EXTH-75. IN VITRO AND IN VIVO EFFICACY OF COMBINATORIAL INHIBITION OF LSD1 AND HDACS IN PATIENT DERIVED GLIOBLASTOMA STEM CELL MODELS

Abstract Inhibitors of histone deacetylases (HDACi) have been tested in glioblastoma (GBM), however, single agent clinical efficacy has not been proven, prompting study of combinatorial approaches. A rational HDACi combination strategy is with inhibition of LSD1, lysine specific demethylase 1, a histone demethylase known to exist in complex with HDAC1/2. We previously showed in vivo efficacy of combining a brain permeant LSD1 inhibitor, tranylcypromine (TCP) together with vorinostat. More selective inhibitors of LSD1 have been developed and were tested in the current study together with the HDACi, vorinostat or panobinostat in a panel of patient derived glioblastoma stem cell (GSC) lines that have been characterized as radio-resistant or radiosensitive. To test in vitro sensitivity, dose response experiments were conducted in nine GSC cell lines using three LSD1 inhibitors or two HDACi as single agents or in combination. Two non-tumor lines were also tested to assess selective cytotoxicity for brain tumor lines. Cell viability and clonogenicity assays were performed. Several radio-resistant GSC lines showed sensitivity to single agent LSD1 inhibitors, whereas some radio-sensitive GSCs did not, indicating distinct molecular mechanisms of response. Intracranial orthotopic xenograft models were used to test in vivo efficacy using the radioresistant GSC20 line. Interestingly, LSD1 inhibition alone promoted tumor burden reduction and better survival rates than combination treatment. RNA-Seq guided biomarkers of sensitivity to the LSD1/HDAC inhibitor combination were assessed in vitro and in vivo. Both in vivo and in vitro data show that GSC20 treated with GSK LSD1 inhibitor had higher levels of HKDC1 gene with in vivo data also showing elevated levels of FTH1. In vivo mouse data for vorinostat and combination treatment groups showed that only the FTH1 gene level was elevated, providing a potential explanation for low in vivo efficacy. Collectively, our data suggests that LSD1 inhibition represents a viable strategy in GBM.