EGR1 drives sex-differences in glioblastoma tumorigenicity
Although progress has been made in treating glioblastoma (GBM), with fewer than 5% of patients surviving more than 5 years after diagnosis. For reasons that are not well understood, females are roughly 60% as likely as males to develop GBM, and female patients consistently respond better to treatment than do males. Understanding the molecular etiology of these sex differences in tumor progression and resiliency to treatment could reveal potent new therapeutic targets, ultimately improving survival of both male and female GBM patients. Here we show that the transcription factor Egr1 is a primary mediator of sex differences in multiple GBM tumorigenic phenotypes. In multivariate analysis, high levels of EGR1 expression are correlated with shortened survival for male GBM patients only. To investigate the molecular mechanisms underlying this sex difference, we performed a genomic analysis in our established ex vivo murine model of sex differences, which showed that the transcription factors Egr1 and Klf5 preferentially recruit the transcriptional activator Brd4 to enhancers in male cells relative to female cells, explaining a previously made observation that Brd4 inhibitors reverse sex differences in GBM. Next, using murine and human primary GBM cells, we demonstrated that the small molecule compound SR18662, which downregulates Egr1 and its downstream target Klf5, abrogates GBM growth, migration, invasion, clonogenicity, and response to radiation in a sex-biased fashion. Finally, we knocked down Egr1 and Klf5 via CRISPRi in both untreated and SR18662-treated GBM cells to reveal the sex-biased anti-tumorigenic effects of SR18662 were largely due to Egr1 downregulation, independent of Klf5 downregulation. This result was replicated in vivo. Our results strongly indicate that an Egr1 regulon is a key determinant of sex differences in GBM. As EGR1 is implicated in the cancer biology of many cancers that also display sex differences in incidence and treatment response, our results are likely to be broadly applicable in oncology.