Constitutive Connective Tissue Growth Factor Expression in Scleroderma Fibroblasts Is Dependent on Sp1*

Fibrotic diseases such as scleroderma (systemic scle-rosis, SSc) are characterized by an excessive production of extracellular matrix and profibrotic proteins such as connective tissue growth factor (CTGF). In normal dermal fibroblasts, CTGF is not expressed unless induced by proteins such as tumor growth factor- (cid:1) (TGF (cid:1) ). Con-versely, in fibroblasts cultured from fibrotic lesions CTGF mRNA and protein are constitutively expressed, even in the absence of exogenously added TGF (cid:1) . Thus, studying the mechanism underlying CTGF overexpression in SSc fibroblasts is likely to yield valuable insights into the basis of the fibrotic phenotype of SSc and possibly other scarring disease. CTGF overexpression is mediated primarily by sequences in the CTGF promoter. In this report, we identify the minimal promoter element involved with the overexpression of CTGF in SSc fibroblasts. This element is distinct from the element necessary and sufficient for the induction of CTGF expression by TGF (cid:1) in normal fibroblasts. Within this region is a functional Sp1 binding site. Blocking Sp1 activity reduces the elevated, constitutive levels of CTGF promoter activity and protein expression observed in SSc fibroblasts. Relative to consensus binding sites for known tran- scription factors was purchased and used as described by the manufac-turer (Panomics). To assess the role of phosphorylation in Sp1 binding, nuclear extracts were incubated with and without 1 unit of calf intestinal alkaline phosphatase (Roche) for 1 h at 37 °C before use in an electrophoretic mobility shift assay.

Wound healing requires the de novo synthesis of connective tissue. If this process is not appropriately terminated, excessive matrix deposition occurs, resulting in pathological fibrosis (1,2). Fibrotic diseases are among the largest groups of disease for which there is no known effective therapy, in part because the cause of these diseases remains elusive. Because TGF␤ 1 promotes fibroblast proliferation and matrix synthesis, attention has long been devoted to the potential role of this factor in initiation and maintenance of fibrosis (for reviews, see Refs. 3 and 4). For example, there is a clear correlation between TGF␤ action and the initiation of fibrosis; in acute drug-or surgeryinduced animal models, anti-TGF␤ strategies, such as a neutralizing TGF␤ antibody or overexpression of Smad7 (for reviews, see Refs. 3 and 4), are effective at attenuating the onset and severity of fibrogenesis.
However, the precise role that TGF␤ plays in human fibrotic disease is not entirely clear. In the chronic fibrosing disorder scleroderma (SSc) TGF␤ mRNA is localized to the leading edge of the fibrotic lesion (i.e. to the region of enhanced inflammatory response that is presumably involved with the expansion of the fibrotic response) but not to the lesional area itself (5). Although a general feature of fibroblasts taken from SSc lesions is that they show elevated levels of collagen relative to their normal counterparts, SSc fibroblasts do not show elevated TGF␤ levels or binding activity or enhanced sensitivity to TGF␤ treatment (6 -8). Thus, TGF␤ expression seems to be clearly associated with the initiation of the fibrotic phenotype of SSc; however, the role of this cytokine in the maintenance of the fibrotic phenotype in SSc, or of other fibrotic disease, for that matter, remains unclear.
In contrast to TGF␤, expression of the profibrotic protein CTGF correlates well with the severity of fibrotic phenotype of SSc (9,10). Although not normally expressed in fibroblasts unless induced by TGF␤ or other proteins, such as thrombin, CTGF is constitutively expressed in many fibrotic disorders, such as scleroderma, liver sclerosis, and pulmonary and renal fibrosis (for reviews, see Refs. 11 and 12). CTGF protein induces proliferation, collagen synthesis, and chemotaxis in mesenchymal cells (13)(14)(15). CTGF seems to be associated with matrix deposition. For example, a CTGF response element lies between nucleotides Ϫ376 and ϩ58 of the type I collagen (Col1A2) promoter (16), which includes the TGF␤ response element of this promoter (17). In addition, whereas subcutaneous injection of TGF␤ into neonatal rats results only in a transient fibrotic response, coinjection of CTGF and TGF␤ results in sustained fibrosis (18). This result could possibly be because CTGF can bind TGF␤ and consequently may enhance the activity of TGF␤, at low concentrations, to bind its receptors (19). Collectively, these results suggest that TGF␤ may initiate, but CTGF may sustain, the fibrotic response. Thus, examining the mechanism underlying the constitutive overexpression of CTGF in fibroblasts from fibrotic lesions should provide insights into the molecular mechanism underlying fibrosis.
CTGF expression seems to be controlled primarily at the level of transcription and involves both TGF␤-dependent and -independent mechanisms (20 -24). For example, the TGF␤ induction of the CTGF promoter requires consensus binding motifs for Smad and TEF transcription factors (22)(23)(24). Similarly, the elevated expression of CTGF protein observed in SSc fibroblasts is paralleled by elevated levels of CTGF promoter activity (20,22). Thus, analyzing the relative contribution of TGF␤-dependent and -independent mechanisms to the elevated level of CTGF promoter activity in lesional SSc fibroblasts should yield valuable insights into the molecular basis of the maintenance of the SSc phenotype.
To gain insights into the molecular mechanism underlying the fibrotic phenotype of SSc, in this study, we identify regions of the CTGF promoter necessary for the overexpression of CTGF in lesional, dermal SSc fibroblasts. Our results provide new insights into the molecular mechanism underlying the maintenance of the fibrotic phenotype of the SSc fibroblast.

MATERIALS AND METHODS
Cell Culture, Reporter Assays, Transfections, and Western Analysis-Dermal fibroblasts from SSc lesions and healthy persons were taken from biopsies of age-, sex-, and anatomical site-matched volunteers, after informed consent was obtained. All patients fulfilled the criteria of the American College of Rheumatology for the diagnosis of diffuse SSc. Fibroblasts were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and used between passages 2 and 5 (18,20). Transfections and reporter assays were carried out as described previously (20,22), using 1 g of a reporter construct and 0.25 g of the control CMV␤-galactosidase plasmid (Clontech). For assays in which the role of trans-acting proteins were to be tested, 0.5 g of reporter and 1 g of empty vector or expression vector encoding the protein of interest were used. TGF␤2 was from R&D Systems or Celtrix. Data presented are means Ϯ S.E. Statistical analysis was performed by the Student's unpaired test. p values less than 0.05 were considered statistically significant. For Western blots, cells were cultured and 25 l of media were electrophoresed through a 12% SDS/polyacrylamide gel (Novex) and blotted to nitrocellulose (Bio-Rad). CTGF protein was detected as described previously (20,22). For studies involving mithramycin, 150 nM mithramycin (Sigma) was added 6 h before harvesting. Anti-actin antibody was from Sigma.
Electrophoretic Mobility Shift Assays-An enzyme-linked immunosorbent assay detecting levels of Sp1 binding to a double-stranded oligomer containing a consensus Sp1 site was purchased (Clontech). A commercially available source of Sp1 was purchased (Promega). Fibroblast nuclear extracts were prepared and quantified as described previously (25). Gel shifts were performed with 2-5 g of protein and a 0.5-ng probe (1-5 ϫ 10 4 cpm) as described previously (22). For competition experiments, 100-fold excess unlabeled competitor was added to the reaction mixture before incubation, when required. The SMAD oligomer was described previously (25); AP2 and Sp1 oligomers were purchased (Promega). A gel-shift microarray that contained doublestranded oligomers bearing consensus binding sites for known transcription factors was purchased and used as described by the manufacturer (Panomics). To assess the role of phosphorylation in Sp1 binding, nuclear extracts were incubated with and without 1 unit of calf intestinal alkaline phosphatase (Roche) for 1 h at 37°C before use in an electrophoretic mobility shift assay.

RESULTS
The Region Lying between Nucleotides Ϫ86 and ϩ17 of the CTGF Promoter Is Required for Its Overexpression in SSc Fibroblasts-To identify regions of the CTGF promoter important for its overexpression in SSc fibroblasts, we transfected a series of CTGF promoter deletion constructs (wt, Ϫ244 bp, and Ϫ86 bp) into normal and SSc dermal fibroblasts. All constructs contained abundant CTGF promoter expression, suggesting that the elevated level of CTGF expression observed in SSc fibroblasts was controlled by the first 86 base pairs upstream of the transcription initiation start site (Fig. 1). This region of the CTGF promoter does not contain the TGF␤ response that is necessary and sufficient for TGF␤ to induce CTGF in fibroblasts (22)(23)(24).
The Region between Ϫ86 and ϩ17 of the CTGF Promoter Contains One Functional Sp1 Site-To begin to identify elements of the CTGF promoter lying between Ϫ86 and ϩ17 that might contribute to the elevated level of CTGF expression in SSc fibroblasts, we examined this region of the CTGF promoter for consensus sites of known transcription factors. We noted a TATA box flanked by two consensus binding sites for transcription factor Sp1, one 5Ј to the TATA box (5ЈSp1; Fig. 2A); the other 3Ј to the TATA box (3ЈSp1; Fig. 2A). To determine whether the CTGF promoter's putative Sp1 elements were functional, we first assessed whether these sites could bind Sp1 protein. To do this, we synthesized and radiolabeled a doublestranded oligomer containing both putative Sp1 sites (the sequence shown in Fig. 2A) and performed gel-shift analyses using a commercially available source of Sp1 protein. We found that addition of Sp1 protein to the probe resulted in a shift that could be competed either by a 100-fold molar excess of either unlabeled probe or by an oligomer containing a consensus Sp1 site (Fig. 2B). However, a 100-fold molar excess of oligomers containing either a consensus AP2 site or the SMAD element of the CTGF promoter (22) did not compete for factor binding.
To test which of the putative Sp1 binding sites in the CTGF promoter could bind Sp1, we generated double-stranded oligomers that contained mutations in either the 3ЈSp1 site or the 5ЈSp1 site but were otherwise identical to the gel-shift probe and used them as competitors in our gel-shift assays. We found that a 100-fold molar excess of an oligomer containing a mutated 3ЈSp1 site did not compete for factor binding; however, an oligomer containing a mutation in the 5ЈSp1 site still competed for factor binding (Fig. 2B). Thus, Sp1 or Sp1-like binding activity binds to the consensus Sp1 site downstream from the TATA box in the CTGF promoter but not to the consensus site upstream of the TATA box.
To verify that the Sp1 binding site downstream of the TATA box was functional, we mutated the 5ЈSp1 or the 3ЈSp1 sites in the context of an otherwise wild-type, full-length CTGF pro-  (20) were transfected into dermal fibroblasts cultured from four healthy subjects as well as lesional areas from four subjects with diffuse SSc (n ϭ 12). Constructs were: the wild-type CTGF promoter (WT, containing nucleotides Ϫ805 to ϩ17), Ϫ244 (containing nucleotides Ϫ244 to ϩ17), and Ϫ86 (containing nucleotides Ϫ86 to ϩ17). Expression values (mean Ϯ S.D.), after adjusting for differences in transfection efficiencies among samples using a cotransfected vector (CMV-␤-galactosidase (Clontech)), are shown. moter, and subcloned the resultant fragments in front of the SEAP reporter gene. These constructs (m5ЈSp1 and m3ЈSp1, respectively; Fig. 2C) were transfected into fibroblasts. When we compared the expression levels of these new constructs with that of the wild-type, full-length CTGF promoter/SEAP reporter construct (wt CTGF; Fig. 2C), we found that mutation of the 3ЈSp1 site, but not the 5ЈSp1 site resulted in a decrease of ϳ50% in basal promoter activity. Thus, only the putative Sp1 site downstream from the TATA box contributed to the basal CTGF promoter activity. We then cotransfected NIH 3T3 fibroblasts with either empty expression vector or expression vector encoding Sp1 along with the wt CTGF, m5Ј, and m3Ј promoter/ reporter constructs. We found that Sp1 activated the wild-type reporter and the m5ЈSp1 mutant construct, but not the m3ЈSp1 mutant construct (Fig. 2D). Collectively, these data suggest that the Sp1 recognition motif 3Ј to the TATA box, but not the Sp1 site 5Ј to the TATA box, is functional and contributes significantly to basal activity of the CTGF promoter.
To verify and extend previous results showing that this region of the CTGF promoter did not contain its TGF␤ response element and to verify that Sp1 binding was not required for this response (22)(23)(24), we assessed whether the 3Ј Sp1 site was necessary for the TGF␤-mediated-induction of the CTGF promoter. To do this, we transfected the wild-type (wt CTGF), m3ЈSp1, or m5ЈSp1 mutant constructs into cell lines (Fig. 2C). After an 18-h incubation in serum-free media, cells were incubated for 24 h with or without 25 ng/ml TGF␤2. TGF␤ significantly activated all the CTGF promoter/reporter constructs, indicating that Sp1 is not required for the TGF␤ induction of the CTGF promoter.
Sp1 Is Required for the Elevated Levels of CTGF Expression Observed in Scleroderma Fibroblasts-To assess the contribution of Sp1 to the elevated level of CTGF expression in SSc fibroblasts, we transfected these cells with the full-length, FIG. 2. Sp1 activates the CTGF promoter. A, sequence of the CTGF promoter. Putative Sp1 sites are underlined (5ЈSp1 and 3ЈSp1) and are defined relative to the TATA box (gray). B, the sequence shown in A was synthesized as a double-stranded oligomer, radiolabeled, and used in a gel-shift assay with HeLa nuclear extracts that contained Sp1 binding activity (Promega). Competitor oligonucleotides were used as indicated at 100-fold molar excess. The 5ЈSp1 and 3ЈSp1 competitors are identical to the probe (WT), other than that they contain a mutation to a HindIII (AAGCTT) site in the appropriate putative binding site. Oligomers bearing consensus binding sites for known transcription factor binding sites (Ap2 and Sp1) were from Promega. Mutation of the 3ЈSp1 site abolishes the ability of the competitor oligomer to compete for Sp1 binding to the labeled probe. C, effect of mutating putative Sp1 sites on basal and TGF␤-induced CTGF promoter activity. The wild-type CTGF promoter/SEAP reporter construct (wtCTGF, containing nucleotides Ϫ805 to ϩ17 (20)) or otherwise identical constructs bearing mutations in either 5ЈSp1 or 3ЈSp1 site of the CTGF promoter (see "Materials and Methods") were transfected into NIH 3T3 cells. After an 18-h incubation in serum-free media, TGF␤2 (25 ng/ml) was added for a further 24 h. Mutating the Sp1 site 3Ј to the TATA box (m3ЈSp1) results in an equivalent reduction of basal promoter activity and TGF␤ induction of CTGF; that is, mutation of the functional Sp1 site has no impact on the fold-induction of the promoter by TGF␤. All experiments were performed in 6-well plates. Cells were cotransfected with a CMV promoter-driven vector encoding ␤-galactosidase (0.25 g) to control for differences in transfection efficiency. Values expressed are mean Ϯ S.E. (n ϭ 6). D, Sp1 transactivates the CTGF promoter. A wild-type CTGF promoter/SEAP reporter construct (wtCTGF (20)) or otherwise identical constructs bearing mutations in the Sp1 sites within the CTGF promoter at positions 5Ј (m5ЈSP1) or 3Ј (m3ЈSP1) of the TATA box (see Fig. 1) were transfected into NIH 3T3 cells. Cells were cotransfected with empty expression vector or vector encoding Sp1, as shown. All experiments were performed in 6-well plates. Cells were cotransfected with a CMV promoter-driven vector encoding ␤-galactosidase (0.25 g) to control for differences in transfection efficiency. Values expressed are mean Ϯ S.D. (n ϭ 6).
wild-type CTGF promoter/reporter construct or with an otherwise identical construct with a mutation in its functional 3ЈSp1 binding site (Fig. 3). Mutation of the 3ЈSp1 site resulted in a 50% decrease of basal CTGF promoter activity in SSc fibroblasts (Fig. 3). Similarly, incubation of cells with the Sp1specific inhibitor mithramycin (26) reduced elevated CTGF promoter activity (Fig. 3) and protein expression (Fig. 4). Mithramycin also blocked the TGF␤-mediated induction of CTGF protein in normal fibroblasts, reflecting the fact that Sp1 is essential for basal CTGF promoter activity in fibroblasts. Collectively, our results suggest that the constitutive overexpression of CTGF in SSc fibroblasts was Sp1-dependent.
Elevated Levels of Sp1 Binding Exist in Nuclear Extracts from SSc Fibroblasts-We then assessed whether the requirement for Sp1 for the elevated levels of CTGF observed in SSc fibroblasts was mirrored by an elevated level of Sp1 binding to its binding site. We addressed this issue by evaluating the total level of cellular Sp1 capable of binding an Sp1 binding element. We compared nuclear protein from fibroblasts from four healthy subjects with fibroblasts with five subjects with SSc using two related methods. First, we used an enzyme-linked immunosorbent assay that measures Sp1 binding to an Sp1 response element (Fig. 5A). Second, we used a direct gel-shift assay with the same Sp1 binding sequence as probe for consistency. According to either assay, subjects with SSc possessed significantly elevated levels of Sp1 binding relative to healthy subjects (Fig. 5B). As a control, we pooled our normal and SSc nuclear extracts and hybridized equal proteins amounts to a gel-shift microarray, which contained double-stranded oligomers bearing consensus DNA binding sites of known transcription factors. We found that, in contrast to our results examining Sp1 binding, nuclear extracts from normal and SSc fibroblasts contained equal amounts of binding to a consensus oligomer containing a binding site for the nuclear factor of activated T cells (Fig. 5C). To assess whether phosphorylation of Sp1 played a role in the Sp1 binding in fibroblasts, we preincubated nuclear extracts from normal and SSc fibroblasts with calf intestinal alkaline phosphatase before use in EMSAs. We found that removal of phosphates enhanced binding of Sp1 to DNA, suggesting that in fibroblasts, phosphorylation of Sp1 normally suppresses Sp1 binding to DNA (Fig. 5D). Thus, our results suggest that the elevated level of CTGF expression observed in SSc fibroblasts is dependent, at least in part, on an elevated level of Sp1 binding present in the nuclei of SSc fibroblasts and is relatively independent of sequences involved with the TGF␤ induction of CTGF. These results are consistent with the notion that the fibrotic phenotype of the SSc fibroblasts is not caused solely by hyperactive or autocrine TGF␤ signaling. DISCUSSION Previously, we identified the CTGF promoter's TGF␤ response element, which is both necessary and sufficient to confer TGF␤ responsiveness to a heterologous promoter (22)(23)(24). However, this TGF␤ response element is dispensable for the elevated, constitutive level of CTGF expression that is the hallmark of the fibrotic phenotype of the SSc fibroblast; removal of this element has no significant impact on the elevated level of CTGF promoter activity in SSc fibroblasts (22, current study). Here, we have found that Sp1 is involved in CTGF promoter activity. Targeting Sp1 markedly reduced CTGF expression in SSc fibroblasts. Furthermore, nuclear extracts from SSc fibroblasts possessed significantly elevated levels of Sp1 binding to DNA. These results suggest that targeting mechanisms involved with Sp1 or its activation may provide useful in developing antifibrotic therapies.
Our results are intriguing in light of recent observations concerning the role of Sp1 in matrix gene regulation. Functional Sp1 binding motifs have been found in the minimal promoters of several collagen genes. For example, mithramycin inhibits collagen ␣1 type I protein expression in normal fibroblasts (28), and Sp1 has been shown to activate expression of this gene's promoter (29). Furthermore, analysis of the collagen ␣2 type I promoter has shown that TGF␤ activates its expression through a complex process involving SMAD3 and -4 and Sp1 (30). In addition, Sp1 was shown to directly contribute to the expression of a wide variety of matrix genes (31). However, none of these studies examined the role of Sp1 and in the regulation of fibrogenic genes in a fibrotic setting. Intriguingly, Sp1 or its family members may contribute directly to the elevated level of expression of target genes in fibrosis. Intriguingly, increased phosphorylation of Sp1 has been shown to be a feature of scleroderma fibroblasts (32). However, the relevance of the phosphorylated residues to Sp1 binding or the elevation of matrix gene expression in SSc fibroblasts is not clear. In this study, we found that phosphorylation of Sp1 had little effect on the ability of Sp1 to bind to its consensus binding element; addition of alkaline phosphatase to nuclear extracts before performing a gel shift with an oligomer bearing a consensus FIG. 3. Role of Sp1 in elevated CTGF promoter activity in dermal, lesional scleroderma fibroblasts. CTGF promoter/SEAP reporter constructs used contained either the wtCTGF (Fig. 3) or an otherwise identical construct containing a mutated SMAD site (⌬SMAD (22)) or a mutated 3ЈSp1 site (m3ЈSp1; Fig. 1). CTGF promoter/reporter constructs (0.5 g/well) alone with expression vector (as indicated; 1 g/well) were transfected into fibroblasts cultured from dermal scleroderma lesions. The construct containing a mutation in the 3ЈSp1 site possesses 50% less activity than the full-length, wild-type construct. Administration of 150 nM mithramycin also caused a reduction in CTGF promoter activity.
FIG. 4. Effect of mithramycin on CTGF expression in dermal fibroblasts. After culturing cells to 85% confluence, media were changed, and the Sp1 inhibitor mithramycin (150 nM) was added for 24 h to normal dermal fibroblasts, in the presence or absence of TGF␤1 (2.5 ng/ml), and to fibroblasts cultured from dermal lesions of patients with diffuse scleroderma. Cell layers were then lysed and equal amounts of protein (25 g) were subjected to SDS/PAGE and Western blot analysis with an anti-CTGF antibody. Addition of mithramycin blocks the basal CTGF expression in scleroderma fibroblasts. Blots were also probed with an anti-actin antibody to establish that lanes were equally loaded.
Sp1 element enhanced Sp1 binding to DNA. Our results are, in fact, consistent with data from other systems showing that Sp1 binding is repressed by phosphorylation and that removal of phosphate groups from Sp1 enhances binding to DNA (33)(34)(35)(36)(37)(38)(39). However, it is possible that phosphorylation of Sp1 or Sp1-like proteins could enhance the ability of Sp1 to form an active transcriptional complex with other nuclear factors. That said, our report is the first to directly assess the potential, functional contribution of Sp1 to the fibrotic phenotype of the SSc fibroblast.
The elevated level of Sp1 binding observed in our studies could arise from an increased amounts of Sp1 protein in SSc nuclear extracts compared with extracts prepared from healthy subjects. However, when we directly addressed this issue by using Western blot analysis of nuclear extracts with an anti-Sp1 antibody, we could show no appreciable differences in Sp1 protein levels between nuclear extracts prepared from healthy subjects and those prepared from persons afflicted with SSc (data not shown). Sp1 is a member of the Kruppel-like family of proteins that effectively bind to a consensus Sp1 binding site (40). Thus, it is possible that the elevated Sp1 binding activity observed in SSc nuclear extracts could be caused by other members of the Kruppel-like Sp1 family and that these members show elevated expression in lesional diffuse SSc fibroblasts relative to normal fibroblasts. In fact, we have shown that other Kruppel-like family members are capable of activating the CTGF promoter (data not shown). The precise identity of the proteins binding to the Sp1 binding site in the CTGF promoter, and whether expression of these proteins is elevated in SSc, is currently under investigation.
In conclusion, by investigating the constitutive CTGF expression from defined fibrotic settings, we have shown that the TGF␤ response element in the CTGF promoter is dispensable for its constitutive up-regulation in SSc fibroblasts; rather, it seems to reflect an elevation of basal promoter activity as visualized by its dependence on Sp1. Sp1 generally acts with other transcription factors to potentiate transcription (for review, see Ref. 40), so it is likely that Sp1 acts with these factors to activate transcription of genes, such as CTGF, expressed in SSc fibroblasts. Identification of these additional factors is currently underway. That said, our results suggest that a simple model of autocrine TGF␤ signaling seems to be insufficient to explain the fibrotic phenotype of the SSc fibroblast. For example, prolonged exposure of fibroblasts to TGF␤ is insufficient to generate fibroblasts that constitutively overexpress collagen relative to their normal counterparts (8). Because CTGF has been reported to be an effective marker of fibrosis, and given the known profibrotic effects of this molecule and that small molecule inhibitors that suppress CTGF expression alleviate fibrotic effects in vivo and in vitro (41,42), further identification of key transcription factors and signaling pathways involved in regulating CTGF promoter activity in SSc is highly likely to have a significant impact on the development of novel therapeutic agents that could combat this debilitating disorder. FIG. 5. Elevated levels of Sp1 binding to DNA exists in nuclear extracts from fibroblasts cultured from dermal lesions of SSc patients. A, an enzyme-linked immunosorbent assay, which detects binding of Sp1 or Sp1-like binding activity to a double-stranded oligonucleotide bearing a consensus Sp1 binding motif, was used to detect relative levels of Sp1 binding activities in nuclear extracts prepared from dermal fibroblasts cultured from four healthy subjects as well as five subjects with diffuse SSc (mean Ϯ S.D., n ϭ 12). B, a gel shift using a doublestranded oligonucleotide identical to that used in A was used to provide a visual representation of the data in A. C, a microarray containing double-stranded oligomer bearing consensus binding elements of known transcription factors was probed using nuclear extracts prepared from healthy subjects and subjects with diffuse SSc. Elevated binding of nuclear factors to a consensus Sp1 binding element, but not a consensus nuclear factor of activated T cells (NFAT) binding element, is shown. D, phosphorylation of Sp1 or Sp1-like binding activity suppresses Sp1 binding to DNA. Nuclear extracts from healthy subjects and subjects with SSc were pretreated for 1 h at 37°C with calf intestinal alkaline phosphatase (CIP) before performing an electrophoretic mobility shift assay with a consensus Sp1 binding site. Dephosphorylation of extracts eliminates the faster mobility DNA/protein complexes, indicating that these complexes represent the phosphorylated form of Sp1 binding to DNA. However, the overall binding of Sp1 to its DNA binding site increases.