P01.37 Impact of C/EBPβ isoforms on brain tumor initiating cells

Abstract

BACKGROUND: The transcription factor C/EBPβ has been identified as a master regulator of the mesenchymal phenotype in malignant glioma. It is transcribed from an intronless gene, but can be translated into three isoforms (LAP, LAP*, LIP). C/EBPβ has been shown to be upregulated in mesenchymal GBMs and its expression correlates to a worse prognosis in patients. Downregulation of C/EBPβ reduces proliferation and migration in vitro and diminishes tumor growth in vivo. In addition, C/EBPβ protein levels are increased in hypoxic and necrotic areas in human GBM tissue samples. However, expression levels and impact of the different isoforms in glioma are unknown. Crosstalk of C/EBPβ and TFGβ has been shown to influence metastatic behavior of breast carcinoma cells. The aim of this project is therefore to investigate the impact of C/EBPβ isoforms LAP and LIP on phenotype, proliferation and migration of brain tumor initiating cells (BTICs) and their possible interaction with transforming growth factor beta (TGFβ). Methods: LAP and LIP isoforms were stably overexpressed in a set of proneural and mesenchymal BTIC lines. Marker expression was investigated by immunocytochemistry and Western blot. Proliferation was monitored by crystal violet staining and migration by spheroid assay. Genome wide transcriptome analysis was performed by next generation RNA sequencing on a HiSeq (Illumina). Results: Overexpression of LAP ­(full-length protein) induced proneural BTICs to switch towards a mesenchymal phenotype, whereas overexpression of LIP had no such effect. Increased LAP led to enhanced migration in the majority of cell lines irrespective of subtype, whereas LIP mostly decreased migration. Rather the reverse effect was observed regarding proliferation. RNASeq analysis revealed highly differential consequences of LAP and LIP overexpression regarding expressed genes and principal component analysis. TGFβ signaling proved to be one of the most significantly altered pathways both in LAP and LIP cells. However, TGFβ treatment had either no effect or inhibited proliferation or migration independent of C/EBPβ isoform levels. Conclusions: We could show for the first time that overexpression of C/EBPβ is sufficient to induce a mesenchymal phenotype in proneural BTICs and that the LAP isoform is responsible for this transdifferentiation. Independent of glioma subtype, LAP and LIP can elicit opposing effects. Nevertheless, both isoforms might contribute to tumor progression, since shifting the balance between LAP and LIP might add to BTICs capability to adapt to microenvironmental cues, e.g. by adopting a more migratory or proliferative phenotype. In contrast to the findings in breast CA, TGFβ acted independently of C/EBPβ isoforms. Ongoing experiments will provide further insight into downstream effector molecules and regulatory mechanisms for each isoform in GBM.