The Hematologist

July-August 2019, Volume 16, Issue 4

RNA Isoforms, Innate Immunity, and Leukemogenesis: IRAK4 As a Novel Therapeutic Target in MDS/AML

Charlotte Brierley, MD Clinical Research Fellow
Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, United Kingdom
Adam Mead, MD, PhD Professor of Hematology
Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, United Kingdom

Published on: June 13, 2019

Smith MA, Choudhary GS, Pellagatti A, et al. U2AF1 mutations induce oncogenic IRAK4 isoforms and activate innate immune pathways in myeloid malignancies. Nat Cell Biol. 2019;21:640-650.

Recurrent somatic mutations in core components of the spliceosome, the central cellular machine that coordinates processing of pre-messenger RNA (pre-mRNA) to mature mRNA during gene transcription, are prevalent in myelodysplastic syndromes (MDSs) and other myeloid neoplasms.1 These mutations are heterozygous, mutually exclusive, and cause aberrant splicing due to altered splice-site recognition. As a result, the three common spliceosome mutations (U2AF1, SF3B1, SRSF2) cause widespread disruption of splicing, resulting in a panoply of differentially expressed isoform variants.2 Functional assessment to identify the key isoform variants that contribute to the development of hematopoietic malignancy represents a major challenge.

In May’s issue of Nature Cell Biology, researchers based at the Cincinnati Children’s Hospital Medical Center, in collaboration with groups in New York, St. Louis, Bethesda, and Oxford, described a powerful experimental approach to identify such functional isoform variants. The researchers first evaluated exon usage in 160 acute myeloid leukemia (AML) samples from The Cancer Genome Atlas (TCGA), specifically focusing on genes that were explicitly regulated at the level of mRNA isoform switching, rather than gene-level expression variance. This identified 887 genes demonstrating high variance of isoform expression that could be used to segregate patients into three groups with distinct clinical prognoses. The group with the most adverse prognosis demonstrated significant enrichment of certain mRNA isoforms in innate immune pathway genes. The most significantly altered gene undergoing isoform switching was interleukin-1 receptor–associated kinase 4 (IRAK4), a serine/threonine kinase acting downstream of the toll-like receptor and interleukin-1 receptor superfamily, interacting with key cell signalling regulators such as MYD88 to control proinflammatory gene expression. The inclusion or exclusion of exon 4 of IRAK4 determines whether a long (IRAK4-L) or short (IRAK4-S) mRNA isoform is generated. Dr. Molly A. Smith and colleagues demonstrated that certain AML human primary cells and AML cell lines preferentially express the IRAK4-L RNA and protein isoform, when compared with healthy cord blood and bone marrow CD34+ cells, which predominantly express IRAK4-S.

Next, they used a range of different experimental approaches to demonstrate that IRAK4-L is a key component of the myddosome, interacting with MYD88 to activate NF-κB and MAPK pathways, whereas IRAK4-S preferentially activates MAPK signalling. Immunoprecipitation assays demonstrated that IRAK4 binds Myd88 directly via the N terminal domain, which is lost in IRAK4-S. Leukemic cell function in vitro and in vivo was shown to be dependent on IRAK4-L, including through treatment with CA-4948, an IRAK4 kinase inhibitor in clinical development. Importantly, neither genetic nor pharmacologic inhibition of IRAK4 had a major impact on normal hematopoiesis.

Finally, the authors evaluated the genetic alterations which result in aberrant IRAK4-L splicing and demonstrate correlation between IRAK4-L expression and the hotspot S34F mutation in U2AF1. Selective IRAK4 inhibition in U2AF1 mutant primary patient cells suppressed serial engraftment of MDS cells in a xenograft model, suggesting that inhibition of IRAK4-L impairs U2AF1-mutant MDS-propagating cells in vivo.

Overall, this comprehensive functional analysis of a targetable oncogenic RNA isoform elegantly demonstrates interactions between innate immunity, pathogen stimulation, and leukemic transformation. The authors also proposed that IRAK4-L expression might underlie aberrant IRAK-1 activation recently reported in MDS and AML. The mechanism of aberrant IRAK4-L expression in non-U2AF1 mutant AML cases remains unclear. These findings provide proof of principle that aberrant splicing associated with spliceosome mutations creates therapeutic vulnerabilities relating to specific isoforms, presenting an alternative strategy to target these poor prognosis malignancies in addition to the use of small-molecule modulators of the spliceosome, currently in clinical development. More broadly, this study also highlights the importance of understanding the mechanisms of leukemogenesis where control of innate immunity is disturbed, and whether there is overlap between pathways activating innate immunity and those regulating hematopoietic stem/progenitor cells during MDS/AML development.

In summary, Dr. Smith and colleagues identified a poor-prognosis subtype of MDS/AML defined by the accumulation of certain mRNA isoform changes in innate immune pathway genes. The preferential expression of IRAK4-L occurs as an event downstream of the known U2AF1 S34F mutation and is seen in numerous MDS and AML cell lines and primary cells. IRAK4-L retains an N terminal domain that interacts directly with Myd88 to promote maximal activation of NFkB when compared to the short isoform, which promotes MAPK activation. A selective IRAK4 inhibitor halts leukemic growth both in vitro and in transplantation experiments of primary cells into immunosuppressed mice, providing a putative novel therapeutic strategy in MDS/AML.


  1. Yoshida K, Sanada M, Shiraishi Y, et al. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature. 2011;478:64-69.
  2. Pellagatti A, Armstrong RN, Steeples V, et al. Impact of spliceosome mutations on RNA splicing in myelodysplasia: dysregulated genes/pathways and clinical associations. Blood. 2018;132:1225-1240.

Conflict of Interests

Dr. Brierley and Dr. Mead indicated no relevant conflicts of interest. back to top