Background: Approximately 15% of patients with newly diagnosed neuroblastoma are at ultra-high-risk for treatment failure and therefore have the most dismal outcomes. Currently utilized clinicobiological variables are insufficient to identify this cohort at diagnosis, and new strategies are needed to refine risk stratification and improve therapy. It is feasible to use computerized image analysis and proteomic profiling on single slides from diagnostic tissue. The nuclear export protein Exportin-1 (XPO1) is more abundant in those at ultra-high-risk for treatment failure, and this druggable target has been implicated in tumorigenesis of multiple cancers.
Methods: Diagnostic tumor samples were obtained from COG and included high-risk neuroblastoma patients defined as INSS stage 4, greater than 18 months. We compared digital histologic features and proteomic profiles from patients at ultra-high-risk for treatment failure, which was defined as death from disease in less than 18 months, with profiles from those with high-risk disease that was successfully treated (survival without recurrence for greater than 5 years). Top targets of interest were further characterized in neuroblastoma cell lines, and the mechanism of action of Selinexor, an XPO1 inhibitor, was evaluated.
Results: Comparative digital histologic analysis identifies distinct features present in those at ultra-high-risk for treatment failure. Comparative proteomics reveals high XPO1 as a most differentially abundant protein associated with inferior outcomes. Correlative studies using immunohistochemical staining intensity are ongoing. XPO1 is variably expressed across cell lines, with proliferative defects and apoptosis observed after treatment with Selinexor. Selinexor results in downregulation of XPO1 and anti-apoptotic proteins BARD1 and Survivin, correlated with a decrease in acetylated STAT3.
Conclusions: Computerized image analysis and proteomics may complement pathologist review and refine contemporary neuroblastoma risk stratification. Identification of patients at ultra-high-risk for treatment failure at diagnosis has potential to improve outcomes. Continued evaluation of XPO1, in addition to other protein targets, will offer insight into the pathogenesis of the most highly aggressive disease and rational combination approaches to therapy.