HIF2α negatively regulates MYCN protein levels and promotes a low-risk noradrenergic phenotype in neuroblastoma

Abstract The role of HIF2α, encoded by EPAS1, in neuroblastoma remains controversial. Here we demonstrate that induction of high levels of HIF2α in MYCN-amplified neuroblastoma cells results in a rapid and profound reduction of the oncoprotein MYCN. This is followed by an upregulation of genes characteristic of noradrenergic cells in the adrenal medulla. Additionally, upon induction of HIF2α, the proliferation rate drops substantially, and cells develop elongated neurite-like protrusions, indicative of differentiation. In vivo HIF2α induction in established xenografts significantly attenuates tumour growth. Notably, analysis of sequenced neuroblastoma patient samples, revealed a negative correlation between EPAS1 and MYCN expression and a strong positive correlation between EPAS1 expression, high expression levels of noradrenergic markers, and improved patient outcome. This was paralleled by analysis of human adrenal medulla datasets wherein EPAS1 expression was prominent in populations with high expression levels of genes characteristic of noradrenergic chromaffin cells. Our findings show that high levels of HIF2α in neuroblastoma, leads to drastically reduced MYCN protein levels, cell cycle exit, and noradrenergic cell differentiation. Taken together, our results challenge the dogma that HIF2α acts as an oncogene in neuroblastoma and rather suggest that HIF2α has potential tumour suppressor capacity in this particular disease. Significance statement HIF2α has been proposed as a neuroblastoma oncogene and a tractable target for clinical intervention, this has been questioned by several studies. Thus, it is necessary to move beyond correlative studies and better determine the function of HIF2α in neuroblastoma. Our study shows that induced expression of HIF2α in MYCN-amplified neuroblastoma substantially reduces MYCN protein levels and attenuates proliferation while it induces several features of noradrenergic differentiation and impedes xenograft tumour growth. Together with bioinformatic analysis of sequenced neuroblastoma patient samples and the developing human adrenal medulla, this couples HIF2α to low-risk neuroblastoma with a substantially better patient outcome. Thus, in neuroblastoma HIF2α exhibit tumour suppressor capacity rather than oncogenic capacity. Competing Interest Statement The authors have declared no competing interest.