Three-dimensional growth sensitizes breast cancer cells to treatment with ferroptosis-promoting drugs

Abstract

Abstract Drugs causing ferroptosis, iron-mediated cell death, represent promising tools for cancer treatment. While exploring the effect of these drugs on breast cancer (BC), we found that a ferroptosis-inducing drug erastin dramatically inhibits tumorigenicity of human BC cells in mice but when used at a concentration known to effectively kill other cell types only modestly reduces such growth in 2D monolayer culture. BCs grow in vivo as 3D masses, and we found that ferroptosis inducers erastin and sulfasalazine inhibit growth of multiple human BC cell lines in 3D culture significantly stronger than in 2D culture. To understand the mechanism of this differential effect, we found that ferroptosis inducers upregulate mRNAs encoding multiple direct and indirect autophagy stimulators, such as ATG16L2, ATG9A, ATG4D, GABARAP, SQSTM/p62, SEC23A and BAX, in tumor cells growing in 2D but not in 3D culture. Furthermore, these drugs promoted autophagy of tumor cells growing in a 2D but not in a 3D manner. We observed that pharmacological inhibition of autophagy-stimulating protein kinase ULK1 or RNA interference-mediated knockdown of autophagy mediator ATG12 significantly sensitized tumor cells to erastin treatment in 2D culture. We also found that ferroptosis-promoting treatments upregulate heme oxygenase-1 (HO-1) in BC cells. HO-1 increases cellular free iron pool and can potentially promote ferroptosis. Indeed, we observed that HO-1 knockdown by RNA interference reversed the effect of ferroptosis inducers on BC cell 3D growth. Hence, the effect of these drugs on such growth is mediated by HO-1. In summary, autophagy triggered by ferroptosis-promoting drugs reduces their ability to kill BC growing in a 2D manner. This protection mechanism is inhibited in BC cells growing as a 3D mass, and ferroptosis-promoting drugs kill such cells more effectively. Moreover, this death is mediated by HO-1. Thus, ferroptosis induction represents a promising strategy for blocking 3D BC growth.