CLU (clusterin) is a tumor suppressor gene that we have previously shown to be negatively modulated by the MYCN proto-oncogene, but the mechanism of repression was unclear. Here, we show that MYCN inhibits the expression of CLU by direct interaction with the non-canonical E box sequence CACGCG in the 5′-flanking region. Binding of MYCN to the CLU gene induces bivalent epigenetic marks and recruitment of repressive proteins such as histone deacetylases and Polycomb members. MYCN physically binds in vitro and in vivo to EZH2, a component of the Polycomb repressive complex 2, required to repress CLU. Notably, EZH2 interacts with the Myc box domain 3, a segment of MYC known to be essential for its transforming effects. The expression of CLU can be restored in MYCN-amplified cells by epigenetic drugs with therapeutic results. Importantly, the anticancer effects of the drugs are ablated if CLU expression is blunted by RNA interference. Our study implies that MYC tumorigenesis can be effectively antagonized by epigenetic drugs that interfere with the recruitment of chromatin modifiers at repressive E boxes of tumor suppressor genes such as CLU.
A significant proportion of patients suffering from acute myeloid leukemia (AML) cannot be cured by conventional chemotherapy, relapsed disease being a common problem. Molecular targeting of essential oncogenic mediators is an attractive approach to improving outcomes for this disease. The hematopoietic transcription factor c-MYB has been revealed as a central component of complexes maintaining aberrant gene expression programs in AML. We have previously screened the Connectivity Map database to identify mebendazole as an anti-AML therapeutic targeting c-MYB. In the present study we demonstrate that another hit from this screen, the steroidal lactone withaferin A (WFA), induces rapid ablation of c-MYB protein and consequent inhibition of c-MYB target gene expression, loss of leukemia cell viability, reduced colony formation and impaired disease progression. Although WFA has been reported to have pleiotropic anti-cancer effects, we demonstrate that its anti-AML activity depends on c-MYB modulation and can be partially reversed by a stabilized c-MYB mutant. c-MYB ablation results from disrupted HSP/HSC70 chaperone protein homeostasis in leukemia cells following induction of proteotoxicity and the unfolded protein response by WFA. The widespread use of WFA in traditional medicines throughout the world indicates that it represents a promising candidate for repurposing into AML therapy.