Clinicopathological and prognostic significance of DDX41 mutation in myeloid neoplasms: a systematic review and meta-analysis
- Open Access
- 21.04.2025
- Review
Erschienen in:
Abstract
Introduction
Myeloid malignancies are a type of clonal disorders that arise from myeloid stem or precursor cells. These encompass the severe stages, like acute myeloid leukemia (AML), as well as the pre-leukemic phases, which include myeloproliferative neoplasms (MPN), myelodysplastic/myeloproliferative neoplasms (MDS/MPN), and myelodysplastic syndromes (MDS). The advent of next-generation sequencing (NGS) in recent decades has facilitated a more profound comprehension of the molecular pathophysiology of myeloid neoplasms. This has led to the generation of valuable insights pertaining to disease diagnosis, classification, prognosis, and treatment [1‐3].
The DEAD/H-box helicase 41 gene (DDX41), situated on chromosome 5q35, encodes a protein that functions as a DEAD-box-type RNA helicase. DDX41 comprises four domains: a Coiled coil domain, the two Rec-A like domains of the DEAD-box protein core, RecA1 and RecA2, termed DEAD and HelicC in DDX41 literature, and a Zn-finger domain. Two RecA-like domains contain ten conserved motifs. These ten conserved motifs are involved in ATP binding, ATP hydrolysis, nucleotide binding, and RNA unwinding activities [4]. DDX41 plays a role in RNA conformational changes, including splicing of pre-mRNAs, excision of snoRNA-containing introns, modification of pre-rRNA processing, and interaction with R-loops [5].
Anzeige
Heterozygous DDX41 mutations in the human germline were first detected in multiple MDS and AML families [6]. Through the use of next-generation sequencing technology, 28 different germline DDX41 variants have been identified in 43 unrelated patients with MDS/AML [7]. Pathogenic mutations in DDX41 are detected in around 5% of instances of myeloid neoplasms [8]. The DDX41 mutation was identified as the underlying cause in 80% of myeloid tumors with a genetic background, representing 13% of the total cases [9]. The majority of germline mutations occur in the N-terminus of the protein, predominantly within the DEAD domain, including nonsense, frameshift and missense mutations, which ultimately result in premature termination and truncated proteins, and give rise to protein-level amino acid alterations. Though one of the most prominent predisposition genes for myeloid malignancies is the DDX41 mutation, the precise implications of DDX41 mutations in myeloid malignancies remain unclear. For the purpose of precisely characterizing the prognostic significance of DDX41 mutations in myeloid neoplasm patients and to clarify the associations between DDX41 mutations and the clinicopathological features of myeloid neoplasms, we conducted a meta-analysis.
Methods
Literature registration
The Preferred Reporting Items for Systematic review and Meta-Analysis (PRISMA) criteria were adhered to in the writing of the protocol. The registration number for the protocol in the database PROSPERO is CRD42024508577, and it is accessible via the following link.
Search strategy
The PubMed, the Cochrane Library, Embase, Web of Science, MEDLINE, and Google Scholar databases were used to conduct a systematic literature retrieval of the literature. Retrieval time of included studies is limited until January, 2024 and no language restrictions were applied in the research. The keywords included the following terms: DDX41 or DEAD-box helicase 41; myeloid neoplasms or acute myeloid leukemia or chronic myeloid leukemia or myelodysplastic syndrome. In order to identify any additional eligible publications, the references of key articles were inspected.
Anzeige
Literature selection criteria
Studies that fulfilled the specified criteria may be incorporated into the meta-analysis: (1)Any type of myeloid malignancies was involved; (2) The relationship between DDX41 mutations and the resulting clinical outcome; (3) Detailed survival information was provided on patients with DDX41 mutations to obtain overall survival (OS) data with a hazard ratio (HR) and 95% confidence interval (CI). If the data were not provided, we extracted survival data using the Kaplan–Meier curve. In light of multiple reports pertaining to a single study, our meta-analysis encompassed the publication of the highest quality.
Data extraction and quality assessment
The data were retrieved by two researchers operating as individual units, and any discrepancies were resolved through discussion among all investigators. The extracted information is as follows: the first author, year, country of the population enrolled, sample types, sample size, percentage of cases with normal karyotype, germline mutation type, subtype of myeloid neoplasms, concomitant mutations, patient’s age and gender demographics, white blood cells, hemoglobin, platelet count, bone marrow blasts, as well as HR and 95% CI of OS.
The methodological quality of the cohort studies included in this analysis was assessed using the Newcastle-Ottawa Quality Assessment Scale (NOS). A NOS score of six or higher signifies that the study is well-conducted and of high quality.
Statistical analysis
Stata 17.0 software (Stata Corporation, College Station, Texas, USA) and Review Manager (version 5.3, the Cochrane Collaboration, Oxford, UK) were employed to perform the meta-analysis. The relationship between DDX41 mutations and survival outcomes was evaluated using HR and 95% CI. In the absence of direct offers, the requisite data were extracted from Kaplan–Meier curves via Engauge Digitizer Version 4.1. The resulting rates of survival at particular intervals were then transcribed into the spreadsheet devised by Tierney JF et al. for the calculation of HRs and 95% CIs [10]. Moreover, OR and 95% CI were chosen to estimate the relevance between DDX41 mutations and clinicopathologic significance. The statistical heterogeneity among the studies was assessed utilizing a chi-squared (χ²) based Q test in conjunction with I² statistics. In the event that heterogeneity is deemed to be significant (p < 0.1 or I² > 50%), a random effects model will be employed; if not, a fixed-effects model will be utilized [11]. Publication bias was identified through the application of Egger’s and Begg’s tests, with a P-value of less than 0.05 indicating a significant presence of bias [12].
Results
Selected studies and characteristic
A preliminary literature search yielded a total of 57,097 publications from the PubMed, the Cochrane Library, Embase, Web of Science, MEDLINE, and Google Scholar databases, of which 278 were duplicates. A further study, identified through a review of the references, was deemed eligible for inclusion. A total of 102 articles were eliminated following an in-depth analysis of their titles and abstracts. A further comprehensive examination of the full texts was conducted on the 63 articles, resulting in the identification of 20 publications that ultimately constituted the final meta-analysis [6, 7, 13‐30]. The detail of the screening process is demonstrated in Fig. 1. Within the scope of these 20 retrospective studies, a cumulative total of 9,058 patients was examined, with individual sample sizes varying from 2 to 6,336. The range of publication years is given as 2015 to 2023. A summary of the primary characteristics of the research examined is displayed in Table 1. The regions represented in the studies include USA [6, 13, 14, 17, 19‐21, 23, 24, 31], China [15, 16], Korea [18], Japan [22], Thailand [26], France [7], Canada [25], Germany [30], Spain [28] and Australia [27]. A variety of myeloid malignancies were observed among the enrolled patients, including MDS, AML, MPN, MDS/MPN, chronic myeloid leukemia (CML), chronic cytopenia of undetermined significance (CCUS), and idiopathic cytopenia of undetermined significance (ICUS)(Table 1) [6, 13, 15, 16, 18, 20‐24, 26].
Fig. 1
Flow chart for the designated approach
Table 1
Characteristics of the included studies
Author | Year | Country | DDX41 mutations (n) | Gender (M/F) | Age(range) | Subtype | Normal karyotype(%) | NOS score |
|---|---|---|---|---|---|---|---|---|
Quesada et al.[13] | 2019 | USA | 34 | 26/8 | 70(45–84) | MDS, AML, MPN, MDS/MPN | 58.82 | 7 |
Alkhateeb et al.[17] | 2021 | USA | 33 | 24/9 | NR | MDS, AML, MPN, CCUS | NR | 8 |
Badar et al.[20] | 2023 | USA | 107 | 66/41 | NR | MDS, AML, MPN | 84.51 | 6 |
Bannon et al.[14] | 2020 | USA | 35 | 28/7 | 65(48–85) | MDS, AML, MDS/MPN | NR | 7 |
Bataller et al.[21] | 2023 | USA | 151 | 114/37 | 69(21–90) | MDS, AML, MDS/MPN | 58.60 | 8 |
Choi et al.[18] | 2020 | Korea | 28 | 272/185 | 66(41–79) | MDS, AML, ICUS | 75.00 | 7 |
Li et al.[19] | 2022 | USA | 195 | 109/48 | 69(7–92) | MDS, AML, MPN | NR | 6 |
Makishima et al.[22] | 2023 | Japan | 346 | 3940/2396 | 68(15–94) | MDS, AML, MDS/MPN, MPN | 66.90 | 8 |
Qu et al.[15] | 2020 | China | 47 | 44/3 | 65(33–78) | MDS, AML | 98.00 | 8 |
S´ebert et al.[7] | 2019 | France | 43 | 30/13 | 69(36–88) | MDS, AML, MDS/MPN | NR | 8 |
Zhang et al.[16] | 2021 | China | 14 | NR | 59(28–78) | AML, MDS/MPN | 57.14 | 7 |
Goyal et al.[23] | 2021 | USA | 20 | 10/10 | 65(45–88) | MDS, AML, MPN, CML, CCUS | 68.00 | 7 |
Li et al.[24] | 2021 | USA | 28 | 78/47 | 68(48–90) | AML | 80.00 | 7 |
Polprasert et al.[6] | 2015 | USA | 27 | 645/400 | 68(44–88) | MDS, AML, MDS/MPN | 69.23 | 7 |
Tierens et al.[25] | 2023 | Canada | 51 | 38/13 | 69(21–91) | MDS, AML, MPN | NR | 6 |
Guijarro et al.[28] | 2023 | Spain | 2 | 21/26 | 61,72 | AML | NR | 6 |
Jahn et al.[29] | 2023 | USA | 33 | NR | NR | AML | NR | 6 |
Maierhofer et al.[30] | 2023 | Germany | 294 | NR | NR | MDS, AML, MPN | NR | 7 |
Polprasert et al.[26] | 2019 | Thailand | 6 | 5/1 | 67(61–77) | MDS, AML | 66.67 | 7 |
Singhal et al.[27] | 2021 | Australia, USA | 4 | 20/10 | 68(56–84) | MDS, AML | NR | 7 |
Significance of DDX41 mutations on prognosis in myeloid neoplasm patients
We evaluated all eligible studies, the estimated HR for OS was 0.70 (95% CI 0.52–0.93, I² = 69%, P = 0.01), indicating that DDX41 mutation-carrying patients with myeloid neoplasms had a favorable OS compared to DDX41 wild-type matched controls (Fig. 2). Additionally, to identify the causes of heterogeneity in the OS analysis, a subgroup analysis was used. (Fig. 3). AML patients’ pooled HR was 0.60 (95% CI 0.41–0.87, I² = 40%, P = 0.008), according to a subgroup analysis of myeloid neoplasm type, while the HR for patients with MDS/AML was 0.66 (95% CI 0.49–0.88, I² = 31%, P = 0.004). With regard to MDS, however, no evidence of a link was discovered (HR 0.97, 95% CI 0.39–2.43, I² = 89%, P = 0.95). In addition, three studies provided the data for OS among patients with DDX41 germline and somatic mutations, though no statistically significant association between them was found (HR 0.56, 95% CI 0.12–2.62, I² = 91%, P = 0.46) (Fig. 4).
Anzeige
Fig. 2
Forest plot for meta-analysis showed the association between myeloid neoplasms patients with DDX41 mutation and and OS in the entire cohort
Fig. 3
Forest plot of subgroup analysis of OS by myeloid neoplasm type
Fig. 4
Forest plot for the association of DDX41 germline and somatic mutations with OS in patients with myeloid neoplasms
Association between DDX41 mutations and clinicopathological characteristics
Six studies with data on DDX41 mutation and gender association were identified with a total of 4976 patients for analysis. A noteworthy correlation was discerned between the DDX41 mutation and male gender (OR 3.41, 95% CI 1.78–6.54, I² = 41%, P = 0.0002) (Fig. 5A). Four studies with data on DDX41 mutation and age of diagnosis of myeloid neoplasms were identified with a total of 6948 patients for analysis. Age did not correlate with the DDX41 mutation (MD 3.90, 95% CI 0.95–6.86, I² = 60%, P = 0.01) (Fig. 5B).
Anzeige
Fig. 5
Meta-analysis of DDX41 mutation association with clinicopathologic. (A) Association between DDX41 mutations and male gender: OR 3.41, 95% CI 1.78–6.54, I² = 41%, P = 0.0002. (B) Association between DDX41 mutations and patient age: MD 3.90, 95% CI 0.95–6.86, I² = 60%, P = 0.01. (C) Association between DDX41 mutations and the prevalence of MDS: OR 1.57, 95% CI 1.29–1.91, I² = 0%, P < 0.00001. (D) Association between DDX41 mutations and the prevalence of AML: OR 1.05, 95% CI 0.6–1.83, I² = 61%, P = 0.86. (E) Association between DDX41 mutations and blast percentage in bone marrow: MD -4.03, 95% CI -22.12-14.05, I² = 97%, P = 0.66. (F) Association between DDX41 mutations and white blood cell count: MD -4.72, 95% CI -9.98-0.55, I² = 60%, P = 0.08
DDX41 mutation related to prevalence of different myeloid neoplasms. The results of five investigations combined showed a strong correlation between prevalence of MDS and the DDX41 mutation (OR 1.57, 95% CI 1.29–1.91, I² = 0%, P < 0.00001), despite no absence of association was demonstrated between DDX41 mutation and AML (OR 1.05, 95% CI 0.6–1.83, I² = 61%, P = 0.86) (Fig. 5C&D). The DDX41 mutation was not associated with either blast percentage in bone marrow or white blood cells (MD -4.03, 95% CI -22.12-14.05, I² = 97%, P = 0.66; MD -4.72, 95% CI -9.98-0.55, I² = 60%, P = 0.08) (Fig. 5E&F).
Among the 20 publications, 10 articles reported details of germline mutations in 376 patients (Fig. 6). Of the germline DDX41 mutations, p.M1I (74/376, 19.68%) was most common, followed by p.D140fs (61/376, 16.22%) and p.R525H (26/376, 6.91%). Apart from DDX41, the somatic myeloid co-mutations that were most commonly detected were ASXL1 (172/1077, 15.97%), DNMT3A (96/1077, 8.91%), TP53 (84/1077, 7.8%), TET2 (60/1077, 5.57%), and SRSF2 (58/1077, 5.39%).
Fig. 6
Percentage of patients with germline DDX41 mutations
Anzeige
Nine articles reported karyotype information. The frequency of the normal karyotype was higher among DDX41 mutant patients (Table 1). Complex karyotype (27/316, 8.54%) is the most common, followed by del 5q/-5 (18/316, 5.70%) and trisomy 8 (8/316, 2.53%) (Fig. 7).
Fig. 7
Percentage of patients with abnormal karyotypes
Sensitivity test and publication bias
With the aim of evaluating the consistency of the findings and addressing heterogeneity, we performed a sensitivity analysis, whereby each investigation was methodically removed. As shown in Fig. 8, the stepwise exclusion of included studies did not result in any significant differences, suggesting that the individual contributions of each study were inconsequential with respect to the stability of the relationship between DDX41 mutations and OS in patients diagnosed with myeloid neoplasms. This underscores the robust nature of our outcomes. Utilizing funnel plots, Egger’s, and Begg’s tests, the studies on OS were evaluated for publication bias. As indicated by the funnel plots, there is minimal evidence supporting publication bias (Fig. 9). The Egger’s and Begg’s tests for linear regression also demonstrated the absence of a substantial publication bias (Egger’s P = 0.469, Begg’s P = 0.210) (Fig. 10).
Fig. 8
Sensitivity analysis of the meta-analysis of the studies included in the present meta-analysis for OS
Fig. 9
Funnel plot for publication bias in terms of the association of DDX41 mutation with OS
Fig. 10
Forest plot of Egger’s test for publication bias of OS
Discussion
Virtually every facet of RNA metabolism is carried out by highly conserved enzymes designated as RNA helicases. The families of RNA helicases that are known to exist are DEAD-box, DEAH/RHA, Ski2-like, Upf1-like, and RIG-I [31]. The highly conserved amino acid sequence Asp-Glu-Ala-Asp derived the name of DEAD-box proteins. These enzymes, which exhibit high degrees of conservation among diverse species, participate in a range of RNA-related processes, including ribosome biogenesis, pre-mRNA splicing, transcriptional control, translation, RNA export, and RNA degradation [32]. Two domains resembling DNA recombination and repair protein A (RecA) make up the core of DEAD-box proteins. The DEAD-box protein DDX41 is distinguished by a disordered N-terminal region, a DEAD domain, and a Helicase domain. In patients with myeloid tumors, hereditary and sporadic mutations of DDX41 have been identified [21, 24]. DDX41 is a highly mutated gene associated with lengthy latency, severe illness, and a poor prognosis in familial AML/MDS [6]. Despite DDX41 having been recognized as a tumor-inhibitory gene in myeloid neoplasms, several experiments have assessed the predictive value of DDX41 mutations in myeloid neoplasms, yet the results remain inconclusive [6, 7].
The investigation was based on an analysis of 20 studies encompassing a total of 9,058 patients. Of the 20 retrospective studies, only 6 studies measured survival. As evidenced by the forest plots, meta-analyses have consistently demonstrated that DDX41 mutations are a powerful prognostic factor for favorable OS. A meta-analysis of existing literature indicates that, when compared with wild-type controls, DDX41 mutations are positively correlated with OS (HR 0.70, 95% CI 0.52–0.93, P = 0.01). Despite the significant heterogeneity (I2 = 69%), the sensitivity analysis shows that the meta-analysis’s outcomes were consistent and trustworthy. A comprehensive analysis was undertaken in order to elucidate the association between DDX41 mutations and myeloid neoplasms prognosis. The data was stratified based on disease type through the undertaking of subgroup analyses. The findings indicated a notable relationship between OS and the AML and MDS/AML cohorts, with pooled HR of 0.60 (95% CI: 0.41–0.87; I2 = 40%) and 0.66 (95% CI: 0.49–0.88; I2 = 31%), respectively. On the other hand, among MDS patients, nothing indicated a statistically significant link (HR 0.97, 95% CI 0.39–2.43, I² = 89%, P = 0.95). Nevertheless, there were only two studies included in the MDS/AML subgroup, and more research is necessary to determine the precise function of mutation detection in OS. It is important to highlight that no statistically significant correlation was identified in patients exhibiting DDX41 germline and somatic mutations (HR 0.56, 95% CI 0.12–2.62, I² = 91%, P = 0.46).
The DDX41 mutation appears to be connected to male sex, according to our meta-analysis (OR 3.41, 95% CI 1.78–6.54, I² = 41%, P = 0.0002). The study showed that DDX41 mutations are strongly linked to MDS prevalence (OR 1.57, 95% CI 1.29–1.91, I² = 0%, P < 0.00001), whereas no statistically significant correlation between DDX41 gene mutation and AML prevalence. This discrepancy may be attributed to the distinct molecular mechanisms underlying MDS and AML, DDX41 mutations may play a more prominent role in the early stages of myeloid malignancies, such as MDS, rather than in the more genetically complex AML. However, the lack of association with AML prevalence should be interpreted with caution, as it may also reflect the underrepresentation of other myeloid phenotypes, such as MPN, CCUS, and MDS/MPN, in our analysis. Future studies with larger and more diverse cohorts are needed to explore the potential role of DDX41 mutations in these phenotypes. In addition, no significant effect was found on age, AML prevalence, bone marrow characteristics, or white blood cell count.
Another prominent feature is that the frequency of normal karyotype was higher in DDX41 mutation patients. Normal karyotype is associated with favorable prognosis in AML [33], this may partially explain the longer survival observed in patients with DDX41 mutations. While among abnormal karyotypes, the most prevalent is the complex karyotype, which is linked to poor prognosis in both AML and MDS [34]. In the current analysis, p.M1I, p.D140fs and p.R525H are the most common germline DDX41 mutations, which have multiple functions in the pathophysiology of a specific category of myeloid neoplasms [35, 36]. Numerous oncogenic mutations have been shown to be prevalent in myeloid neoplasms by genomic profiling investigations. We have included the most often reported co-mutations in patients who have myeloid neoplasms accompanied by DDX41 mutations. ASXL1, DNMT3A, and TP53 were the three most often found co-mutations. It is presently unknown how precisely these co-mutations in DDX41-mutated myeloid malignancies work. Nonetheless, a number of investigations have suggested that TP53-mutated DDX41 mutations may be able to avoid triggering the DNA damage response system, cell cycle arrest, and death [37, 38]. As a result, the ability to survive and undergo more clonal selection may facilitate malignant transformation.
Although this meta-analysis methodically compiles data emphasizing the critical function of DDX41 mutation as a prognostic factor in myeloid neoplasms, it is imperative to recognize its inherent limitations. Firstly, instead of being prospective research, all of the recruited studies were retrospective designs. A further limitation is that even though we reached out to the authors to request information, there was not enough data to provide statistical power to check for a link. Third, the direct calculation of the HR through the use of the analysis of variance may yield a more accurate result than the HR derived from the survival curve.
In conclusion, we found that OS for myeloid neoplasms is significantly impacted by DDX41 mutations. AML and MDS/AML were found to be substantially correlated with OS by subgroup screening. Moreover, male sex and DDX41 mutations were correlated. However, in order to draw a stronger conclusion, prospective randomized controlled trials with a greater diversity of myeloid neoplasm types and bigger sample numbers are required.
Declarations
Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
