No abstract

Idiopathic pulmonary fibrosis (IPF) is a progressive lung disease. Although the pathogenesis is poorly understood, evidence suggests that genetic and epigenetic alterations, such as DNA methylation, may play a key role. Bone morphogenetic proteins (BMPs) are members of the transforming growth factor-β (TGF-β) superfamily and are important regulators in IPF. Here we identified BMP endothelial cell precursor-derived regulator (BMPER) as a key regulator of fibroblast activation. BMPER is a secreted glycoprotein that binds directly to BMPs and may regulate TGF-β/BMP signaling, but its role in lung fibrosis is not clear. BMPER is highly expressed in human IPF lung fibroblasts compared to normal lung fibroblasts. Demethylation agent 5′-azacytidine decreased BMPER expression in fibroblasts, and attenuated the invasion and migration of IPF lung fibroblasts. Furthermore, siRNA-mediated reduction of BMPER in the human lung fibroblasts impaired cell migration and invasion. 5′-azacytidine treatment additionally regulated BMPER expression and reduced lung fibrosis in mice in vivo. These findings demonstrate that methylation of specific genes in fibroblasts may offer a new therapeutic strategy for IPF by modulating fibroblast activation.

BackgroundAlterations in epigenetic marks, including methylation or acetylation, are common in human cancers. For many epigenetic pathways, however, direct measures of activity are unknown, making their role in various cancers difficult to assess. Gene expression signatures facilitate the examination of patterns of epigenetic pathway activation across and within human cancer types allowing better understanding of the relationships between these pathways.MethodsWe used Bayesian regression to generate gene expression signatures from normal epithelial cells before and after epigenetic pathway activation. Signatures were applied to datasets from TCGA, GEO, CaArray, ArrayExpress, and the cancer cell line encyclopedia. For TCGA data, signature results were correlated with copy number variation and DNA methylation changes. GSEA was used to identify biologic pathways related to the signatures.ResultsWe developed and validated signatures reflecting downstream effects of enhancer of zeste homolog 2(EZH2), histone deacetylase(HDAC) 1, HDAC4, sirtuin 1(SIRT1), and DNA methyltransferase 2(DNMT2). By applying these signatures to data from cancer cell lines and tumors in large public repositories, we identify those cancers that have the highest and lowest activation of each of these pathways. Highest EZH2 activation is seen in neuroblastoma, hepatocellular carcinoma, small cell lung cancer, and melanoma, while highest HDAC activity is seen in pharyngeal cancer, kidney cancer, and pancreatic cancer. Across all datasets studied, activation of both EZH2 and HDAC4 is significantly underrepresented. Using breast cancer and glioblastoma as examples to examine intrinsic subtypes of particular cancers, EZH2 activation was highest in luminal breast cancers and proneural glioblastomas, while HDAC4 activation was highest in basal breast cancer and mesenchymal glioblastoma. EZH2 and HDAC4 activation are associated with particular chromosome abnormalities: EZH2 activation with aberrations in genes from the TGF and phosphatidylinositol pathways and HDAC4 activation with aberrations in inflammatory and chemokine related genes.ConclusionGene expression patterns can reveal the activation level of epigenetic pathways. Epigenetic pathways define biologically relevant subsets of human cancers. EZH2 activation and HDAC4 activation correlate with growth factor signaling and inflammation, respectively, and represent two distinct states for cancer cells. This understanding may allow us to identify targetable drivers in these cancer subsets.

Software for our approach is available for download at: http://www.bioconductor.org/packages/release/bioc/html/ASSIGN.html and https://github.com/wevanjohnson/ASSIGN.