minneapolis, MN], and AF-307-NA, a polyclonal goat anti-human PDGFRα antibody [R&D Systems]) were shown to bind this receptor, by ELISA and ECL.
PDGFRα and MAP kinase phosphorylation assays were performed as previously described (9), with the following modifications. Briefly, porcine aortic endothelial cells stably expressing human PDGFRα were seeded in 12-well tissue culture plates (100,000 cells/well; Becton Dickinson, Franklin Lakes, NJ) and allowed to grow overnight. Wells were then rinsed and incubated in serum-free media. After overnight incubation to render cells quiescent, the cells were treated with purified Ig (200–1,000 μg/ml) for 30 minutes at 37°C. Cells were then detached and lysed in 90 μl lysis buffer. Cell lysates (12.5 μl) were analyzed by sandwich ELISA (R&D Systems) to measure phosphorylated PDGFRα. Lysates were also analyzed by Western blotting for phosphorylated MAP kinase and PDGFRα. Human PDGF-AA and PDGF-BB (Austral Biologicals, San Ramon, CA) were used to stimulate receptor phosphorylation. The mitogenic assay was performed as previously described (9).
Forty-nine patients with SSc were included in this study. Forty-six met the ACR criteria for the diagnosis of scleroderma, while the other 3 had at least 3 features of the CREST syndrome. The mean ± SD age at the time of sample collection was 53.2 ± 13.8 years, with a mean ± SD disease duration (from first non-Raynaud's symptom) of 8.35 ± 8 years. Eighty-four percent of the SSc patients were female. Seventy-eight percent were white, 18% were African American, and 4% were Asian American. Fifty-five percent had diffuse cutaneous SSc (dcSSc) and 45% had lcSSc. Serum samples from 10 additional scleroderma patients (5 with dcSSc and 5 with lcSSc; all meeting the ACR criteria) were included in antibody purification studies.
Presence of PDGFR binding activity in some normal and scleroderma serum samples.
Thirty-five serum samples from normal subjects were evaluated for binding activity to PDGFRα and EGFR. A proportion of these samples showed binding activity to PDGFRα but not to EGFR (Figures 1A and B). Among the 49 samples from scleroderma patients, there were also some that bound to PDGFRα but not EGFR (Figures 1C and D). With 2 times the median value set as the cutoff for designation of positive binding activity, 34.3% of the normal sera (12 of 35) and 32.7% of the scleroderma sera (16 of 49) bound to PDGFRα. Some serum samples from normal subjects also exhibited binding to PDGFRβ (11 of 35) but not to CSFR1 (data not shown).
PDGFRα binding activity in normal serum samples attributed to non-agonist antibodies.
The 5 normal serum samples showing the highest binding activity and the 5 showing the lowest binding activity (no binding) were subjected to antibody purification. Purified immunoglobulins from the sera with the highest binding and the sera with the lowest binding (designated groups a and b, respectively) were tested in the PDGFRα binding assay, at 8 mg/ml (Figure 2A). (This concentration was the maximum amount possible given the yield following antibody purification of the 10 serum samples.) Although binding activity was reduced in the purified immunoglobulins compared with serum, possibly due to an antibody concentration of >8 mg/ml in serum, the overall data confirmed that Ig in the serum in the bridge-type assay, and not dimeric PDGFR ligands, was the binding agent. Each of the 10 purified Ig samples from healthy subjects was then tested for the ability to induce PDGFRα phosphorylation in a cell-based assay. At a concentration of 200 μg/ml, none of the purified Ig samples induced receptor phosphorylation (Figure 2B).
PDGFRα used in the above-described binding assay was derived from a mammalian cell expression system and is glycosylated (PDGFRα–NS-0; Figure 2C). PDGFRα expressed in a baculovirus system is not glycosylated, as evidenced by its faster gel migration relative to PDGFRα–NS-0 (Figure 2C). As seen in Figure 2D, 2 of the 5 sera from group b (with no binding activity) did exhibit binding to PDGFRα expressed in a baculovirus system. This suggested that changes in the receptor may allow binding of serum autoantibodies (see below).
PDGFRα binding activity in scleroderma serum samples attributed to non-agonist antibodies.
Ten additional scleroderma samples were subjected to antibody purification. In accordance with the proportion of the initial 49 SSc sera that exhibited binding activity, 4 of the 10 additional SSc sera showed PDGFRα binding activity (Figure 3A). The 4 samples that exhibited binding and the 6 that did not exhibit binding were designated groups c and d, respectively. After antibody purification, 3 of the 4 sera from group c maintained PDGFRα binding activity at 6.3 mg/ml (Figure 3B). (This concentration was the maximum amount possible given the yield following antibody purification of the 10 SSc serum samples.) Similar to findings in the samples from the normal subjects, none of the 10 purified Ig samples from SSc patients showed agonist activity at 200–1,000 μg/ml in a cell-based phosphorylation assay (Figure 3C [Western blotting data not shown]). Purified Ig from SSc sera that were positive and SSc sera that were negative for binding to PDGFRα were also assessed for agonist activity in a mitogenic assay. The purified Ig samples did not induce a mitogenic response when incubated with PDGFRα-expressing cells (Figure 3D). A phosphorylation assay was performed to assess MAP kinase activation after PDGFRα-expressing cells were incubated with purified Ig from SSc sera that were positive and SSc sera that were negative for PDGFRα binding. The purified Ig samples did not induce MAP kinase phosphorylation (Figure 3E).
We have demonstrated that antibodies to PDGFR are present in approximately one-third of patients with scleroderma, but these antibodies are not a specific finding for the disease in that they are also detectable in normal subjects. In addition, these antibodies purified from normal and scleroderma serum were not agonistic at concentrations of 1,000 μg/ml in a cell-based PDGFRα phosphorylation assay. The purified antibodies were also shown to not exert PDGFRα agonist activity in a mitogenic assay and to not induce MAP kinase phosphorylation after incubation with PDGFRα-expressing cells. While we detected antibodies to PDGFRα and PDGFRβ, antibodies against EGFR and CSFR1 were not found using a sensitive ECL binding assay. Our data also demonstrate recognition of a nonglycosylated form of PDGFRα by serum antibodies that are unable to recognize the normally glycosylated receptor.
Whereas our findings showed that anti-PDGFRα antibodies from normal and SSc sera were not agonistic in cell-based assays, it was previously reported that pure Ig from scleroderma patients at 100–300 μg/ml did induce phosphorylation of PDGFRβ (7). The same group of investigators also reported that stimulatory autoantibodies to PDGFR were present in patients with extensive chronic graft-versus-host disease (10). The discrepancy between our findings and theirs might be attributed to the use of different cell lines and differences in receptor levels on the cell lines used. The authors of the previous report also generated clones from immortalized scleroderma lymphocytes. Antibodies from these clones induced the expression of collagen genes and reactive oxygen species from fibroblasts with a higher specific activity than was observed with total scleroderma IgG (7). It would be of interest to test these clonal antibodies in the same cell-based assays used in our study.
The presence of anti-PDGFR antibodies in a significant number of randomly selected normal sera suggests that these antibodies are not disease specific and may not have any relevant biologic consequence in some individuals. Likewise, we were unable to detect anti-PDGFR antibodies in every SSc patient, suggesting that the presence of the antibody is not a uniform explanation for triggering of fibrosis in SSc via the PDGFR.
It may be speculated that the finding of anti-PDGFR antibodies in ∼34% of normal and scleroderma patient sera suggests that any role these antibodies might play in scleroderma has to do with unique changes in PDGFRs. This hypothesis would be an extension of a previous proposal regarding “intrinsic alterations” in scleroderma fibroblasts that allow the fibrotic response to cytokine stimulation (11). PDGFRα is thought to be overexpressed in vivo in scleroderma, since in vitro experiments have shown up-regulation of this receptor by transforming growth factor β1 in scleroderma but not normal skin fibroblasts (12). Elevated cell surface levels of PDGFRs might be more susceptible to potential agonist effects of autoantibodies. In addition to elevated receptor levels, alterations in the receptors might be recognized by autoantibodies. The data presented here demonstrate the recognition of a nonglycosylated form of PDGFRα by serum antibodies that are unable to recognize the normally glycosylated receptor. Future studies might focus on changes in PDGFRα expressed by scleroderma fibroblasts compared with normal fibroblasts that make it susceptible to autoantibody recognition and activation. Although porcine aortic endothelial cells expressing human PDGFRα show a robust mitogenic response to PDGF, they probably lack several of the potential intrinsic properties of SSc fibroblasts.
In conclusion, we did not find positivity for autoantibodies to PDGFRα in every scleroderma patient studied, and we did identify the presence of these autoantibodies in some normal subjects. The anti-PDGFRα fraction within 1,000 μg/ml of purified Ig did not activate receptor in a cell-based assay. The present data therefore do not support the conclusion that autoantibodies to PDGFRα are pathologically important in scleroderma.
Dr. Loizos 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. Loizos, LaRiccia, Boin, Wigley, Kussie.
Acquisition of data. Loizos, LaRiccia, Weiner, Hummers, Wigley.
Analysis and interpretation of data. Loizos, LaRiccia, Boin, Hummers, Wigley, Kussie.
Manuscript preparation. Loizos, Griffith, Boin, Hummers, Wigley, Kussie.
Statistical analysis. Loizos, LaRiccia.
Serum purification and characterization. Griffith.