At least 300ng of RNA from each sample was reverse-transcribed using the Bio-Rad iScriptTM cDNA Synthesis Kit (Cat# 170-8891; Bio-Rad, Hercules, CA). is usually associated with two impartial chromatin-remodeling complexes at its amino (N-) and carboxy (C-) terminus [31C33]. It is regulator of brain development and most studies have focused on its function as a negative regulator of neuronal lineage specification in embryonic stem cells and neural progenitors [34C43]. EB 47 REST expression is dysregulated in various tumors of neural or neural crest origin including medulloblastoma [44, 45], glioblastoma [46, 47], Ewings sarcoma [48, 49] and neuroblastoma [50C52]. Previous work from our group as well as others has shown that REST is usually important for medulloblastoma progression and maintenance [53]. However, REST biology in DIPG has not been evaluated thus far. Here we show that REST gene and protein expression is usually elevated in DIPG samples compared to normal controls. It is also expressed to various levels in DIPG cell lines. REST loss diminished DIPG cell growth and formation of intracranial tumors. This was due to EB 47 a decrease in cell proliferation. In addition, DIPG tumors resulting from cells with REST loss exhibited a decrease in CD31, an endothelial marker, and vascular endothelial growth factor receptor 2 (VEGFR2) staining. assays revealed a significant decrease in the ability of human umbilical vascular endothelial cells (HUVEC) to form tubes when cultured in medium harvested from DIPG cells where REST expression was knocked down. This change in tube formation was not due to endothelial cell proliferation. In mechanistic studies, we observed that levels of REST and that of the pro-angiogenic protein and ligand for VEGFR2, Gremlin-1 (GREM-1), were directly correlated in DIPG xenografts. REST knockdown caused a decline in secreted GREM-1 as measured by ELISA. Knockdown of decreased the ability of DIPG cells to support the formation of tubes EB 47 by both HUVEC and human brain micro-vascular endothelial cells (HBMECs). The ability of GREM-1 to promote downstream AKT activation in HUVEC and HBMECs was confirmed using recombinant GREM-1. Thus, our study is the first to implicate REST in DIPG tumors. We also demonstrate an autocrine and paracrine function for REST in DIPG development. The latter involves upregulation of GREM-1 and AKT activation. RESULTS REST is usually expressed at variable levels in human DIPG To evaluate REST expression in DIPG, we obtained microarray datasets made up of gene expression values in human DIPG tumors from Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo) and analyzed through the GEO2R interface. REST mRNA levels were significantly elevated in DIPG tumor samples (n=35) compared to normal brain (n=10). This elevation was particularly significant in DIPGs with H3K27M mutation (Physique ?(Figure1A).1A). Further, human formalin-fixed paraffin-embedded (FFPE) DIPG specimens (n=19) obtained at autopsy were subjected to immunohistochemical (IHC) analyses. REST expression was scored by a neuropathologist as a negative (0)/ poor and focal (+)/ poor, diffuse or multifocal (++)/ strong and focal (+++)/or strong, diffuse or multifocal (++++). Normal brainstem samples are from patients with DIPG tumors, but from a region where tumor was thought not to be present. Approximately, 21% of tumors showed increased REST expression compared to total number of samples analyzed (Physique ?(Physique1B;1B; Table ?Table1).1). REST transcript and protein levels in three human DIPG (SU) cell lines were determined by q-RT-PCR and western blotting. As shown in Figure ?Physique1C,1C, REST mRNA levels were higher in SU-DIPG-IV and SU-DIPG-VI compared to SU-DIPG-XIII. However, REST protein levels were higher in SU-DIPG-IV and SU-DIPG-XIII relative to SU-DIPG-VI (Physique ?(Figure1D1D). Open in a separate window Physique 1 REST expression is elevated in human DIPG(A) Gene expression profiles measured by microarray. Gene expression datasets deposited in GEO were retrieved and analyzed using GEO2R as described in Materials and Methods. A comparison between normal brain samples and a total of 35 DIPG patient samples were shown around the left side. The same set of patient samples were subdivided into three distinct subgroups (H3-K27M, silent and MYCN) [16] and were compared with samples of an unknown subgroup on the right side. Each dot corresponds to one individual patient. Bars represent mean with standard deviations. *p 0.05; ns=non-significant. (B) Hematoxylin-eosin (H&E) and immunohistochemical analysis (IHC) for REST in FFPE DIPG tumor specimens (n=19) and Acvrl1 normal pons (n=2) was performed as described.