DDX41: the poster child for familial AML

In this issue of Blood, Li et al¹ and Duployez et al² define the clinical and genetic landscape of DDX41 germline variants (DDX41^MutGL) detected in 275 patients across the 2 studies, representing the largest collection of patients with DDX41^MutGL reported to date. These studies strengthen claims that acute myeloid leukemia (AML) with DDX41^MutGL represents a unique clinicopathological entity. It is typically characterized by late‐onset disease, normal karyotype, male sex skewing (2.7:1), and favorable outcome. DDX41^MutGL is likely to contribute to the etiology of >5% of all patients with AML, making it the most common predisposing event by a significant margin that has been reported in AML (see figure panel B).

DDX41 is located on the long arm of chromosome 5 (5q35.3). It encodes a member of the DEAD‐box helicase family implicated in messenger RNA splicing and is essential for myeloid differentiation of hematopoietic stem and progenitor cells.^3,DDX41^MutGL was first described as a causative genetic event for inherited predominantly myeloid hematopoietic neoplasms in 2015.⁴ It was immediately included in the World Health Organization classification of myeloid neoplasms.⁵ The studies by Duployez et al and Li et al are timely because they offer guidance for interpreting germline variants and for managing disease in patients with these variants.

Altogether, these studies tested the mutational status of DDX41 in more than 3000 patients with AML. Comparison of patients with DDX41^MutGL and patients with DDX41^WT enrolled on the ALFA and PHYLLO trials by Duployez et al demonstrated significantly higher rates of complete remission and lower rates of relapse after hematopoietic stem cell transplantation in patients with DDX41^MutGL than in patients with DDX41^WT (15% vs 44%). However, this did not translate into prolonged overall survival (hazard ratio, 0.77; 95% confidence interval, 0.35‐1.68; P = .5). The study by Li et al extended previous reports that linked DDX41^MutGL to other hematologic neoplasms,^6,7 including myelodysplastic syndromes, myeloproliferative neoplasms, cytopenia, B‐cell lymphomas, and multiple myeloma.

More than 60% of patients with DDX41^MutGL harbored loss‐of‐function mutations (nonsense, frameshift, or those affecting the initiation codon), with one-third of the variants being either p.M1? or p.D140fs. Although DDX41 mutations were rarely detected in sporadic AMLs (<1%), the majority of patients with DDX41^MutGL AMLs (70%‐80%) acquired an additional DDX41 variant that seems to be distinctive from their germline counterparts, with the p.R525H variant detected in >70% of patients (see figure panel A). This observation is reminiscent of CEBPA and RUNX1 familial cases in which the onset of overt symptoms seems to coincide with the acquisition of an additional and distinctive mutation in the same gene (eg, CEBPA N‐terminal vs C-terminal).^8,9 Conversely, Duployez et al reconstructed the clonal architecture of 10 patients with DDX41^MutGL AML, and they detected shared genetic alterations between matched samples at diagnosis and relapse. This distinguishes DDX41 AML from familial AML with CEBPA^MutGL, in which patients are typically cured of the initial disease but are prone to develop a new leukemia.⁹ 

Li et al noted that the nature of DDX41^MutGL varied between white and Asian patients with p.D140fs and p.M1?, which are common in whites but are much rarer in Asians (62:1) in whom p.A500f and p.Y592C predominate (21:1). As was recently reported for ETV6,¹⁰ all germline variants are not created equal, and indeed, Li et al classified DDX41^MutGL variants into 39 causal (ie, pathogenic/likely pathogenic) variants (CVs) and 43 variants of unknown significance (VUS), according to the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) guidelines.¹¹ In their cohort, 116 patients had germline CVs (66%) and 60 patients had VUS (34%).² Patients with germline VUS represented a distinct group of patients with younger age of onset, lower male predominance, and higher mutational burden. Moreover, only 1 patient with VUS DDX41^MutGL AML (5.5%; 1 of 18) presented with an acquired DDX41 variant in contrast to more than 80% of patients with CV DDX41^MutGL. The nature of additional secondary mutations also differed between CV and VUS DDX41MutGL patients. FLT3 (internal tandem duplication and tyrosine kinase domain) and NPM1 mutations were conspicuously absent in patients with CV DDX41MutGL (<2%) whereas enriched in patients with VUS DDX41MutGL (50%)² (see figure panel B). Together these differences suggest, but do not prove, that the VUS DDX41^MutGL are not related to AML causality.

The late age of onset, incomplete penetrance, and lack of family history make it difficult to recognize inherited forms of DDX41^MutGL‐driven disease. For this reason, it is essential to test bone marrow donors to detect asymptomatic carriers, because several examples of donor‐derived DDX41 leukemia have been described in the literature.^12,13 The studies of Duployez et al and Li et al will be reassuring to centers in which DDX41 screening is now routinely performed and in which there is a degree of uncertainty regarding the impact of DDX41^MutGL variants. The importance of VUS DDX41^MutGL and whether they represent bona fide risk factors or are inconsequential is an open question (see figure panel B), awaiting a determination of the prevalence of VUS DDX41^MutGL in normal populations and a rigorous assessment of the functional consequences of specific variants and their expression.

The perception that familial forms of AML are rare certainly needs to be revisited to ensure that DDX41 and emerging predisposing loci are included as a standard part of routine testing for AML. These genes are set to provide a new perspective on the origins of AML.