Novel Pathogenic Defects Of Dead/H-Box Helicases In Myeloid Neoplasms

Familial MDS is rare and usually associated with early presentation in childhood. We encountered identical twins presenting with concordant MDS (RCMD-subtype) and a strong family history of leukemia. Their diseases demonstrated several unifying features, including a clinical response to lenalidomide despite neither twin expressing the del(5q) abnormality, resulting in transfusion-independence and restoration of normal counts. Whole exome sequencing (WES) for confirmatory deep targeted sequencing identified the same somatic heterozygous missense mutation of DDX41, c.G1574A (p.R525H) in both cases. In addition, we found diversifying accessory somatic mutations; in case 1 the clonal architecture consisted of DDX41, PHF6 (c.59delG, p.C20fs) and DNMT3A (c.T1180C, p.C394R) in 45%, 14%, and 15% reads, respectively, while in the 2^nd case the clone consisted of 37% DDX41 and 50% JAK2 (c.G1849T, p.V617F) mutant reads. We also found a rare non-synonymous alteration of DDX41, c.T1187C (p.I396T) present in the germline DNA in both twins (the prevalence of this germline minor allele in the general population is <.05%).

DDX41 is a member of the DEAD-box helicase family, the largest of the helicase families. RNA helicases are involved in RNA metabolism, spliceosomal function, ribosome biogenesis, pre-mRNA splicing and translation initiation. To assess the clinical importance of DDX41 and other RNA helicase mutations, we screened a cohort of 763 patients (pts) with myeloid neoplasms and also analyzed the TCGA AML cohort. We found 1.04% (N=10/960) of DDX41 mutations among MDS/AML pts and other RNA helicases mutations including DDX23 as follows : 0.47% (N=2/419), DDX4 : 0.24% (N=1/419), DDX54 : 0.1% (N=1/960) and DHX33 : 0.48%(N=2/419) were demonstrated. We also found 2 rare germline events of DHX29: c.G1627A, p.V543M and c.G1561A, p.E521K.

We further focused on the most prevalent DDX41 lesions located on the long arm of chromosome 5. SNP-array identified deletions of 5q involving the DDX41 locus (5q35.3) in 16% of cases. Expression analysis found that DDX41 mRNA levels were significantly lower in cases with del(5q) affecting the DDX41 locus than in those without deletions (P=.0004). The majority of cases with the DDX41 defect showed advanced disease (higher-risk MDS and sAML). Cases with lower expression of DDX41 showed a significantly worse OS than those with higher expression (HR=1.6, 95%CI=1.04-2.19).

As both twins showed a remarkable responsiveness to lenalidomide despite lack of del(5q), we analyzed an additional 94 pts (MDS,MPN,MDS/MPN) who received lenalidomide for the presence of DDX41. In lower-risk group pts, those who demonstrated DDX41 locus deletions/DDX41 mutations showed a significantly higher sensitivity to lenalidomide than those who did not (P=.024). When we compared expression levels of DDX41 in responders (n=10) and refractory patients (n=9), lenalidomide responders showed significantly lower expression of DDX41 (P=.048), indicating a relationship between DDX41 and lenalidomide sensitivity.

To study the biological consequences of DDX41 low expression, we knocked down DDX41 expression in leukemic cell lines (HL-60 and K562 cell lines) via lentiviral transduced shRNA. Significantly higher proliferation rates were observed in both DDX41 knockdown cell lines compared to control. Based on a synthetic lethal approach, we used the DDX41 knockdown cells to test the effects of helicase inhibitors (C₁₄H₁₅N₃O): DDX41 low expressing cells showed differential sensitivity to helicase inhibitors as compared with control cells.

In sum, DDX41 mutant or haploinsuffcient cases are associated with MDS (as a secondary driver) and may signal lenalidomide responsiveness. DDX41 and other helicases represent a novel class of genes found to carry mutations in MDS. DDX41 and possibly other helicases may be targeted by RNA helicase inhibitors.