Hypoxia-induced increase in the production of extracellular matrix proteins in systemic sclerosis.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 56, No. 12, December 2007, pp 4203–4215 DOI 10.1002/art.23074 © 2007, American College of Rheumatology Hypoxia-Induced Increase in the Production of Extracellular Matrix Proteins in Systemic Sclerosis Jörg H. W. Distler,1 Astrid Jüngel,2 Margarita Pileckyte,3 Jochen Zwerina,4 Beat A. Michel,2 Renate E. Gay,2 Otylia Kowal-Bielecka,5 Marco Matucci-Cerinic,6 Georg Schett,4 Hugo H. Marti,7 Steffen Gay,2 and Oliver Distler2 Objective. Insufficient angiogenesis with tissue ischemia and accumulation of extracellular matrix are hallmarks of systemic sclerosis (SSc). Based on the severely decreased oxygen levels in the skin of patients with SSc, we aimed to investigate the role of hypoxia in the pathogenesis of SSc. Methods. Subtractive hybridization was used to compare gene expression in dermal fibroblasts under hypoxic and normoxic conditions. Dermal fibroblasts were further characterized by exposure to different concentrations of oxygen and for different time periods as well as by interference with hypoxia-inducible factor 1␣ (HIF-1␣). The systemic normobaric hypoxia model in mice was used for in vivo analyses. Results. Several extracellular matrix proteins and genes involved in extracellular matrix turnover, such as thrombospondin 1, pro␣2(I) collagen, fibronectin 1, insulin-like growth factor binding protein 3, and transforming growth factor ␤–induced protein, were induced by hypoxia in SSc and healthy dermal fibroblasts. The induction of these genes was time- and dose-dependent. Experiments with HIF-1␣–knockout mouse embryonic fibroblasts, deferoxamine/cobalt ions as chemical stabilizers of HIF-1␣, and HIF-1␣ small interfering RNA consistently showed that extracellular matrix genes are induced in dermal fibroblasts by HIF-1␣–dependent, as well as HIF-1␣–independent, mechanisms. Using the systemic normobaric hypoxia mouse model, we demonstrated that dermal hypoxia leads to the induction of the identified extracellular matrix genes in vivo after both short exposure and prolonged exposure to hypoxia. Conclusion. These data show that hypoxia contributes directly to the progression of fibrosis in patients with SSc by increasing the release of major extracellular matrix proteins. Targeting of hypoxia pathways might therefore be of therapeutic value in patients with SSc. The cellular signaling pathways involved in the response to hypoxia have been elucidated in detail during the last several years (1). Many downstream effects of hypoxia are mediated via stabilization of the transcription factor hypoxia-inducible factor 1␣ (HIF1␣) (2). Under normoxic conditions, hydroxylation and acetylation of the oxygen-dependent domain of HIF-1␣ promote binding of the von Hippel-Lindau tumor suppressor protein (pVHL) and rapid degradation in proteasomes. With decreasing levels of oxygen, hydroxylation and acetylation of HIF-1␣ do not occur due to the lack of molecular oxygen, pVHL does not bind, and the HIF-1␣ protein is stabilized. After translocation into the nucleus, HIF-1␣ binds with its dimerization partner HIF-1␤/aryl hydrocarbon receptor nuclear translocator (ARNT) to defined hypoxia-responsive elements in regulatory regions of target genes, such as vascular endothelial growth factor (VEGF), and increases their transcription. While HIF-1␣ is an important mediator of 1 Jörg H. W. Distler, MD: Center of Experimental Rheumatology and Zurich Center of Integrative Human Physiology, University Hospital Zurich, Zurich, Switzerland, and University of Erlangen– Nuremberg, Erlangen, Germany; 2Astrid Jüngel, PhD, Beat A. Michel, MD, Renate E. Gay, MD, Steffen Gay, MD, Oliver Distler, MD: Center of Experimental Rheumatology and Zurich Center of Integrative Human Physiology, University Hospital Zurich, Zurich, Switzerland; 3Margarita Pileckyte, MD, PhD: Kaunas Medical University Hospital, Kaunas, Lithuania; 4Jochen Zwerina, MD, Georg Schett, MD: University of Erlangen–Nuremberg, Erlangen, Germany; 5Otylia Kowal-Bielecka, MD: Medical University of Bialystok, Bialystok, Poland; 6Marco Matucci-Cerinic, MD: University of Florence, Florence, Italy; 7Hugo H. Marti, MD: University of Heidelberg, Heidelberg, Germany. Address correspondence and reprint requests to Oliver Distler, MD, Center of Experimental Rheumatology and Zurich Center of Integrative Human Physiology, University Hospital Zurich, Gloriastrasse 25, CH-8091 Zurich, Switzerland. E-mail: Oliver.Distler@ usz.ch. Submitted for publication April 22, 2007; accepted in revised form August 13, 2007. 4203 4204 DISTLER ET AL hypoxia signaling, HIF-1␣–independent mechanisms, such as messenger RNA (mRNA) stabilization and increased transcription by other HIF family members, also contribute to the cellular responses to hypoxia (1,2). Systemic sclerosis (SSc) is a chronic fibrotic disorder of unknown cause that affects the skin and a variety of internal organs (3). The hallmark of SSc is an excessive accumulation of extracellular matrix proteins, which are released by activated interstitial fibroblasts. Extracellular matrix proteins that are increased in the skin include collagens, fibronectin, glycosaminoglycans, and thrombospondin (4). The resulting progressive fibrosis of the skin and involved organs is a major cause of morbidity and mortality in SSc patients (5). The mechanisms leading to the activation of interstitial fibroblasts are incompletely understood, despite their central role in the pathogenesis of SSc. Abnormalities of the microvascular system are another key feature of SSc. Endothelial cell damage is among the earliest changes in the disease and results in disorganization of the capillary architecture and loss of capillaries. The reduction of capillary density is associated with an insufficient formation of new vessels via angiogenesis (6). Together with the increased distance to blood vessels caused by the accumulation of extracellular matrix proteins, the decreased capillary density reduces the supply of oxygen and leads to tissue hypoxia. We recently showed that patients with SSc have severely reduced levels of oxygen in fibrotic skin as compared with the skin of healthy volunteers (7). However, the molecular effects of hypoxia in SSc and their role in the fibrotic process in vivo have not been analyzed. Thus, the aim of the present study was to identify and characterize genes regulated by hypoxia, using subtractive hybridization as a screening technique, and to address the induction of these genes in a systemic normobaric hypoxia model in mice. MATERIALS AND METHODS Patients and fibroblast cultures. SSc and control fibroblasts were derived from biopsies of affected skin obtained from SSc patients or normal skin obtained from healthy volunteers. Fibroblasts were obtained from biopsy samples by enzymatic digestion, as described previously (8). All SSc patients fulfilled the criteria for SSc as suggested by LeRoy and Medsger (9). Fibroblasts from passages 3–8 were used for the experiments. All patients and controls signed a consent form that had been approved by the local Institutional Review Boards. Mouse embryonic fibroblasts (MEFs) from HIF1␣⫹/⫹ (wild-type) and HIF-1␣–/– (knockout) mice were kindly provided by R. Johnson (10). Culture conditions and induction of hypoxia. Fibroblasts were grown in Dulbecco’s modified Eagle’s medium (Life Technologies, Basel, Switzerland) as described elsewhere (8). For exposure to hypoxia, fibroblasts were transferred into an incubator (Forma Scientific, Illkirch, France) and exposed to a humidified atmosphere containing 5% CO2 and between 1% and 16% O2 volume/volume (hypoxia) as indicated below (7). For controls, cells were cultured under the same conditions except that the atmosphere contained 20% O2 v/v (normoxia). For all experiments, cells were used when they reached 30–50% confluence. In another set of experiments, SSc and normal dermal fibroblasts were grown to 50% confluence in 12-well plates. Deferoxamine mesylate and cobalt chloride (both from Sigma, Deisenhofen, Germany), dissolved in distilled water, were added to the medium for 12 hours at a final concentration of 100 g/ml. Western blotting. Cultured cells were removed from the incubator and rinsed immediately with ice-cold phosphate buffered saline. Extraction of nuclear proteins and transfer onto nitrocellulose membranes was performed according to standard protocols (11). HIF-1␣ protein was detected using monoclonal mouse anti–HIF-1␣ mgc3 antibodies (12), as described elsewhere (7). Suppressive subtractive hybridization. SSc fibroblasts exposed to hypoxic or normoxic conditions for 24 hours were used for these experiments. Suppressive subtractive hybridization was performed using the PCR-Select system (Clontech, Palo Alto, CA). The oxygen concentration under the hypoxic condition (1% O2) is equivalent to a PO2 value of 7 mm Hg, which is close to the 10% percentile measured in the fibrotic skin of SSc patients (6). Isolation of total RNA was performed with TRIzol LS reagent (Gibco BRL, Basel, Switzerland). After analysis of the RNA quality on agarose gel, 1 g of total RNA from SSc dermal fibroblasts exposed to hypoxia for 24 hours or from normoxic controls were reverse transcribed into hypoxic and normoxic complementary DNA (cDNA) pools using the SMART cDNA synthesis kit, which ensures full-length transcription of mRNA (Clontech). Suppressive subtractive hybridization was then performed. Briefly, cDNA from an SSc fibroblast culture exposed to hypoxic conditions for 24 hours was used as tester. After creation of blunt-ended fragments by digestion with Rsa I, tester cDNA was separated into 2 pools, which were ligated to 2 different adaptors. The cDNA pool from normoxic fibroblasts digested with Rsa I, but not ligated to any adaptor, was used as a driver. Hybridization of tester cDNA from hypoxic fibroblasts with an excess of driver cDNA from normoxic controls leads to equalization and enrichment of differentially expressed sequences among tester cDNA molecules, since only cDNA molecules with 2 different adaptors at both ends are amplified exponentially, whereas cDNA molecules present in both hypoxic and normoxic fibroblasts are not amplified. After 2 hybridization steps, hypoxia-induced sequences were further amplified by nested polymerase chain reaction (PCR) using primers against sequences of the 2 different adaptors. For creation of a subtractive cDNA library, the PCR mixture enriched for differentially expressed genes was digested with Rsa I to cleave the adaptors from the cDNA strands. After purification with the High Pure PCR Product Purification kit (Boehringer, Mannheim, Germany), cDNA HYPOXIA-INDUCED EXTRACELLULAR MATRIX PROTEINS IN SSc fragments were ligated into the pPCR-Script Amp SK(⫹) vector (Stratagene, Basel, Switzerland) with T4-ligase at 16°C for 10 hours. Vectors were then amplified in Epicurian Coli XL10-Gold ultracompetent cells (Stratagene). Plasmid DNA was prepared with the Concert High Purity Plasmid Miniprep system (Gibco BRL). Isolated plasmids containing sequences from the library of hypoxia-induced genes were analyzed by automatic dideoxy-sequencing (Microsynth, Balgach, Switzerland). For each gene, the homology with published sequences was analyzed by searching GenBank databases. From the cDNA library of hypoxia-induced genes obtained by subtractive hybridization, 58 clones were randomly selected, sequenced, and identified using National Center for Biotechnology Information BLAST databases. Because similar to other differential expression screening techniques, suppression subtractive hybridization produces false-positive results, the differential expression of the identified genes was confirmed and quantified by real-time PCR using SYBR Green. By using this strategy, the induction of 48% of the identified genes (28 clones) could be confirmed. This percentage is well within the range of the percentages from other studies using suppressive subtractive hybridization and confirms the validity of the experimental approach (13). Quantitative real-time PCR. TRIzol LS reagent was used for RNA isolation from tissues and cultured cells. Tissue samples were homogenized with the Dispergierstation T8.10 (IKA Labortechnik, Wohlen, Switzerland). Tissue samples from the mouse experiments were homogenized with the Dispergierstation T8.10, and TRIzol LS reagent was used for RNA isolation of tissues and cultured cells. For quantification of mRNA, SYBR Green real-time PCR was performed using the ABI Prism 7700 Sequence Detection system (PE Applied Biosystems, Rotkreuz, Switzerland) as described previously (14). Specific primer pairs for each gene were designed with Primer Express software (PE Applied Biosystems). (A table of primers used in these experiments can be obtained by contacting the authors.) Samples without enzyme in the reverse transcription were used as a control (non–reverse transcriptase control) to exclude genomic contamination. Nonspecific signals caused by primer dimers were excluded by dissociation curve analysis, by analysis of the reaction products on agarose gels, and by use of no-template controls. For quantification of VEGF mRNA, TaqMan realtime PCR was performed with the TaqMan probe and primer sequences for VEGF previously described (7). To normalize for the amounts of loaded cDNA, ␤-actin was used as an endogenous control. After confirmation that the amplification efficiency of the genes of interest and the endogenous control ␤-actin was equal, differences were calculated according to the threshold cycle (Ct) and the comparative Ct method for relative quantification. All measurements were performed in duplicate. Transfection of human dermal fibroblasts with small interfering RNAs (siRNA) against HIF-1␣ and HIF-2␣. Three distinct predesigned siRNA against HIF-1␣ and HIF-2␣, as well as control siRNA, were purchased from Ambion (Huntingdon, UK). Transfection of cells was performed by nucleofection using an Amaxa system (Amaxa, Cologne, Germany) as described previously (6). Eight hours after transfection, the medium was changed, and the cells were either exposed to hypoxia (1% oxygen) or were cultured under normoxic condi- 4205 tions as described previously (6). After 48 hours, RNA was isolated and analyzed by real-time PCR as described above. Inhibition of transforming growth factor ␤ (TGF␤) signaling. To investigate the contribution of TGF␤ signaling to the induction of extracellular matrix proteins by hypoxia, human dermal fibroblasts were cultured in the presence of a neutralizing mouse anti-human TGF␤ antibody (R&D Systems, Wiesbaden, Germany). The median neutralization dose of the neutralizing TGF␤ antibody in the presence of 0.25 ng/ml of TGF␤ was 30 ng/ml. Concentrations of the neutralizing TGF␤ antibody from 20 ng/ml to 500 ng/ml were used for our experiments. Cells incubated with isotype mouse antihuman antibodies (R&D Systems) at the same concentrations were used as controls. After 48 hours under either hypoxic or normoxic conditions, the expression of extracellular matrix proteins was analyzed. Collagen protein measurements. Collagen protein was measured with the Sircol collagen assay as described elsewhere (8). This assay detects protein from types I–XIV collagen using a quantitative Sirius Red binding method. Animals. All experiments were performed according to protocols approved by the local Animal Research Ethics Committee. Female C57BL/6 mice ages 4–6 weeks (n ⫽ 4 animals per group) were exposed to systemic normobaric hypoxia by substitution of oxygen with nitrogen in a closed Persplex chamber using a Digamix 2M 302/a-F pump (H. Wösthoff Messtechnik, Bochum, Germany) at a flow rate of 37 liters/minute (11). Mice were allowed to adapt to hypoxia over a period of 1 hour, with gradually decreased inspiratory O2 fractions, from 21% to 6%, and were then maintained at a fraction of inspired oxygen (FiO2) level of 6% for 24 or 48 hours. The animals were killed immediately after exposure to hypoxia. Stereotactic microdissection of the dermis. The dermis was separated from mouse skin specimens by stereotactic microdissection as described previously (15). Briefly, freshly frozen skin specimens from mice exposed to hypoxia and to normoxia (controls) were cut into 15-m sections and placed on glass slides. Using a histologic stereomicroscope (Stemi DV4; Zeiss, Oberkochen, Germany) at 40⫻ magnification, 10–15 sections from each animal were microdissected by removing the epidermal and subcutaneous parts of the tissue. The remaining dermis was scraped from the slide and immediately placed on dry ice to prevent degradation of RNA. Statistical analysis. Data are expressed as the mean ⫾ SEM. Wilcoxon’s signed rank test for related samples and the Mann-Whitney test for unrelated samples were used for statistical analyses. P values less than 0.05 were considered statistically significant. RESULTS Expression of HIF-1␣ and VEGF in dermal fibroblast cultures. We first compared the induction of HIF-1␣ by immunoblotting of cultured dermal fibroblasts from SSc patients and healthy controls after exposure to hypoxic conditions. While no expression of HIF-1␣ protein was observed in the normoxic controls, a strong and stable expression of HIF-1␣ was found in 4206 DISTLER ET AL Table 1. Extracellular matrix proteins and genes involved in extracellular matrix regulation in fibroblasts from SSc patients and normal controls* SSc fibroblasts Sequence identity Accession number Thrombospondin 1 Pro␣2(I) collagen Fibronectin 1 Microfibrillar-associated protein 4 TGF␤-induced protein IGFBP-3 XM_031616 NM_000089 XM_030549 XM_045044 XM_038210 XM_038123 Up-regulated cultures 6 6 6 5 6 6 of of of of of of 6 6 6 6 6 6 Mean fold induction 3.30 ⫾ 0.93† 2.10 ⫾ 0.27† 1.46 ⫾ 0.27† 2.38 ⫾ 0.54 2.22 ⫾ 0.38† 6.26 ⫾ 1.08† Normal fibroblasts Up-regulated cultures 3 5 5 4 5 5 of of of of of of 5 5 5 5 5 5 Mean fold induction 2.70 ⫾ 0.45 1.97 ⫾ 0.35† 2.13 ⫾ 0.41† 2.63 ⫾ 0.86 2.24 ⫾ 0.53† 7.48 ⫾ 0.88† * Hypoxia-induced genes that were identified by suppressive subtractive hybridization were confirmed in additional samples by real-time polymerase chain reaction analysis. SSc ⫽ systemic sclerosis; TGF␤ ⫽ transforming growth factor ␤; IGFBP-3 ⫽ insulin-like growth factor binding protein 3. † P ⬍ 0.05 for dermal fibroblasts cultured under hypoxic conditions compared with normoxic control fibroblasts. dermal SSc and normal fibroblasts cultured under hypoxic conditions in 1% oxygen. Consistent with recent observations (6), we found no differences in the levels of HIF-1␣ protein between SSc and normal fibroblasts after 24 hours of hypoxia (data not shown). Because VEGF is one of the best-characterized downstream targets of HIF-1␣ and is further increased by hypoxia due to mRNA stabilization, we next analyzed the levels of VEGF mRNA by TaqMan real-time PCR. Consistent with the results of the HIF-1␣ Western blot analysis, a mean ⫾ SEM 2.0 ⫾ 0.1–fold induction of VEGF mRNA was observed in SSc fibroblasts exposed to hypoxia as compared with normoxic controls (P ⬍ 0.05). Again, there was no significant difference between SSc and normal fibroblasts (2.1 ⫾ 0.2–fold upregulation). Together, these results show that hypoxia can be sufficiently induced in dermal fibroblasts under these experimental conditions, with no differences in the expression of HIF-1␣ and VEGF between SSc and normal dermal fibroblasts. Hypoxia-induced genes in cultured dermal fibroblasts. We next aimed to identify downstream targets of hypoxia signaling in dermal fibroblasts using suppression subtractive hybridization. The differential expression of genes identified by subtractive hybridization was confirmed in additional cultures of fibroblasts from SSc patients and was compared with that in normal fibroblasts. The identified and confirmed genes were grouped according to their main biologic function. One group consisted of genes known to inhibit cell proliferation and included basic leucine transcription factor, B cell translocation factor, cyclin T1, and peripheral myelin protein 22/growth arrest–specific 3. Furthermore, induction of enzymes with functions in metabolic pathways, such as triosephosphate isomerase 1 (TPI-1), phosphoglycerate kinase 1 (PGK-1), and NADH dehydrogenase B8 subunit, was observed. Among these enzymes, TPI-1 and PGK-1 are well characterized as hypoxia-driven HIF-1␣ target genes (1), which further confirms that the experimental setting used for the differential screening approach successfully activated cellular hypoxia pathways in the dermal fibroblasts. Six clones of the hypoxia-induced cDNA library were homologous to recently identified gene sequences, for which biologic functions have not yet been established. Moreover, a number of genes with different biologic functions, such as Toll-like receptor 4 and cadherin 13, were found to be induced by hypoxia. (A table showing the genes identified in these experiments can be obtained by contacting the authors.) Consistent with the findings of the HIF-1␣ and VEGF analyses, hypoxia induced the majority of the identified genes to a similar extent in SSc and normal fibroblasts. Hypoxia-induced extracellular matrix proteins. The most interesting group of genes was identified from the hypoxia-induced cDNA library encoded for extracellular matrix proteins and for genes that are involved in extracellular matrix regulation (Table 1). Genes significantly induced by hypoxia in SSc fibroblasts included fibronectin 1 (1.46 ⫾ 0.27–fold induction compared with normoxic controls; P ⬍ 0.05), thrombospondin 1 (3.30 ⫾ 0.93–fold induction; P ⬍ 0.05), pro␣2(I) collagen (COL1A2) (2.10 ⫾ 0.27–fold induction; P ⬍ 0.05), insulin-like growth factor binding protein 3 (IGFBP-3) (6.26 ⫾ 1.08–fold induction; P ⬍ 0.05) and TGF␤induced protein (TGF␤i) (2.22 ⫾ 0.38–fold induction; P ⬍ 0.05). Similar to the other genes, the extracellular matrix proteins were induced in SSc fibroblasts and normal fibroblasts to a similar extent by hypoxia (Table 1). In additional experiments, another extracellular matrix protein, cartilage oligomeric matrix protein (COMP; thrombospondin 5), was induced by hypoxia (2.8 ⫾ 0.4–fold induction; P ⬍ 0.05). We also confirmed these results on the protein level by showing a significant HYPOXIA-INDUCED EXTRACELLULAR MATRIX PROTEINS IN SSc 4207 Figure 1. Time course of the up-regulation of extracellular matrix proteins and related genes induced by exposure to hypoxia for 12–48 hours. Prolonged exposure to hypoxia resulted in an even stronger induction of mRNA for A, thrombospondin 1, B, fibronectin 1, C, pro␣2(I) collagen (col 1A2), D, transforming growth factor ␤–induced protein (TGF␤i), and E, insulin-like growth factor binding protein 3 (IGFBP-3) in cultured dermal fibroblasts from patients with systemic sclerosis (SSc; n ⫽ 5) and normal controls (n ⫽ 5) as compared with normoxic controls (defined as a value of 1). Values are the mean and SEM. ⴱ ⫽ P ⬍ 0.05 for dermal fibroblasts cultured under hypoxic conditions (1% oxygen) compared with the same fibroblasts cultured under normoxic conditions (20% oxygen) for the same time period. induction of collagens after exposure to hypoxia using the Sircol collagen assay (data not shown). Further increases in the production of extracellular matrix proteins after prolonged exposure to hypoxia. In vivo, chronic hypoxia is present in the dermis of patients with SSc because of reduced capillary density and the accumulation of extracellular matrix proteins (6). Thus, the exposure of fibroblasts to hypoxia for 24 hours, as used for the suppressive subtractive hybridization experiments, might not fully reflect the situation in vivo. To address this issue, SSc and normal dermal fibroblast cultures were exposed to hypoxic conditions for prolonged times and were compared with control cultures exposed to normoxic conditions. Notably, after prolonged exposure to hypoxia, most of the extracellular matrix proteins and related genes were significantly increased to levels higher than those achieved after shorter exposure times (Figure 1). After 12 hours, there was only minor up-regulation, but after 24 hours of hypoxia, up-regulation of all genes reached statistical significance as compared with controls. The levels of fibronectin 1, thrombospondin 1, 4208 DISTLER ET AL Figure 2. Oxygen concentration–dependent induction of extracellular matrix proteins and related genes in dermal fibroblasts. Expression of A, fibronectin 1, B, thrombospondin 1, C, pro␣2(I) collagen (col 1A2), D, transforming growth factor ␤–induced protein (TGF␤i), and E, insulin-like growth factor binding protein 3 (IGFBP-3) in cultured dermal fibroblasts from patients with systemic sclerosis (SSc; n ⫽ 5) and normal controls (NH; n ⫽ 5) was analyzed by real-time polymerase chain reaction after 48 hours of exposure to 20%, 16%, 11%, 8.5%, 6%, or 1% oxygen. Values are the mean and SEM. ⴱ ⫽ P ⬍ 0.05 for dermal fibroblasts cultured under hypoxic conditions (as indicated) compared with the same fibroblasts cultured under normoxic conditions (20% oxygen) for the same time period. COL1A2, and IGFBP-3 increased further after 36 hours in a time-dependent manner, and the degree of induction by hypoxia was even more pronounced after 48 hours (Figure 1). The induction of TGF␤i reached its maximum between 24 hours and 36 hours, with a constant up-regulation at later time points. In contrast, the induction of COMP decreased from a mean ⫾ SEM of 2.8 ⫾ 0.4–fold after 12 hours to 1.3 ⫾ 0.1–fold after 96 hours. Again, we did not observe significant differences between SSc and normal fibroblasts for any of these extracellular matrix proteins. These data suggest that prolonged exposure to hypoxia at levels similar to those in vivo has even more profound effects on the induction of most extracellular matrix proteins. Correlation of the induction of extracellular matrix proteins with the levels of oxygen. To further examine the functional association between hypoxia and extracellular matrix proteins, dermal fibroblasts from HYPOXIA-INDUCED EXTRACELLULAR MATRIX PROTEINS IN SSc SSc patients and normal controls were exposed to different oxygen concentrations, ranging from 21% to 1%. The expression of IGFBP-3 did not differ at oxygen concentrations between 16% and 11% as compared with SSc fibroblasts under normoxic conditions. However, at lower oxygen levels, there was a significant concentration-dependent increase in the levels of IGFBP-3 (Figure 2E). At 8.5% oxygen, IGFBP-3 levels were up-regulated by 1.9 ⫾ 0.1–fold and increased further to 2.8 ⫾ 0.4–fold at 6% oxygen. The greatest induction of IGFBP-3, with an increase of 6.3 ⫾ 1.1– fold, was measured at 1% oxygen (Figure 2E), which is consistent with the reduced oxygen levels measured in vivo in the dermis of patients with SSc (6). Similar results were obtained with normal fibroblasts. Analogous to the results for IGFBP-3, significant oxygen concentration– dependent induction was also observed for fibronectin 1, thrombospondin 1, COL1A2, and TGF␤i (Figures 2A– D). These data further confirm the functional correlation between hypoxia and extracellular matrix proteins by showing a concentration-dependent induction, with the strongest effects at reduced oxygen concentrations equivalent to those present in vivo in patients with SSc (6). Role of HIF-1␣ in hypoxia-mediated induction of extracellular matrix proteins. Based on the dose- and time-dependent induction of several extracellular matrix proteins and related genes by hypoxia, we hypothesized that targeting of the major hypoxia transcription factor HIF-1 might be a valuable approach to inhibiting the hypoxia-mediated accumulation of extracellular matrix proteins. To test this hypothesis, embryonic fibroblasts from HIF-1␣–/– and HIF-1␣⫹/⫹ mice were cultured under hypoxic and normoxic conditions. A strong, HIF1␣–dependent regulation of TGF␤i was observed, with a mean ⫾ SEM reduction of 85 ⫾ 8% in MEFs from HIF-1␣–/– mice, as compared with TGF␤i expression in MEFs from HIF-1␣⫹/⫹ mice, under hypoxic conditions (Figure 3). This reduced expression in HIF-1␣–/– MEFs was in the range of the reduced expression of PGK-1 (94 ⫾ 3% reduction in HIF-1␣–/– MEFs), a well-defined HIF-1␣ target that was used as a positive control in these experiments (Figure 3). In contrast, the mean induction of IGFBP-3 and fibronectin decreased by only 35 ⫾ 12% and 47 ⫾ 10%, respectively, in MEFs from HIF-1␣–/– mice compared with those from HIF-1␣⫹/⫹ mice under hypoxic conditions, suggesting that the transcriptional activation of these genes is only partially mediated by HIF-1␣, but is also mediated via HIF-1␣–independent pathways (Fig- 4209 Figure 3. Role of the transcription factor hypoxia-inducible factor 1␣ (HIF-1␣) in the induction of extracellular matrix components by hypoxia. The contribution of HIF-1␣ to the up-regulation of fibronectin 1, insulin-like growth factor binding protein 3 (IGFBP-3), and transforming growth factor ␤–induced protein (TGF␤i) was analyzed using mouse embryonic fibroblasts (MEFs) from HIF-1␣⫹/⫹ (wildtype [WT]) and HIF-1␣–/– (knockout) mice. Phosphoglycerate kinase 1 (PGK-1), an oxygen-sensitive gene known to be HIF-1␣–dependent, was used as a positive control. Mean induction by hypoxia in MEFs from WT mice was defined as 100%, and the mean induction in MEFs from HIF-1␣–/– mice is shown relative to this value. Values are the mean and SEM of 4 independent experiments. ure 3). The expression of Col1a2 was below the level of detection in mouse embryonic fibroblasts, and thrombospondin 1 was not induced by hypoxia in the mouse cell line. To confirm these results in a human system and with adult cells, human dermal fibroblasts were stimulated with deferoxamine mesylate (DFX) and cobalt ions (Co2⫹). DFX and Co2⫹ mimic hypoxia by preventing the degradation of HIF-1␣ protein under normoxic conditions (16). In addition to the genes mentioned above, COL1A2 and thrombospondin 1 were induced by DFX and Co2⫹ in human dermal fibroblasts, confirming the induction of these genes by hypoxia and suggesting at least partial regulation by HIF-1␣ (Figure 4A). The results with the chemical stabilizers of HIF-1␣ were confirmed by knockdown experiments using specific siRNA against HIF-1␣. PGK-1 and VEGF, two well-established HIF-1–dependent genes, were used as positive controls in our analysis of the efficacy of the suppression of HIF-1␣ by the siRNA. The induction of PGK-1 by hypoxia was reduced to 40 ⫾ 8%, and the induction of VEGF was reduced to 26 ⫾ 8% Figure 4. Hypoxia-inducible factor 1␣ (HIF-1␣)–dependent induction of extracellular matrix genes induced by hypoxia in human dermal fibroblasts. A, Induction of extracellular matrix proteins by deferoxamine mesylate (DFx) and cobalt chloride (CoCl2) in fibroblasts from patients with systemic sclerosis (SSc). Similar to hypoxic conditions, the HIF-1␣ stabilizers DFx and CoCl2 induced extracellular matrix proteins and related genes in dermal fibroblasts. Thrombospondin 1, pro␣2(I) collagen (col 1A2), and transforming growth factor ␤–induced protein (TGF␤i) were induced to a similar extent as under hypoxic conditions. In contrast, fibronectin 1 was only slightly induced and insulin-like growth factor binding protein 3 (IGFBP-3) was induced to a much lesser extent than under hypoxic conditions, suggesting that HIF-1␣–independent mechanisms are involved. B–F, Modulation of the induction of thrombospondin 1 (B), pro␣2(I) collagen (C), TGF␤i (D), IGFBP-3 (E), and fibronectin 1 (F) in SSc fibroblasts by 3 different small interfering RNAs (siRNA) against HIF-1␣. Consistent with the results obtained with the chemical stabilizers of HIF-1␣, siRNA against HIF-1␣ strongly reduced the induction of thrombospondin, pro␣2(I) collagen, and TGF␤i by hypoxia, whereas the expression of fibronectin 1 and IGFBP-3 was not significantly reduced. Values are the mean and SEM. 4210 DISTLER ET AL HYPOXIA-INDUCED EXTRACELLULAR MATRIX PROTEINS IN SSc 4211 Figure 5. Transforming growth factor ␤ (TGF␤)–dependent induction of extracellular matrix proteins by hypoxia in human dermal fibroblasts. To study the role of TGF␤ in the induction of extracellular matrix genes by hypoxia, dermal fibroblasts from patients with systemic sclerosis were incubated with neutralizing antibodies against TGF␤ (aTGFb AB) at 20 ng/ml, 100 ng/ml, or 500 ng/ml and cultured for 48 hours under hypoxic or normoxic conditions. Neutralizing antibodies against TGF␤ reduced the induction of A, pro␣2(I) collagen, B, fibronectin 1, C, thrombospondin 1, and D, transforming growth factor ␤–induced protein under both hypoxic and normoxic conditions. The reductions were more pronounced in hypoxic fibroblasts. These results suggest that TGF␤ plays a major role in the induction of extracellular matrix proteins by hypoxia. Values are the mean and SEM. (data not shown); these findings confirm that HIF-1␣ was efficiently suppressed in our system. The 3 siRNA against HIF-1␣ strongly suppressed the induction of COL1A2, thrombospondin 1, and TGF␤i, to 26 ⫾ 12%, 32 ⫾ 7%, and 56 ⫾ 14%, as compared with normoxic control siRNA, suggesting HIF-1␣–dependent regulation of these genes (Figures 4B–D). Consistent with the results with MEFs from HIF-1␣–/– mice, the induction of IGFBP-3 and fibronectin 1 were only slightly reduced (Figure 4E and F), suggesting that HIF-1␣ does not play a major role in the induction of these genes under hypoxic conditions. In contrast to HIF-1␣, transfection of human dermal fibroblasts with siRNA against HIF-2␣ did not reduce the induction of COL1A2, thrombospondin 1, fibronectin 1, IGFBP-3, or TGF␤i (Figures 4B–F). These findings are evidence against an important role of HIF-2␣ in the induction of these extracellular matrix genes in dermal fibroblasts from SSc patients and healthy controls. These data show on different experimental levels that the effects of hypoxia on the induction of extracellular matrix genes are mediated by both HIF-1␣– dependent and HIF-1␣–independent pathways. Thus, targeting of the HIF-1␣ system could be a strategy by which to inhibit the hypoxia-driven synthesis of extracel- 4212 lular matrix proteins, but it might not completely block the hypoxia-mediated effects on the activation of dermal fibroblasts. Moreover, inhibition of HIF-2␣ seems to be inefficient to prevent the induction of extracellular matrix proteins under hypoxic conditions. Dependence of extracellular matrix protein induction on TGF␤. TGF␤ is a major stimulus for the induction of extracellular matrix proteins in SSc fibroblasts. In addition, hypoxia has recently been shown to induce the expression of connective tissue growth factor in SSc fibroblasts (17). To investigate whether the up-regulation of extracellular matrix proteins is mediated by TGF␤-dependent pathways, we incubated SSc fibroblasts under hypoxic and normoxic conditions with neutralizing antibodies against TGF␤. Neutralizing antibodies against TGF␤ completely abrogated the induction of COL1A2, fibronectin 1, thrombospondin 1, and TGF␤i under hypoxic conditions (Figures 5A–D). In fibroblasts subjected to normoxic conditions, neutralizing antibodies against TGF␤ also reduced the production of extracellular matrix proteins. However, the inhibitory effects were not as pronounced as those seen with fibroblasts exposed to hypoxic conditions, suggesting that TGF␤ plays a key role in the induction of extracellular matrix proteins by hypoxia. Induction of extracellular matrix proteins in the dermis in an in vivo mouse model of hypoxia. We next aimed to confirm the in vitro effects of hypoxia on dermal fibroblasts in vivo by using a mouse model of systemic normobaric hypoxia. This mouse model has been validated for the induction of systemic hypoxia in various tissues (11), but it is unknown whether it induces cellular hypoxia in the skin. Thus, we first analyzed the expression of the hypoxia-driven genes VEGF and PGK-1 in RNA extracts from dermal specimens from the back of mice subjected to hypoxic conditions. In fact, VEGF as well as PGK-1 mRNA increased significantly after exposure to hypoxia, by 2.0 ⫾ 0.4–fold and 2.3 ⫾ 0.4–fold, respectively, as compared with mice maintained under normoxic conditions (P ⬍ 0.05) (data not shown). The induction of VEGF and PGK-1 after exposure to hypoxia was also found in dermal RNA extracts from the skin of the ear (2.1 ⫾ 0.4–fold and 2.4 ⫾ 0.1–fold induction, respectively; P ⬍ 0.05) (data not shown). The induction of these 2 established markers of hypoxia at different sites confirmed that hypoxiainduced pathways were activated in the mouse skin. We next analyzed the expression of the identified hypoxia-induced extracellular matrix proteins and related genes in total RNA extracts from the dermis of hypoxia-treated mice. After 24 hours of hypoxia, all DISTLER ET AL Figure 6. Synthesis of extracellular matrix proteins and related genes in the dermis of hypoxia-treated mice. The expression of the hypoxiainduced genes identified by subtractive hybridization in cultured dermal fibroblasts from the dermis of mice treated with hypoxia for 24 hours and 48 hours was analyzed by real-time polymerase chain reaction using SYBR Green. Values are the mean and SEM of 4 mice per experimental condition. col 1A2 ⫽ pro␣2(I) collagen; TGFbi ⫽ transforming growth factor ␤–induced protein; IGFBP-3 ⫽ insulin-like growth factor binding protein 3. genes identified by subtractive hybridization were induced in the dermis, as compared with control mice maintained under normoxic conditions (Figure 6). For example, hypoxic mice produced 3.0 ⫾ 0.2–fold more IGFBP-3 and 2.1 ⫾ 0.2–fold more TGF␤i than did normoxic controls in dermis samples obtained from their back. Similar results were found in studies of dermis samples from the ears of hypoxia-treated mice. Since prolonged exposure to hypoxia resulted in a further up-regulation of these genes in vitro, we speculated that this might also be true for the situation in vivo. Indeed, a stronger induction of fibronectin 1, thrombospondin 1, and Col1a2 was detected in the dermis of mice exposed to hypoxia for 48 hours (Figure 6). Again, similar results were obtained for dermis samples from the back and ears of hypoxia-treated mice. Together, these data confirm in an in vivo model that hypoxia in the skin results in an increased production of several extracellular matrix proteins and related genes. This induction by hypoxia is time-dependent, with stronger effects after prolonged exposure to hypoxia, as was seen in the skin of patients with SSc. HYPOXIA-INDUCED EXTRACELLULAR MATRIX PROTEINS IN SSc DISCUSSION Severe tissue hypoxia and activated interstitial fibroblasts are characteristic features of systemic sclerosis (SSc). The present study was performed with the objective of identifying molecular pathways that are induced by hypoxia and that contribute to the chronic activation of dermal fibroblasts in SSc. The most intriguing finding was the induction of extracellular matrix proteins in dermal fibroblasts. The up-regulation of extracellular matrix proteins by hypoxia demonstrates that hypoxia directly contributes to the development and progression of fibrosis in this disease and therefore provides a novel link between the vascular changes and fibrosis in SSc. The extracellular matrix proteins we identified (fibronectin 1, thrombospondin 1, and collagens) are produced in excessive amounts by SSc fibroblasts and form a major part of the fibrotic material in SSc skin (4). TGF␤i is an extracellular matrix protein that is expressed at high levels in fibrotic lesions, such as arteriosclerotic plaques (18), and in zones of thickened extracellular matrix in the bladder (19). While its role in SSc has not yet been addressed, TGF␤i has been demonstrated to bind to types I, II, and IV collagen (20) and to promote the attachment and spreading of fibroblasts, activities similar to those of fibronectin 1 (21). Another hypoxia-induced protein identified was IGFBP-3, which is involved in the regulation of extracellular matrix proteins. IGFBP-3 binds to extracellular matrix components, protects IGF-1 from degradation, and modulates the effects of IGF-1 on target cells (22). The ligand IGF-1 is profibrotic through its stimulation of the synthesis of collagen and down-regulation of the production of collagenases in fibroblasts. Moreover, IGFBP-3 is overexpressed in fibroblasts from the fibrotic skin of patients with SSc and directly induces the synthesis of fibronectin in lung fibroblasts (22,23). The findings of our differential expression screening analysis of hypoxia-induced genes have a direct impact on virtually all in vitro studies dealing with the pathogenesis of SSc. Hypoxia has a wide range of molecular effects on cells through both HIF-1␣– dependent as well as HIF-1␣–independent mechanisms (1). This is underlined by our results in dermal fibroblasts, which identified hypoxia-induced molecules with a role in such diverse biologic processes as angiogenesis, proliferation, cellular metabolic pathways, and extracellular matrix accumulation. Cell culture studies performed under normoxic conditions at 20% oxygen clearly do not resemble the hypoxic environment present 4213 in vivo in SSc tissues (7). Thus, the interference with multiple hypoxia-driven pathways is overlooked with normoxic culture conditions and has not been taken into account in the majority of in vitro experiments in the past. While the hypoxic stimulus is only present in the skin of patients with SSc, the hypoxia-induced expression profile was not different between normal and SSc dermal fibroblasts in our experimental setting. These results suggest that the induction of extracellular matrix proteins by chronic hypoxia represents a common biologic reaction pattern rather than disease-specific effects. However, since the oxygen levels are severely reduced in the skin of SSc patients as compared with healthy controls (7), the activation of dermal fibroblasts by hypoxia is operative only in SSc, but not in healthy controls. Although it was beyond the scope of the present study to analyze the effects of hypoxia in other organ systems and disease settings, the induction of extracellular matrix proteins by hypoxia might also play a role in other fibrotic diseases associated with vascular abnormalities. Indeed, consistent with our data, chronic hypoxia, for example, has been shown to increase matrix components in renal tubular interstitial cells, fibroblasts, and hepatic stellate cells in vitro (24–26). However, an up-regulation of extracellular matrix proteins and the role of HIF-1␣ in this process have not yet been investigated. Short-term hypoxia such as that which occurs in patients with vasospasms due to primary Raynaud’s phenomenon is not sufficient to induce fibroblast activation. Exposure of dermal fibroblasts to hypoxia for less than 6 hours has no effects on the stabilization of HIF-1␣ protein (7), and in the present study, this did not induce the production of extracellular matrix proteins. This observation is important, because it explains why patients with primary Raynaud’s phenomenon do not progress to a state of fibroblast activation with increased extracellular matrix protein deposition despite their short-term hypoxia caused by vasospasm, a situation opposite that in SSc patients with chronic hypoxia caused by fibrosis and capillary reduction. The induction of major extracellular matrix proteins by hypoxia raises the question of whether hypoxia signaling pathways could be a target for therapeutic approaches. In this regard, it should be emphasized that the hypoxia-induced release of extracellular matrix proteins worsens the hypoxia of interstitial fibroblasts by further increasing the distance to blood vessels. This finally results in a vicious circle of extracellular matrix 4214 DISTLER ET AL accumulation and hypoxia. Consistent with previous studies, we demonstrated that the induction of extracellular matrix proteins is dependent on TGF␤ (25). The stimulation of downstream effects of TGF␤ seems not to depend on the induction of TGF␤, because we did not observe a consistently increased synthesis of TGF␤ mRNA and protein under hypoxic conditions (data not shown). Alternative explanations include increased activation of latent TGF␤ complexes or optimized signaling of TGF␤. Our experiments with MEFs from HIF-1␣–/– mice and with siRNA against HIF-1␣ suggest that targeting of the HIF-1␣ system could be a strategy for inhibiting, at least in part, the hypoxia-driven synthesis of extracellular matrix proteins. Along this line, Kung and coworkers (27) identified a small-molecule inhibitor that blocks the interaction of HIF-1 with the transcriptional coactivator p300, thereby attenuating hypoxiainducible transcription in vitro and in vivo. Unfortunately, this compound proved to be too toxic to allow further development for its use in humans (28). A number of drugs, however, have been identified that indirectly down-regulate HIF-1. This includes inhibitors of histone deacetylases such as trichostatin A, which are in clinical trials in patients with cancers and have recently been shown to reduce the release of collagen from SSc dermal fibroblasts in vitro (29,30). Whether this also holds true in vivo and whether histone deacetylase inhibitors could be a potential therapy for patients with SSc remains to be analyzed. In conclusion, the present study is the first to show that hypoxia induces multiple extracellular matrix proteins in dermal fibroblasts in vitro as well as in systemic normobaric hypoxia in vivo. The induction of extracellular matrix proteins by hypoxia was dose- and time-dependent, with higher inductions under conditions of chronic hypoxia and under low oxygen concentrations as are present in the skin of SSc patients. Analysis with MEFs from HIF-1␣–/– mice indicated that the induction of extracellular matrix genes by hypoxia is at least partly driven by HIF-1␣. Considering that progressive fibrosis is a major cause of morbidity and mortality in SSc patients, targeting of hypoxia pathways such as HIF-1 might be a promising novel approach for the treatment of the late fibrotic stages of SSc. AUTHOR CONTRIBUTIONS Drs. J. H. W. Distler and O. Distler had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study design. J. H. W. Distler, Schett, S. Gay, O. Distler. Acquisition of data. J. H. W. Distler, Jüngel, Pileckyte, KowalBielecka, Marti, O. Distler. Analysis and interpretation of data. J. H. W. Distler, Jüngel, Michel, R. E. Gay, Marti, S. Gay, O. Distler. Manuscript preparation. J. H. W. Distler, Jüngel, Zwerina, R. E. Gay, Matucci-Cerinic, S. Gay, O. Distler. Statistical analysis. J. H. W. Distler, O. Distler. REFERENCES 1. Distler JH, Wenger RH, Gassmann M, Kurowska M, Hirth A, Gay S, et al. Physiologic responses to hypoxia and implications for hypoxia-inducible factors in the pathogenesis of rheumatoid arthritis [review]. Arthritis Rheum 2004;50:10–23. 2. Wenger RH. Cellular adaptation to hypoxia: O2-sensing protein hydroxylases, hypoxia-inducible transcription factors, and O2-regulated gene expression. FASEB J 2002;16:1151–62. 3. Generini S, Matucci-Cerinic M. Raynaud’s phenomenon and vascular disease in systemic sclerosis. Adv Exp Med Biol 1999;455: 93–100. 4. Varga J, Bashey RI. Regulation of connective tissue synthesis in systemic sclerosis. Int Rev Immunol 1995;12:187–99. 5. Hesselstrand R, Scheja A, Akesson A. Mortality and causes of death in a Swedish series of systemic sclerosis patients. Ann Rheum Dis 1998;57:682–6. 6. Distler O, Distler JH, Scheid A, Acker T, Hirth A, Rethage J, et al. Uncontrolled expression of vascular endothelial growth factor and its receptors leads to insufficient skin angiogenesis in patients with systemic sclerosis. Circ Res 2004;95:109–16. 7. Distler JH, Wenger RH, Gassmann M, Kurowska M, Hirth A, Gay S, et al. Physiologic responses to hypoxia and implications for hypoxia-inducible factors in the pathogenesis of rheumatoid arthritis. Arthritis Rheum 2004;50:10–23. 8. Distler JH, Jungel A, Caretto D, Schulze-Horsel U, KowalBielecka O, Gay RE, et al. Monocyte chemoattractant protein 1 released from glycosaminoglycans mediates its profibrotic effects in systemic sclerosis via the release of interleukin-4 from T cells. Arthritis Rheum 2006;54:214–25. 9. LeRoy EC, Medsger TA Jr. Criteria for the classification of early systemic sclerosis. J Rheumatol 2001;28:1573–6. 10. Ryan HE, Lo J, Johnson RS. HIF-1␣ is required for solid tumor formation and embryonic vascularization. EMBO J 1998;17: 3005–15. 11. Stroka DM, Burkhardt T, Desbaillets I, Wenger RH, Neil DA, Bauer C, et al. HIF-1 is expressed in normoxic tissue and displays an organ-specific regulation under systemic hypoxia. FASEB J 2001;15:2445–53. 12. Camenisch G, Tini M, Chilov D, Kvietikova I, Srinivas V, Caro J, et al. General applicability of chicken egg yolk antibodies: the performance of IgY immunoglobulins raised against the hypoxiainducible factor 1␣. FASEB J 1999;13:81–8. 13. Zhang J, Underwood LE, D’Ercole AJ. Hepatic mRNAs upregulated by starvation: an expression profile determined by suppression subtractive hybridization. FASEB J 2001;15:1261–3. 14. Distler JH, Jungel A, Kowal-Bielecka O, Michel BA, Gay RE, Sprott H, et al. Expression of interleukin-21 receptor in epidermis from patients with systemic sclerosis. Arthritis Rheum 2005;52: 856–64. 15. Barghorn A, Komminoth P, Bachmann D, Rutimann K, Saremaslani P, Muletta-Feurer S, et al. Deletion at 3p25.3-p23 is frequently encountered in endocrine pancreatic tumours and is associated with metastatic progression. J Pathol 2001;194:451–8. 16. Wanner RM, Spielmann P, Stroka DM, Camenisch G, Camenisch I, Scheid A, et al. Epolones induce erythropoietin expression via hypoxia-inducible factor-1␣ activation. Blood 2000;96:1558–65. 17. Hong KH, Yoo SA, Kang SS, Choi JJ, Kim WU, Cho CS. Hypoxia HYPOXIA-INDUCED EXTRACELLULAR MATRIX PROTEINS IN SSc 18. 19. 20. 21. 22. 23. induces expression of connective tissue growth factor in scleroderma skin fibroblasts. Clin Exp Immunol 2006;146:362–70. O’Brien ER, Bennett KL, Garvin MR, Zderic TW, Hinohara T, Simpson JB, et al. ␤ig-h3, a transforming growth factor–␤– inducible gene, is overexpressed in atherosclerotic and restenotic human vascular lesions. Arterioscler Thromb Vasc Biol 1996;16: 576–84. Billings PC, Herrick DJ, Kucich U, Engelsberg BN, Abrams WR, Macarak EJ, et al. Extracellular matrix and nuclear localization of ␤ig-h3 in human bladder smooth muscle and fibroblast cells. J Cell Biochem 2000;79:261–73. Hashimoto K, Noshiro M, Ohno S, Kawamoto T, Satakeda H, Akagawa Y, et al. Characterization of a cartilage-derived 66-kDa protein (RGD-CAP/␤ig-h3) that binds to collagen. Biochim Biophys Acta 1997;1355:303–14. LeBaron RG, Bezverkov KI, Zimber MP, Pavelec R, Skonier J, Purchio AF. ␤IG-H3, a novel secretory protein inducible by transforming growth factor-␤, is present in normal skin and promotes the adhesion and spreading of dermal fibroblasts in vitro. J Invest Dermatol 1995;104:844–9. Pilewski JM, Liu L, Henry AC, Knauer AV, Feghali-Bostwick CA. Insulin-like growth factor binding proteins 3 and 5 are overexpressed in idiopathic pulmonary fibrosis and contribute to extracellular matrix deposition. Am J Pathol 2005;166:399–407. Knauer AV, Medsger TA Jr, Wright TM, Feghali CA. Insulin-like growth factor binding proteins contribute to the increased extracellular matrix production by fibroblasts from patients with sys- 24. 25. 26. 27. 28. 29. 30. 4215 temic sclerosis (SSc) [abstract]. Arthritis Rheum 2001;44 Suppl 9:S378. Corpechot C, Barbu V, Wendum D, Kinnman N, Rey C, Poupon R, et al. Hypoxia-induced VEGF and collagen I expressions are associated with angiogenesis and fibrogenesis in experimental cirrhosis. Hepatology 2002;35:1010–21. Falanga V, Tiegs SL, Alstadt SP, Roberts AB, Sporn MB. Transforming growth factor-␤: selective increase in glycosaminoglycan synthesis by cultures of fibroblasts from patients with progressive systemic sclerosis. J Invest Dermatol 1987;89:100–4. Orphanides C, Fine LG, Norman JT. Hypoxia stimulates proximal tubular cell matrix production via a TGF-␤1-independent mechanism. Kidney Int 1997;52:637–47. Kung AL, Zabludoff SD, France DS, Freedman SJ, Tanner EA, Vieira A, et al. Small molecule blockade of transcriptional coactivation of the hypoxia-inducible factor pathway. Cancer Cell 2004;6:33–43. Kaelin WG Jr. The von Hippel-Lindau protein, HIF hydroxylation, and oxygen sensing. Biochem Biophys Res Commun 2005; 338:627–38. Huber LC, Distler JH, Michel BA, Gay RE, Kalden JR, MatucciCerinic M, et al. Inhibition of histone deacetylases reduces the TGF␤-stimulated production of extracellular matrix proteins of skin fibroblasts from patients with systemic sclerosis [abstract]. Arthritis Rheum 2005;52 Suppl 9:S463. Kim MS, Kwon HJ, Lee YM, Baek JH, Jang JE, Lee SW, et al. Histone deacetylases induce angiogenesis by negative regulation of tumor suppressor genes. Nat Med 2001;7:437–43.