Glioblastoma mutliforme (GBM) is one of the most lethal and untreatable human malignancies. Although genetic alterations in EGFR are observed in almost half of patients with GBM, targeting EGFR has demonstrated limited efficacy in the clinic. Mechanisms of resistance, either upfront or acquired, ultimately lead to therapeutic failure. Recent evidence has suggested the presence of molecular heterogeneity can contribute to EGFR inhibitor resistance. While GBM is characterized by striking heterogeneity, whether this phenomenon occurs in GBM is unknown. To address this, we developed a novel microfluidic platform capable of characterizing cell surface protein expression and oncogenic signaling within individual cancer cells. This technology allowed us to molecularly characterize individual subpopulations of cells in a heterogeneous model of GBM and GBM patients. In addition, we were able to identify a biochemical response to targeted therapy within cell subpopulations in patient-derived GBM xenografts and multiple myeloma patients.

Finally, using this microfluidic device, we examined the biological consequences of cellular hetereogeneity in the context of resistance to EGFR inhibition in GBM. Through modeling therapeutic failure with sustained treatment with an EGFR inhibitor in vivo, we observed selection for a preexisting population that was had undetectable levels of EGFRvIII and was MDM4 amplified. The EGFRvIII- /MDM4 amplified cells remained viable despite continuous EGFR inhibition, indicating a role in tumor maintenance. As MDM4 contributes to the negative regulation of p53, we observed specific ablation of these cells following p53 reactivation with nutlin-3a. These results indicate a previously undescribed mechanism whereby an EGFR inhibitor tolerant subpopulation can contribute to tumor maintenance through p53 regulation. Furthermore, these studies highlight the importance of single cell characterization, especially in cancers with significant intratumor diversity.