IRAKs are a family of related kinases that operate at the nexus of multiple innate immune and inflammatory pathways implicated in myeloid malignancies. IRAK4 inhibitors have advanced into clinical trials for MDS and AML validating IRAK4 as a therapeutic target. Early data from these trials has been encouraging; however, the overall responses remain modest. Herein, we explored mechanisms of resistance to IRAK4-selective inhibitors in MDS/AML. Our findings uncovered non-canonical IRAK1/4 signaling paradigms driving leukemic stem and progenitor cells (LSPC) and yielded novel therapeutic strategies for MDS/AML.

Consistent with the initial observations from the clinical trials, inhibition or deletion of IRAK4 in a panel of MDS/AML cell lines and patient samples resulted in incomplete suppression of LSPCs and a corresponding activation of innate immune pathways. Given the evolutionary conserved redundancy of IRAK-dependent pathways, we first examined the expression of IRAK paralogs. In IRAK4 knockout (KO) MDS/AML cells, we only observed overexpression and activation of the IRAK4 paralog, IRAK1. IRAK4 kinase inhibitors or PROTACs resulted in a similar compensatory activation of IRAK1 in MDS/AML. To validate the compensation of IRAK1, we deleted IRAK1 in MDS/AML cells and found that concomitant suppression of IRAK1 led to a significant reduction of LSPC function in IRAK4-KO MDS/AML cells. Co-deletion of IRAK1 and IRAK4 similarly reduced leukemic engraftment and extended survival in an AML xenograft model. Hence, IRAK1 mitigates the efficacy of IRAK4-inhibitors in MDS/AML.

Canonical IRAK signaling depends on recruitment of MyD88 and IRAK4 to activated receptors, which results in the subsequent recruitment and activation of IRAK1. Based on this canonical model, IRAK1 and IRAK4 are thus independently essential for signaling. Since dual inhibition of IRAK1/4 is obligatory for suppression of LSPCs, we posited that conventional IRAK signaling is not sufficient nor essential in MDS/AML. Therefore, we first examined the role of proximal upstream (MyD88) and downstream (TRAF6) effectors of IRAK1/4. Deletion of TRAF6 resulted in suppression of LSPCs as observed upon inhibition of IRAK1/4. In contrast, MyD88-KO AML cells did not exhibit a functional defect but did retain sensitivity to deletion of IRAK1 and IRAK4, indicating canonical MyD88-dependent signaling is not operational in MDS/AML. To investigate non-canonical IRAK1/4 signaling we performed RNA- and ATAC-seq and identified gene expression programs dependent on IRAK1 and/or IRAK4, yet independent of MyD88. IRAK1/4 deletion correlated with dysregulation of transcription factors involved in stem cell maintenance and myeloid differentiation. The requirement for IRAK1/4 in preserving an undifferentiated LSPC state was corroborated by morphological assessment of IRAK1/4-KO MDS/AML cells. To delineate the signaling mechanisms by which IRAK1/4 governs stem programs, we performed mass spectrometry proteomics and identified unique IRAK4- and IRAK1-interacting proteins. Integration of the proteomic and transcriptomic studies identified pathways (HIF1A, STAT3/IRFs, E2F4) implicated in stem cell and undifferentiated states as effectors of IRAK1 and IRAK4.

To extend these studies to improve the efficacy of IRAK4-targeted therapy in MDS/AML, we developed a novel series of structurally related inhibitors targeting IRAK1 and IRAK4 (KME-2780: IRAK1 IC50 = 32 nM; IRAK4 IC50 = 0.9 nM) or IRAK4 alone (KME-3859: IRAK1 IC50 = >500 nM; IRAK4 IC50 = 5 nM). In human AML cell lines and patient samples, the dual IRAK1/4 inhibitor was significantly more effective at inducing cell death and differentiation, and attenuating LSPC function as compared to the IRAK4 inhibitor. Thus, the mandate for targeting both IRAK1 and IRAK4 established in our genetic studies translates to pharmacologic interventions.

Overall, we demonstrate that compensation by IRAK1 is a barrier to IRAK4-directed therapy and reveal that dual IRAK1/4 inhibitors are needed for achieving optimal clinical response in MDS/AML. In the process, we also uncovered novel and non-canonical signaling paradigms governing the oncogenic role of IRAK1/4.

Starczynowski:Treeline Biosciences: Research Funding; Tolero Therapeutics: Research Funding; Kymera Therapeutics: Consultancy; Kurome Therapeutics: Consultancy, Current holder of stock options in a privately-held company, Membership on an entity's Board of Directors or advisory committees, Patents & Royalties, Research Funding; Captor Therapeutics: Consultancy; Sumitomo Pharma Oncology, Inc: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution