Sera from patients with scleroderma (systemic sclerosis [SSc]) contain anti–endothelial cell antibodies (AECAs) capable of inducing endothelial cell apoptosis. We sought to determine whether SSc sera containing anticentromere antibodies (ACAs) or anti–topoisomerase I antibodies (or, anti–Scl-70 antibodies) contain subsets of AECAs that trigger distinct pathways of apoptosis and gene expression in normal adult human dermal endothelial cells (HDECs).
Adult HDECs were grown to subconfluence and treated with control or SSc patient sera. Apoptosis was investigated by differential interference contrast (DIC) microscopy, microarrays of proapoptotic gene expression, caspase 3 protease activity, and flow cytometry for phosphatidyl serine translocation.
Flow cytometry and DIC microscopy demonstrated that HDECs exposed to SSc sera containing either SSc autoantibody underwent apoptosis at much higher levels than those treated with control sera. While unique gene expression profiles were induced in HDECs by stimulation with SSc sera containing the respective autoantibody, similar patterns of increased gene expression of transcripts for the proapoptotic protease caspase 3 as well as the SSc autoantigen fibrillin 1 were demonstrated. Caspase 3 gene expression correlated with increased protease activity, and targeted inhibition of this protease partly blocked SSc serum–induced apoptosis. Immunohistochemistry studies of serum-stimulated HDECs demonstrated the aberrant expression of fibrillin 1 protein only in apoptotic endothelial cells treated with SSc sera containing AECAs.
There are distinct AECA subsets in the sera of patients with limited SSc (with ACAs) and diffuse SSc (with anti–Scl-70) that induce unique patterns of HDEC gene expression in the setting of apoptosis associated with increased caspase 3 activity and the reexpression of endothelial cell fibrillin 1.
Scleroderma (systemic sclerosis [SSc]) is characterized by circulating autoantibodies to nuclear and nucleolar antigens, an obliterative vasculopathy, and widespread cutaneous and visceral fibrosis. The vascular complications include Raynaud's phenomenon, telangiectasia formation, digital ulceration, pulmonary hypertension, and renal crisis and are characterized by proliferative intimal lesions of the small arteries and widespread capillary dropout (1). SSc is a complex autoimmune disorder in which a variety of genes interact with as yet unknown environmental factors to produce the disease phenotype (2). There is increasing evidence that endothelial cell (EC) apoptosis, which is initiated by immunoglobulin binding, is an early feature of the disease (3, 4), but it is still unclear whether immune alterations precede or follow endothelial changes (5, 6).
In SSc, antinuclear antibodies (ANAs) are present in more than 95% of patients with anticentromere antibodies (ACAs), and anti–topoisomerase I (or, anti–Scl-70) antibodies being those most commonly observed (7). ACAs are typically found in patients with limited skin and internal organ involvement who are more likely to develop severe Raynaud's phenomenon, digital ulcers, and pulmonary hypertension (8). Anti–Scl-70 antibodies occur in SSc patients who are more likely to develop diffuse cutaneous disease and pulmonary fibrosis (9). Recent reports suggest a role of these autoantibodies in vascular disease associated with SSc. ACAs derived from the sera of SSc patients with ischemic loss of digits have been shown to recognize fragments of centromeric proteins (CENP-C), thus, potentially implicating anti–CENP-C antibodies in the pathogenesis of ischemic loss of digits (10). Another study demonstrated reactivity of ACAs with an 18–19-kd epitope from membrane extracts of human umbilical vein endothelial cells (HUVECs) (11) suggesting that ACAs may be among the heterogeneous group of anti–endothelial cell antibodies (AECAs) detected in SSc sera that are capable of binding to and inducing EC pathology.
While anti–Scl-70 antibodies have not been directly implicated in SSc vascular dysfunction, several studies suggest a role of these antibodies in disease development. Studies of twins have shown anti–Scl-70 antibody production in SSc-affected twins but not in unaffected twins, suggesting that anti–Scl-70 may be linked to the development of SSc (12, 13). A recent study demonstrated the direct binding of purified human anti–Scl-70 to the cell surface of fibroblasts, suggesting a role for anti–Scl-70 antibodies in fibroblast dysfunction (14). Finally, IgG titers of anti–Scl-70 antibodies have been shown to fluctuate with disease activity and skin fibrosis, implying a pathogenic role of these autoantibodies in SSc (7).
While the etiology of the generation of SSc-specific autoantibodies is not known, data suggest that autoimmunity may result from posttranslational modifications of autoantigens during apoptosis, which exposes cryptic epitopes as targets for autoantibody generation (15). Apoptosis-specific posttranslational modifications of a subgroup of autoantigens have been described, including proteolysis by caspases and granzyme B (16–18). The presence of autoantibodies to fibrillin 1 (a major component of microfibrils in the extracellular matrix) has also been reported in the sera of most SSc patients (19), and these antibodies appear to be specific for SSc (20). Whether fibrillin 1 antibodies are pathogenic has yet to be proven, but one could hypothesize that distinct AECAs are present in the sera of patients with limited and diffuse SSc that are capable of inducing EC apoptosis (4, 21) and contributing to the respective clinical phenotype.
In this study, we hypothesized that SSc sera containing autoantibodies to centromere or topoisomerase I contain subsets of AECAs that trigger distinct pathways of apoptosis and gene expression in normal adult human dermal endothelial cells (HDECs). Vascular cell apoptosis could, in turn, expose autoantigens to immune surveillance, evoking an autoimmune response and perpetuating autoimmunity to blood vessels in SSc. The data from this study provide evidence that distinct patterns of HDEC gene expression are induced by AECAs from patients with limited and diffuse SSc, respectively, that culminate in apoptosis associated with increased caspase 3 protease activity and aberrant expression of the microfibrillar protein fibrillin 1.
MATERIALS AND METHODS
Endothelial cell cultures.
HDECs (originating from the abdominal skin of a healthy 36-year-old female donor) at cell passage 3 were purchased from Cambrex (Walkersville, MD). Cells were then passaged twice and grown to subconfluence in T75 flasks (Fisher Scientific, Pittsburgh, PA) using Microvascular Endothelium Cell Medium 2 (Cambrex) in a humidified incubator containing 5% CO2/95% air and at a temperature of 37°C. Incubations were conducted with autoimmune and control sera using a concentration of 15% serum.
Autoimmune and control sera.
Sera from female SSc patients with ACAs (n = 6) and anti–Scl-70 antibodies (n = 7), systemic lupus erythematosus (SLE) patients with cardiolipin antibodies and associated thrombosis (n = 3), and age- and sex-matched controls (n = 6) were selected from frozen (–70°C) sera stored at the Division of Rheumatology, University of Texas–Houston Medical School. SSc patients fulfilled the 1980 criteria of the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) for the diagnosis of SSc (22). SLE patients fulfilled the 1982 revised criteria of the ACR for the diagnosis of SLE (23). Clinical data and results of antibody testing are shown in Table 1. The median age at onset of the first non–Raynaud's phenomenon symptoms was 44.5 years in SSc patients with ACAs and 43.0 years in SSc patients with anti–Scl-70 antibodies. ANA titers ranged from 1:640 to 1:20,480 for ACAs and from 1:160 to 1:2,560 for anti–Scl-70 antibodies.
Table 1. Clinical data on scleroderma patients and effects of their sera on HDEC apoptosis and fibrillin 1 expression*
Scleroderma autoantibody group
Clinical features of scleroderma
DLCO, % predicted
Fibrillin 1 expression
HDEC = human dermal endothelial cell; DLCO = diffusing capacity for carbon monoxide; AECAs = anti–endothelial cell antibodies.
Total immunoglobulin was purified from patient sera by affinity chromatography, using the NAb Protein L Spin Chromatography kit (Pierce, Rockford, IL).
Detection of apoptotic cells.
Apoptosis of HDECs stimulated with SSc sera was determined by morphologic and biochemical assessments. Differential interference contrast (DIC) microscopy was performed on living HDECs stimulated with autoimmune (n = 6) and control (n = 3) sera, and characteristic morphologic changes of apoptosis were counted per high-power field from 3 separate experiments using patient sera. Flow cytometry was performed on HDECs (Beckman Coulter flow cytometer; Fullerton, CA) for translocated phosphatidylserine using the Vybrant Apoptosis Assay kit (Molecular Probes, Eugene, OR).
Since previous studies have demonstrated changes in adhesion molecules and the occurrence of apoptosis at 3 hours (maximal at 6 hours), with protein synthesis occurring by 12 hours (4, 24), we used a 2,635-feature complementary DNA microarray, based on the gene set available from Research Genetics (Huntsville, AL), for time-dependent HDEC gene expression studies. Human Universal Reference RNA (Stratagene, La Jolla, CA) was used as a common probe across all the arrays (reference-experimental design). The entire data set and detailed microarray methodology are available for review at the National Center for Biotechnology Information Gene Expression Omnibus (GEO) gene expression and hybridization array data repository (online at http://www.ncbi.nlm.nih.gov/geo/; GEO accession no. GSE2710). Gene Map Annotator and Pathway Profiler (GenMAPP; online at http://www.genmapp.org/) software was used to visualize gene expression data on maps representing biologic pathways and groupings of genes pertinent to apoptosis, angiogenesis, and autoantigens.
To investigate the role of SSc AECAs in modulating gene expression, subsequent experiments were performed with fractionated serum (3 control and 4 SSc patient sera) using a 16,659-feature 70-mer oligoarray from the Qiagen/Operon Array-Ready Oligo Set (version 1.1; Qiagen, Chatsworth, CA) designed from the UniGene Database Builds 119 and 133, and conducted as described previously (25).
Quantitative data regarding the genes of interest (caspase 3 [Hs00263337] and fibrillin 1 [Hs00171191]) were determined by TaqMan techniques, using respective probes and Universal PCR Master Mix (Applied Biosystems, Foster City, CA) for RT-PCR, with analysis on an ABI 7900 instrument (Applied Biosystems).
Global gene expression analysis.
Since there were clusters of genes that were uniquely modulated in HDECs stimulated with SSc sera, MAPPFinder software (available at http://www.genmapp.org) was used to determine the biologic processes containing the largest correlated gene expression changes in the data. These biologic processes are ranked by Z score (see below), which provides a meaning value, thereby identifying those areas of biology that warrant a more detailed examination.
Functional studies of caspase 3.
Caspase 3 protease activity was ascertained in HDECs stimulated with SSc sera containing ACAs (n = 3), SSc sera containing anti–Scl-70 antibodies (n = 3), or SLE sera containing anticardiolipin antibodies (n = 3), using a RediPlate 96 EnzChek Caspase 3 assay kit (Molecular Probes).
HDECs were grown to subconfluence in 24-well glass-bottomed SensoPlates (Greiner Bio-One, Longwood, FL) coated with 0.1% gelatin (Cascade Biologics, Portland, OR) and then exposed for 18 hours to SSc sera containing ACAs (n = 6), SSc sera containing anti–Scl-70 antibodies (n = 7), SLE sera (n = 3), and control sera (n = 6). Serum was then removed, and the HDECs were rinsed 3 times with phosphate buffered saline (PBS) and fixed with 4% paraformaldehyde (pH 7.4; Sigma-Aldrich, St. Louis, MO). Fixed cells were incubated with primary antibody in 2% albumin/PBS for 1 hour (in the case of fibrillin 1, incubation overnight was necessary) and with secondary antibody in 2% albumin/PBS for 30 minutes.
Detection of fibrillin 1.
Fibrillin 1 protein was confirmed by a mouse anti-human fibrillin 1 monoclonal antibody (clone 11C1.3; LabVision, Fremont, CA) at a 1:100 dilution, followed by a goat anti-mouse secondary antibody (highly cross-adsorbed against bovine, goat, rabbit, rat, and human antibodies) conjugated to Alexa 488 (Molecular Probes) at a dilution of 1:200. Immunofluorescence studies were performed with an inverted fluorescence microscope using 4′,6-diamidino-2-phenylindole (DAPI; Molecular Probes) for nuclear counterstaining. Human adult dermal fibroblasts (Cambrex) were simultaneously grown to subconfluence to serve as a positive control for fibrillin 1 protein expression.
Detection of AECAs.
The presence of AECAs was ascertained by staining with chicken anti-human IgG conjugated to Alexa 594 (Molecular Probes) at a dilution of 1:200. Chicken antibodies were highly cross-adsorbed against mouse and goat antibodies to prevent the unlikely reaction of the chicken antibodies with the mouse and goat antibodies used for fibrillin 1 detection. Furthermore, since chicken antibodies lack a classic Fc domain, this would avoid nonspecific labeling of HDEC “Fc” receptors and accurately reflect immunoglobulin binding to HDECs.
The 24-well glass-bottomed SensoPlates used for immunohistochemistry were also used for quantitative fluorescence measurements in a GENios microplate reader (GENios, Raleigh, NC). Nuclear staining by DAPI (excitation/emission 360/465 nm) was used as a proxy for the cell count to normalize fluorescein isothiocyanate–labeled anti–fibrillin 1 fluorescence (excitation/emission 485/535 nm) and Texas Red–labeled AECA fluorescence (excitation/emission 590/630 nm). Fluorescence measurements were normalized to control.
Confocal scanning microscopy of AECAs.
For the confocal microscopy experiments, the available filters were for the 488-nm and 594-nm spectral line. Therefore, HDECs were incubated with 10 nM Sytox (Molecular Probes) for 10 minutes to counterstain the nucleus, and AECAs were detected as described above. Serial sections (1 μm) were captured on an Olympus fluorescence scanning confocal microscope (Olympus, Lake Success, NY) to determine the cellular location of AECAs.
Data were analyzed for significance by Student's paired t-test, assuming equal variance unless the F-test indicated unequal variance. P values less than 0.05 were considered significant. With regard to data analysis of biologic processes by MAPPFinder, the Z score was calculated using the mean and SD of the hypergeometric distribution. The difference in the observed number of genes meeting the criterion and the expected number of genes meeting the criterion was divided by the SD of the observed number of genes, normalizing for the size of the observation. To avoid making assumptions about data conforming to a hypergeometric distribution, assignment of P values to this statistic are not performed.
Induction of apoptosis in living HDECs by SSc serum and association with proapoptotic gene expression.
Sera from controls (n = 3), SSc patients with ACAs (n = 3), and SSc patients with anti–Scl-70 antibodies (n = 3) were examined by DIC microscopy to detect morphologic changes of apoptosis in living HDECs. HDECs stimulated with control sera demonstrated normal morphology and cell function (mitosis) (Figure 1A). HDECs stimulated with SSc sera containing ACAs demonstrated apoptotic blebs/bodies and cells with cytoplasmic vacuoles (Figure 1B), similar to that observed with SSc sera containing anti–Scl-70 antibodies, which additionally resulted in gross distortion of endothelial cell morphology and significant cell dropout (Figure 1C). Figure 1D shows the average morphologic changes of apoptosis counted per high-power field from 3 separate experiments using patient sera. Time-course studies of serum-stimulated HDEC gene expression were then performed with pooled sera and demonstrated that 478 genes were uniquely regulated (>3-fold change) in HDECs stimulated with either type of SSc sera (ACAs or anti–Scl-70) as compared with normal human sera.
MAPPFinder software was then used to provide an unsupervised global view of the biologic processes, where the largest correlated gene expression changes were occurring in HDECs exposed to SSc sera. Examples of biologic processes (with Z scores) that were up-regulated include DNA replication involving centromeric protein E and DNA polymerase (z = 3.2), RAS protein signal transduction (z = 2.3), and induction of apoptosis via death domain receptors (z = 2.2). Apoptosis pathway analysis, using GenMAPP software, illustrated that HDEC exposure to either type of SSc sera led to the up-regulation of caspase 3 and requiem and the down-regulation of survivin, NF-κB (p100), and Bcl-2 genes (pathway not shown). Furthermore, gene expression pertaining to EC activation was down-regulated (vascular cell adhesion molecule 1 [VCAM-1], intercellular adhesion molecule 1 [ICAM-1], and ICAM-2), as were a majority of interferon-inducing genes (except for interleukin-6 [IL-6] and IL-18) (data not shown).
To further validate the microarray results, caspase 3 was selected for verification by RT-PCR and was shown to be overexpressed in HDECs stimulated with SSc sera containing either ACAs (1.65-fold; P = 0.02) or anti–Scl-70 antibodies (1.52-fold; P = 0.002), as compared with control sera, at 1, 6, and 12 hours (data not shown).
Triggering of caspase 3 protease activity in HDECs by SSc serum.
Since caspase 3 gene expression was up-regulated in HDECs stimulated with SSc sera and since the respective protease is a terminal executor of apoptosis, functional studies of caspase 3 were performed in HDECs stimulated with SSc sera (n = 3 per SSc autoantibody type) and control sera (n = 3) (Figure 1E). SLE sera (n = 3) and camptothecin (Sigma-Aldrich), a chemical inducer of caspase 3–mediated apoptosis, served as positive controls for autoimmune serum–mediated and chemically induced HDEC apoptosis, respectively. Relative to HDEC protease activity induced by camptothecin treatment, stimulation with autoimmune sera (SSc and SLE) resulted in 75% caspase 3 activity, as compared with 56% caspase 3 activity occurring with control sera stimulation (P = 0.02–0.003). The caspase 3 inhibitor Ac-DEVD-CHO (Molecular Probes) significantly attenuated caspase 3 activity to control levels under all experimental conditions (P < 0.001).
Involvement of caspase 3 in HDEC apoptosis induced by SSc sera.
HDEC apoptosis was next assessed (in the presence or absence of the caspase 3 inhibitor Ac-DEVD-CHO) by flow cytometry in HDECs treated with control sera, ACA-positive SSc sera, anti–Scl-70–positive SSc sera, and SLE sera (Figures 2A–D; each a representative sample from assays conducted on 3 patient sera per study group). SLE sera and SSc sera containing either ACAs or anti–Scl-70 antibodies resulted in levels of HDEC apoptosis (85%, 86%, and 89%, respectively) that were similar to the level induced by camptothecin (90%), a chemical inducer of caspase 3–mediated apoptosis, and were significantly increased compared with control sera (6%) (P < 0.001) (Figure 2E). As expected, the addition of Ac-DEVD-CHO resulted in a 91% attenuation in camptothecin-induced HDEC apoptosis (P < 0.001). In contrast, the addition of Ac-DEVD-CHO resulted in ∼27% reduction in HDEC apoptosis induced by autoimmune sera (P = 0.028–0.009).
Induction of HDEC apoptosis and fibrillin 1 expression by SSc sera, but not SLE sera.
Cluster software analysis and a TreeView-generated heat map from gene profiling experiments (TreeView software; Stanford University, Stanford, CA) show the qualitative differences in gene expression pertinent to apoptosis and autoantigens (Figure 3). Of the autoantigen genes, a >3-fold increase in fibrillin 1 was consistently observed in HDECs stimulated with SSc sera. Since fibrillin 1 is targeted by autoantibodies in SSc, fibrillin 1 gene expression was selected for verification by RT-PCR and was shown to be overexpressed in HDECs stimulated with SSc sera containing ACAs (1.25-fold; P = 0.01) or anti–Scl-70 antibodies (1.28-fold; P = 0.002) as compared with control sera, at 1, 6, and 12 hours (data not shown).
SSc serum–specific induction of fibrillin 1 expression in HDECs was verified by immunohistochemistry (Figure 4). Fibrillin 1 protein localized to HDECs undergoing apoptosis in response to SSc sera (6 of 6 SSc sera tested) (Figures 4C1 and C2), but was not induced in HDECs exposed to control sera (n = 3) (Figure 4B) or those undergoing apoptosis in response to stimulation with SLE sera (n = 3) (Figures 4D1 and D2) or in response to treatment with camptothecin (data not shown).
Additional control sera (n = 3) and SSc sera (n = 7) were then tested for their ability to induce apoptosis and fibrillin 1 expression. Findings in 5 of the 7 additional SSc sera tested were similar. (Table 1 summarizes the findings for all 13 SSc sera tested.) Figure 5A presents a quantitative assessment of the average fibrillin 1 expression in HDECs stimulated with 3 different sera per experimental condition.
HDEC apoptosis induced by autoimmune sera attributable to presence of AECAs.
Since AECAs have been demonstrated to bind to ECs and induce apoptosis (4, 21), we examined whether AECAs were present in SSc sera associated with HDEC apoptosis. HDECs were treated with SSc sera containing ACAs (n = 6), SSc sera containing anti–Scl-70 antibodies (n = 7), SLE sera containing anticardiolipin antibodies (n = 3), and control sera (n = 6). AECA binding and HDEC apoptosis were associated with 11 of the 13 SSc sera (Figures 4C1 and C2), with all 3 of the SLE sera, but apoptosis was seen less often and was not associated with fibrillin 1 protein expression (Figures 4D1 and D2), and with none of the 6 control sera. Confocal microscopy demonstrated the presence of AECAs in apoptotic blebs/bodies of apoptotic HDECs stimulated with either SSc or SLE sera (Figures 4E1 and E2). The SSc serum samples that tested negative for AECAs but positive for anti–Scl-70 antibodies (n = 2) failed to induce HDEC apoptosis and fibrillin 1 expression (Table 1).
A quantitative assessment of the average AECA binding to HDECs stimulated with 3 different sera per experimental condition is shown in Figure 5A. Subsequent studies with individual fractionated SSc sera (Figure 5B) confirmed the role of the immunoglobulin fraction of SSc sera in mediating the effects of whole SSc sera, and subsequent experiments with the non-Ig fraction of fractionated SSc sera demonstrated the absence of HDEC apoptosis and fibrillin 1 induction when AECAs were absorbed from SSc sera (data not shown, but were similar to respective experiments with control sera). As shown in Figure 5B, distinct patterns of gene expression were demonstrated in HDECs stimulated with SSc sera containing ACAs as compared with SSc sera containing anti–Scl-70 antibodies and were attributable to the respective immunoglobulin fraction (AECAs).
The association of serum ACAs and anti–Scl-70 antibodies with distinct clinical phenotypes and outcomes in SSc patients led us to hypothesize that SSc sera containing autoantibodies to centromere or topoisomerase I contain subsets of AECAs that trigger distinct pathways of apoptosis and gene expression in normal adult HDECs. Indeed, in this study using adult HDECs, the overall gene expression pattern observed in response to stimulation with SSc sera containing ACAs and anti–Scl-70 antibodies was a down-regulation of numerous genes, including those responsible for angiogenesis and a proadhesive EC phenotype (VCAM-1, ICAM-1, ICAM-2) and an up-regulation of genes for apoptosis (e.g., caspase 3) and the SSc autoantigen fibrillin 1. Furthermore, studies with fractionated SSc sera provide additional support for the role of AECAs in SSc serum–mediated HDEC pathology and demonstrated distinct gene expression profiles for serum from patients with limited SSc (ACA positive) and diffuse SSc (anti–Scl-70 positive).
The dysregulation of EC genes related to angiogenesis or involved in pulmonary vessel development, for example, by serum from patients with limited SSc (and ACAs), bring to mind the telangiectasia/digital ulcer formation and pulmonary hypertension associated with limited SSc. Similarly, one could argue that there are distinct AECAs in patients with diffuse SSc (and anti–Scl-70) that dysregulate EC genes, for example, that signal cytotoxic T lymphocytes to dermal antigens or that inhibit angiogenesis leading to the diffuse skin and lung fibrosis seen in patients with diffuse SSc. Investigations using sera from larger cohorts of SSc patients are currently under way to confirm the association of these gene expression profiles with clinical phenotype in SSc.
While AECAs from SSc (24) and SLE (26, 27) patients have been reported to induce EC activation, other studies have demonstrated EC activation in the absence of AECAs (28) or, instead, have demonstrated the induction of apoptosis by AECAs (29). The down-regulation of genes favoring a proadhesive EC phenotype are inconsistent with the findings of Carvalho et al (24) and may reflect differences in the source of ECs used (umbilical cord versus adult skin) Interestingly, IL-6 (interferon-β2) and IL-18 (interferon-α–inducing factor) were up-regulated independently of cellular adhesion molecules, reminiscent of the proinflammatory phenotype triggered in HUVECs exposed to AECAs from vasculitis sera (30). One possible explanation for these findings would be that there are different subsets of AECAs in autoimmune diseases that have a proclivity for the vasculature (SSc, SLE, vasculitis) that can induce a proinflammatory state and/or proadhesive state.
While distinct differences in gene expression could be demonstrated in HDECs stimulated with sera from patients with limited or diffuse SSc, exposure to either serum resulted in apoptosis. Caspase 3 is a terminal executor of programmed cell death, and significant protease activity was demonstrated in HDECs exposed to SSc sera independently of clinical disease or SSc autoantibody status. Caspase 3 inhibition led to partial blockade of the apoptosis-inducing activity of either SSc or SLE sera, suggesting the additional involvement of other caspase or noncaspase pathways in autoimmune sera–mediated programmed cell death. Alternatively, the increased activity of caspase 3 may be a reflection of its role in the cleavage of class 4 substrates, rather than autoimmune serum–mediated EC apoptosis (31). Interestingly, class 4 substrates include topoisomerase I and the 70-kd protein in U1 small nuclear RNP, both of which are autoantigens to which autoantibodies are known to develop in SSc and SLE, respectively.
Our investigation into EC autoantigen expression yielded similar results when HDECs were stimulated with sera from patients with limited or diffuse SSc, namely, the overexpression of fibrillin 1. Since adult HDECs do not express fibrillin 1, this investigation is, to our knowledge, the first to show the induction of fibrillin 1 in adult HDECs stimulated with SSc sera. The gene encoding fibrillin 1 has been implicated in the pathogenesis of human SSc (32, 33), and disease-specific anti–fibrillin 1 antibodies have been reported in the sera of most SSc patients independently of SSc autoantibody status (19, 20). Fibrillin 1 is a 350-kd glycoprotein that is a major constituent of microfibrils in the extracellular matrix (34) and is believed to play a critical role in fibrosis through its regulation of transforming growth factor β (TGFβ) signaling in the extracellular matrix through TGFβ sequestration and activation (35). From a vascular perspective, fibrillins have been demonstrated in anchoring filaments from human lymphatic vessel ECs and transmit chemical/mechanical stimuli from the EC cytoskeleton to the extracellular matrix (36). They are also found in association with type VI collagen in human vascular subendothelium (37) and in association with fibronectin in the filaments extending from ECs in the developing mouse aorta (38).
Our data demonstrating the reexpression of this glycoprotein in adult HDECs exposed to SSc sera are reminiscent of tenascin expression in skin fibroblasts (39). Tenascin is transiently expressed during embryonic development, but is absent from most normal adult tissues. However, in healing wounds and in fibrotic diseases such as SSc, tenascin is reexpressed and is up-regulated in SSc fibroblasts by high levels of IL-4. Similarly, the development of AECA-mediated vasculopathy in SSc may trigger the reexpression of fibrillin 1 in HDECs. The absence of fibrillin 1 expression in HDECs undergoing apoptosis in response to SLE sera or to treatment with camptothecin suggests that fibrillin 1 induction is specific to SSc and is less likely a secondary response of HDECs undergoing programmed cell death. The induction of this protein in apoptotic ECs could unmask cryptic epitopes and perpetuate an immune response (e.g., anti–fibrillin 1 antibody). One could speculate about a role of these anti–fibrillin 1 antibodies (generated during vascular apoptosis in SSc) in binding to the microfibrils of the extracellular matrix and triggering fibrosis through the activation of sequestered TGFβ, as was described recently (40). Such speculation is intriguing, since it might explain why fibrosis in patients with SSc is preceded by vasculopathy (e.g., Raynaud's phenomenon, nailfold capillary dropout, and telangiectasia formation).
Our demonstration of SSc AECA–dependent HDEC apoptosis is consistent with previous investigations demonstrating apoptosis of HUVECs upon exposure to SSc AECAs (4, 21) and the in vivo apoptosis-inducing effects of AECAs in the microvasculature of normal chicken embryos following intravenous injections of AECA-positive sera obtained from UCD-200 chickens (an avian model of SSc) (41). However, a recent study demonstrated that SSc serum alone is incapable of inducing either HDEC or HUVEC apoptosis and that activated natural killer cells are required, in addition to SSc sera, to trigger apoptosis in HDECs via CD95 (42). Possible factors that may explain these differences include differences in EC stimulation media (<5% serum versus 15% serum) or the source of microvascular ECs (foreskin versus abdominal skin).
Because an investigation evaluating the activity of SSc AECAs against ECs from the micro- and macrovasculature demonstrated significantly higher binding to the microvasculature (thus emphasizing the distinct differences in ECs obtained from different tissues ), we used ECs obtained from the abdominal skin of a normal adult female as a substrate for AECA detection. With this method, we have shown, for the first time, the colocalization of AECAs with blebs resulting from autoimmune sera–mediated HDEC apoptosis which, in the case of SSc sera, is additionally associated with fibrillin 1 expression. The inability of control sera and AECA-negative sera (from SSc patients with anti–Scl-70 and minimal vascular disease) to induce apoptosis suggests that pathogenic AECAs are absent in control sera and are distinct from SSc autoantibodies in their ability to induce HDEC apoptosis. This is consistent with a recent study that demonstrated the binding of purified human SSc anti–Scl-70 antibodies to fibroblasts but not to HUVECs or dermal microvascular ECs (14).
In conclusion, the investigations described herein demonstrate that distinct AECA subsets exist in the sera of patients with limited SSc (with ACAs) and patients with diffuse SSc (with anti–Scl-70 antibodies) and can differentially modulate EC gene expression. Despite these serum-specific differences in HDEC gene expression, both types of SSc sera trigger apoptosis and, as identified herein for the first time, are also associated with increased caspase 3 activity and the reexpression of EC fibrillin 1. This pathologic recapitulation of fibrillin 1 expression in the setting of apoptosis may provide autoantigenic substrates that lead to autoimmunity in SSc.
Drs. Robert Lafyatis, Raphael Lemaire, and Russ Widom (Boston University School of Medicine) are acknowledged for their critical reading of the manuscript.