Blocking TNF and IL-1 Sensitizes AML Stem Cells to NF-κB Inhibitor Treatment

Background: NF-κB inhibition selectively eliminates leukemia stem cells (LSCs) but has less effect on healthy hematopoietic stem cells (HSCs), suggesting an advantageous target for leukemia therapy. However, NF-κB inhibition alone does not have a significant effect on in vivo models of leukemia. We have reported that many types of AML cells produce tumor necrosis factor-α (TNF) which appears to protect LSCs from NF-κB inhibition by stimulating JNK-AP1 survival/proliferative signaling parallel to NF-κB. Complete inactivation of TNF signaling cannot fully reproduce the effects of JNK-AP1 inhibition, suggesting that some other inflammatory cytokine(s) might also compensate for NF-κB inhibition through activation of JNK-AP1. We found that IL-1α and β (IL-1) is such a cytokine.

Methods: TNF and IL-1 expression was examined in 430 AML samples by microarray and verified by qRT-PCR. IL-1α and β protein levels in the sera of AML patients and medium from cultured AML cells were examined by ELISA. MLL-AF9-transduced murine AML cells, human AML cell lines and primary patient samples were used in this study. TNF signal was blocked by genetic deletion of receptors Tnfr1 and 2, or treatment with a TNF-specific monoclonal antibody. IL-1 signal was blocked by shRNA knockdown of IL-1R or treatment with an IL-1R antagonist. The effects of IL-1 and TNF signal co-inactivation on growth and response to NF-κB inhibition in AML cells were evaluated by in vitro culture and colony-forming assay, and in vivo transplantation/leukemogenesis assays.

Results: We found that most types of human AML cells both express and produce TNF and IL-1. Treatment with these cytokines stimulated the growth and colony-forming ability of AML cells but repressed the growth of normal HSPC in in vitro culture. Inactivation of either the TNF or IL-1 signaling axes suppressed the growth of and significantly sensitized AML stem cells to NF-κB inhibition both in vitro by targeting clonogenic AML progenitor cells, and in vivo through compromising leukemogenesis. Simultaneous inactivation of both TNF and IL-1 resulted in an enhanced repressive effect on AML cells compared to inactivation of either individual cytokine’s receptors alone. AML cells with inactivation of both TNF and IL-1 signals showed significantly lower clonogenic ability in vitro, further reduced leukemogenic capacity in vivo and enhanced sensitivity to NF-kB inhibitor treatment both in vitro and in vivo. Mechanistically, we found that both TNF and IL-1 stimulate JNK-AP1 and NF-κB, two parallel survival signaling pathways in AML cells, but induce JNK-mediated death signaling in normal HSPCs.

Conclusion: Both TNF and IL-1 provide protection to LSCs from the effects of NF-κB inhibitor treatment in an autocrine manner and contribute to repression of normal hematopoiesis in a paracrine fashion. This protection is mediated by JNK-AP1. Our studies show inhibiting both TNF and IL-1 inflammatory signals can block this effect, and may be a better strategy to successfully treat AML when combined with NF-κB inhibitors.