Transcription factor Fli1 regulates collagen fibrillogenesis in mouse skin

Abstract

Biosynthesis of fibrillar collagen in the skin is precisely regulated to maintain proper tissue homeostasis; however, the molecular mechanisms involved in this process remain largely unknown. Transcription factor Fli1 has been shown to repress collagen synthesis in cultured dermal fibroblasts. This study investigated the role of Fli1 in regulation of collagen biosynthesis in mice skin in vivo using mice with the homozygous deletion of the C-terminal transcriptional activation (CTA) domain of the Fli1 gene (Fli1(DeltaCTA/DeltaCTA)). Skin analyses of the Fli1 mutant mice revealed a significant upregulation of fibrillar collagen genes at mRNA level, as well as increased collagen content as measured by acetic acid extraction and hydroxyproline assays. In addition, collagen fibrils contained ultrastructural abnormalities including immature thin fibrils and very thick irregularly shaped fibrils, which correlated with the reduced levels of decorin, fibromodulin, and lumican. Fibroblasts cultured from the skin of Fli1(DeltaCTA/DeltaCTA) mice maintained elevated synthesis of collagen mRNA and protein. Additional experiments in cultured fibroblasts have revealed that although Fli1 DeltaCTA retains the ability to bind to the collagen promoter in vitro and in vivo, it no longer functions as transcriptional repressor. Together, these results establish Fli1 as a key regulator of the collagen homeostasis in the skin in vivo.

FIG. 1.

Fli1 ΔCTA modulates mRNA and protein expression of fibrillar collagens in mice skin. (A) Diagram of wild-type Fli1 and Fli1 ΔCTA constructs (left). Primers A and B are located in the N terminus, while primers C and D are located in Ets domain and CTA domain, respectively. Reverse transcription-PCR analysis was performed of RNAs isolated from cultured dermal fibroblasts derived from the back skin of 12-week-old WT and Fli1^ΔCTA/ΔCTA mice. Two distinct primer sets shown in left panel were used. The arrow points to the expected size of the product. (B and D) Quantitative real-time PCR analysis of gene expression of Fli1 (B) and fibrillar collagens (D) in WT (white bars) and Fli1^ΔCTA/ΔCTA (black bars) mice. Total RNA was isolated from skin punches. Quantitative real-time PCR analysis was performed with Sybr green and β2-microglobin as an internal control. The expression level of each gene in control mice was set at 1. Data were obtained from duplicate samples from at least four mice in each group. Values are presented as means ± standard deviations. *, statistically significant value. (C) The expression levels of Fli1 protein were determined in cultured dermal fibroblasts by immunoblotting following immunoprecipitation (IP). An equal amount of cell extracts was used for immunoprecipitation with mouse monoclonal anti-Fli1 antibody using mouse immunoglobulin G-specific immunoprecipitation matrix. Precipitated proteins were subjected to immunoblotting with mouse monoclonal anti-Fli1 antibody and ExactaCruz E Western blot reagent (horseradish peroxidase-conjugated secondary antibody). (E) Pepsin-soluble collagen was prepared as described in Materials and Methods. Equal aliquots from each sample were resolved by 6% SDS-PAGE and stained with Coomassie blue (left panel). Arrows indicate collagen α1(I) and α2(I) subunits. β-Components represent cross-linked α-chain dimers. Graphical representation of collagen levels was quantitated using NIH Image densitometry software. Values are the means ± standard deviations from at least 3 mice in each group. *, statistically significant values.

FIG. 2.

Expression of matrix-related genes differs in Fli1^ΔCTA/ΔCTA mice. (A) Quantitative PCR analysis of gene expression of decorin (Dcn), fibromodulin (Fmod), lumican (Lum), PLOD2, and lysyl oxidase (Lox) in control (white bars) and Fli1^ΔCTA/ΔCTA (black bars) mice. The expression level of each gene in control mice was set at 100. Data were obtained from duplicate samples from at least 5 mice in each group. Values are presented as means ± standard deviations. *, statistically significant values. (B) Immunohistochemistry was performed on paraffin-embedded, formalin-fixed tissue from control and Fli1^ΔCTA/ΔCTA mice (original magnification, ×200; scale bar, 50 μm). To expose core proteins, sections were treated with the appropriate enzyme (chondroitinase ABC for decorin and β-endogalactosidase for lumican).

FIG. 3.

Abnormal fibril structure in the skin of Fli1^ΔCTA/ΔCTA mice. Transmission EM of fibril structure in WT and Fli1^ΔCTA/ΔCTA mice (original magnification, ×20,000; scale bar, 500 nm). In Fli1^ΔCTA/ΔCTA skin, fibrils with increased diameter and irregular shape were consistently seen.

FIG. 4.

Collagen production is increased in skin fibroblasts isolated from Fli1^ΔCTA/ΔCTA mice. (A) Quantitative PCR analysis of fibrillar collagen mRNAs in fibroblasts isolated from skin of WT (white bars) and Fli1^ΔCTA/ΔCTA (black bars) mice. Values are the means ± standard deviations from at least 3 mice in each group. *, statistically significant values. (B) Western blotting of type I collagen produced by fibroblasts isolated from skin of WT and Fli1^ΔCTA/ΔCTA mice. (C) Formaldehyde cross-linked chromatin from Fli1^+/+ and Fli1^−/− MEFs (left) and Fli1^ΔCTA/ΔCTA skin fibroblasts (right) was subjected to ChIP experiments. Immunoprecipitations were performed using polyclonal Fli1 antibody. Immunoglobulin G (IgG) was used for negative control. Input indicates PCR performed on DNA without any immunoprecipitation. After isolation of bound DNA, PCR was performed using primers spanning the region of bp −256 to + 2 of the endogenous Col1a1 promoter, bp −889 to −727 of the Col1a2 promoter, bp −780 to −604 of the Col3a1 promoter, bp −383 to −234 of the Col5a1 promoter, and bp −314 to −153 of the Col5a2 promoter. (D) Expression vectors encoding untagged human WT Fli1 and human Fli1 ΔCTA were transfected into 293T cells for 48 h. Nuclear extracts were incubated with COL1A2 biotin-labeled oligonucleotides. Proteins bound to these nucleotides were isolated with streptavidin-coupled agarose beads, and Fli1 was detected by immunoblotting with anti-calmodulin binding peptide antibody. The levels of Fli1 in cell lysates were determined by Western blotting. (E) Chromatin was isolated from cultured mouse dermal fibroblasts and immunoprecipitated using the indicated antibodies or control IgG. After isolation of bound DNA, PCR amplification was carried out using mouse α1(I) collagen (Col1a1) promoter-specific primers. Input DNA was taken from each sample before addition of an antibody. Data are representative of three independent experiments.

FIG. 5.

Absence of C-terminal domain alters Fli1-dependent protein-DNA and selected protein-protein interactions. (A) Expression vectors encoding human WT Fli1 and human Fli1 ΔCTA were cotransfected with PCAF or PCAF/ΔHAT (PCAF with a deletion of the histone acetyltransferase domain) into 293T cells for 48 h. Total cell extracts were subjected to immunoprecipitation using streptavidin-coupled agarose beads, followed by immunoblotting using rabbit anti-acetylated lysine antibody (AcK). To visualize the total levels of ectopically expressed Fli1, the same membrane was stripped and reprobed with the anti-calmodulin binding antibody. (B) Expression vectors encoding human WT Fli1 and human Fli1 ΔCTA were transfected with or without PCAF into 293T cells for 48 h. Whole-cell extracts were incubated with biotin-labeled oligonucleotides. Proteins bound to these nucleotides were isolated with streptavidin-coupled agarose beads, and Fli1 was detected by immunoblotting with mouse monoclonal anti-Fli1 antibody. The levels of Fli1 in cell lysates were determined by Western blotting. (C) Fibroblasts were transfected with the −772 COL1A2/CAT construct (2 μg), along with the indicated amount of empty vector or Fli1 constructs for 48 h. Values represent the CAT activities relative to those of cells transfected with empty vector only (100 arbitrary units [AU]). The mean and standard deviation from three separate experiments are shown. *, P < 0.05 versus control cells. (D) Nuclear extracts were prepared from cultured mouse fibroblasts. Fli1 was immunoprecipitated using mouse monoclonal anti-Fli1 antibody (BD Bioscience) and IP Matrix from ExactaCruz E kit (Santa Cruz). For blotting, anti-Sp1 (Santa Cruz) and mouse monoclonal anti-Fli1 antibodies were used.