Addressing limits of Emx2 therapy of glioblastoma multiforme by transgene insulation and epigenetic pharmacological intervention

Activation of a therapeutic transgene encoding for the transcription factor Emx2 acutely suppressed a variety of glioblastoma tumors in vitro. However, employed in TetOFF configuration, this transgene was not able to eradicate glioblastoma in vivo, in homotopically xenotransplanted, immunocompromised mice, because of its silencing, as assessed by an associated EGFP activity reporter. To fix this issue, we decided to model such silencing in vitro. For this purpose, we took again advantage of an IRES-EGFP reporter, associated to the Emx2 transgene or its control, driven here by TetON machinery. To ease the emergence of silencing, we kept these transgenes inactive for a long time past their lentiviral delivery to tumor cells, and we finally activated them by mild doxycycline supplementation. Moreover, to prevent intricacies originating from unexpected leakage of the transgene prior to its intentional activation, we replaced the Emx2-encoding transgene by a surrogate of it, unable to drive synthesis of Emx2 protein, however still harboring sequences supposed to evoke silencing. Resulting, engineered cultures of tumor cells were profiled by cytofluorometry and, finally, the fraction of them expressing the EGFP reporter below a given threshold was employed as an index of transgene silencing. We detected a generalized and progressive decline in activity of all our lentiviral transgenes (both the Emx2-type one and its control), as well as a supplemental decline, specifically affecting Emx2-type constructs since the very early steps of the analytical procedure. We also found that deposition of H3K27me3 and MeCP2 epigenetic marks could contribute to such phenomena. Next, to mimic two key factors supposed to promote silencing in vivo, low O2 peculiar to the glioblastoma niche and "healthy" glia in contact with tumor, we repeated the assays in a hypoxic environment, in the presence of murine neocortical glia. It turned out that both early, generalized transgene dampening, and supplemental Emx2-specific transgene silencing, were enhanced by coculterd glia and, even more, by combined glia and low oxygen. Finally, to mitigate/prevent transgene silencing, we tested two specific devices. First, we insulated the transgene on its 5' side by means of a Ubiquitous Chromatin Opening element (UCOE). Second, we run our assays in the presence of a dedicated drug mix (M3), supposed to lower levels of silencing epigenetic marks H3K27me3, H3K9me3, and 5meCDNA. Both interventions were fruitful, allowing to counteract specific aspects of generalized and Emx2-transgene-peculiar inactivation. We inferred that their combined use might be enough to restore the ability of the Emx2 transgene to acutely suppress glioblastoma in vivo. Building on these findings, now we plan to employ transgene insulation (as described above or further improved by a supplemental 3' element), as well as the M3 mix, to shield the therapeutic Emx2 transgene, and then assess its hopefully improved ability to eradicate glioblastoma - even in a silencing-prone environement - first in vitro and then in vivo.