Identification of the TAK1-NF-κB Axis As Critical Regulator of AML Stem and Progenitor Cell Survival.

Abstract 2982

A small population of leukemic stem cells is resistant to chemotherapy and is responsible for the leukemic out-growth and relapse in acute myeloid leukemia (AML) patients. Evasion of apoptosis might be one of the essential mechanisms involved in this process. In order to gain more insight into the differences in the apoptotic programming between normal and leukemic (stem) cells, we recently performed gene array analysis by comparing CD34⁺ AML cells versus CD34⁺ normal bone marrow (NBM) cells. Gene ontology (GO) analysis of the differentially expressed genes between AML and NBM cells revealed differences in GO terms metabolic processes and apoptosis. In order to characterize differences in apoptotic programming in more detail 429 apoptotic related genes were selected and cluster analysis showed that CD34⁺ AML and CD34⁺ NBM cells could be separated into two distinct groups. In particular TGF-β activated kinase 1 (TAK1)/MAP3K7 was one of the apoptosis-related genes that was significantly higher expressed in CD34⁺ AML cells compared to CD34⁺ NBM cells (p = 1.8e⁻⁷). This increased expression of TAK1 could be confirmed by Q-PCR, showing an increase of on average 5.8 fold in TAK1 expression in the studied CD34⁺ AML cells. In mice it has been demonstrated that TAK1 is required for the survival of hematopoietic cells which is largely dependent on TNFR1 and TNFR2. In accordance with these data, we showed that TAK1 is also necessary in human hematopoiesis. Colony-forming cell (CFC) assays showed that inhibition of TAK1 in human cord blood CD34⁺cells, either by shRNAs targeting TAK1 or the TAK1 inhibitor 5z–7-oxozeaenol, resulted in a 2 fold reduction in CFU-GM and BFU-E frequencies compared to control cells. The efficacy was strongly further enhanced by the addition of TNFα, which resulted in a 9.4 fold decrease in CFC colonies upon TAK1 inhibition.

Subsequently, we questioned whether TAK1 inhibition would affect CD34⁺ AML cell survival. Treatment of the AML cell lines MOLM13, OCI-M3 and HL60 with the TAK1 inhibitor 5z–7-oxozeaenol alone only induced modest effects, but in combination with TNFα for 24 hrs a strong induction of apoptosis was observed (IC₅₀ respectively = 23nM, 215nM and 60 nM). Comparable results were observed in HL60 cells transduced with shRNAs targeting TAK1 whereby a downmodulation of TAK1 resulted in a 5.4 fold increase in Annexin V⁺ cells upon TNFα addition. In accordance with previous data, Western blot analysis showed that TAK1 inhibition reduced the levels of p-IκBα, p-p38, p-ERK and p-C-JUN. To test which of these pathways would be important for cell survival, AML cell lines were treated with either the p38 inhibitor SB203580, MEK/ERK inhibitor U0126, JNK inhibitor SP600125 and the NF-κB inhibitor BMS-345541, alone or in combination with TNFα. Addition of the NF-κB inhibitor BMS-345541 induced apoptosis in OCI-M3 and MOLM13 which was significantly increased in combination with TNFα (2.4 fold, p = 0.02). In contrast, inhibition of p38, MEK/ERK and JNK, either alone or in combination with TNFα, did not induce cell death in the AML cell lines. These data suggest that cell death induced by TAK1 inhibition is mainly due to inhibition of the NF-κB pathway. To determine the effects of TAK1 inhibition on primary AML cells, long-term expansion of the leukemic stem cell enriched CD34⁺ AML cell fraction was evaluated in MS5 stromal co-cultures in the absence or presence of TAK1 inhibitor and/or TNFα. Combined treatment for a period of 2 weeks completely abrogated the out-growth of CD34⁺ AML cells, indicating that both leukemic progenitors as well as leukemic stem cells were targeted. In contrast, addition of the single agents did not efficiently reduce cell growth. Similarly, downmodulation of TAK1 using shRNAs strongly sensitized primary CD34+ AML cells for TNFα-induced apoptosis, showing a 6 fold increase in Annexin V⁺ cells compared to control cells. Results on the in vivo efficacy of TAK1 inhibition on primary AML cells are in progress.

In conclusion, our results show that TAK1 is frequently overexpressed in CD34⁺ AML cells, and that inhibition of TAK1 in combination with TNFα is highly efficient in inducing apoptosis of leukemic stem/progenitor cells in a NF-κB-dependent manner.