The impact of mutational burden, spliceosome and epigenetic regulator mutations on transfusion dependency in dysplastic neoplasms

Introduction

In the updated 5th edition of the World Health Organization (WHO) Classification of Hematolymphoid Tumors, genetic-based diagnosis has become more incorporated and refined. Defining genetic abnormalities is integral to this classification in significant conditions, such as myelodysplastic neoplasm (MDS) and chronic myelomonocytic leukemia (CMML). In recent years, massive parallel sequencing or next-generation sequencing (NGS) evolved from a sophisticated scientific technique to a widely available method for routine practice. This advancement has improved diagnostic precision and enabled more targeted therapeutic options [1], 2]. Patients with MDS or CMML often require the supportive therapy, including red blood cell concentrates (RBC) or platelet concentrates (PC) along with disease-specific treatments [3].

In MDS, the leading cytopenia is macrocytic anemia, which can be regarded as the principal symptom of MDS and is a major contributor to impairments in patients’ health-related quality of life [4], 5]. The estimated prevalence of thrombocytopenia in MDS ranges from 40 to 65 %. At platelet counts <50 × 10⁹/L, hemorrhagic events can occur in half of MDS patients [6]. Tang et al. reported severe thrombocytopenia (defined as platelets <20 × 10⁹/L) in 26.5 % of patients, which was associated with life-threatening complications, including gastrointestinal and intracranial bleeding events, resulting in poor overall survival outcomes [7]. Rozema et al. investigated the transfusion frequency in MDS patients and found that almost 50 % had a high transfusion burden, which was defined as ≥8 units/16 weeks. The study identified older age, a high-risk MDS status, and MDS with multilineage dysplasia or MDS with blast excess as risk factors for higher transfusion requirements. However, none of the above studies investigated the association between genetic lesions and transfusion dependency [8].

CMML is divided into myelodysplastic (MD-CMML) and myeloproliferative (MP-CMML) types depending on the absolute leukocyte count (cut off: MD-CMML <13 × 10⁹/L). MD-CMML and MDS share similar phenotypes with predominant cell lineage myelodysplasia and consequent cytopenia. However, MD-CMML exhibits an absolute monocytosis of ≥0.5 × 10⁹/L. The criteria of absolute monocytosis were refined in the updated WHO classification and lowered to a new cut-off value [9]. In CMML, as with MDS, anemia and thrombocytopenia are highly prevalent, which are also standard criteria for eligible prognostic scoring systems for the disease [10], 11].

Considering the mutational landscape of MDS and CMML, evidence regarding their typical mutations has grown in recent years. In CMML, TET2 mutations are the most frequent. TET2 mutations have also been identified as early events in CMML, affecting the hematopoietic stem cell level. Second-order mutations that accumulate in progenitor cells are mutations encoding the spliceosome apparatus, such as SRSF2, SF3B1, U2AF1, and ZRSR2, or mutations in genes that are relevant to epigenetic control of transcriptional processes involving histone modification (EZH2, ASXL1) and DNA methylation (TET2, DNMT3A, IDH1, and IDH2). Mutations of cell signaling pathways, such as JAK2, KRAS, NRAS, CBL, PTPN11, and FLT3, are typically associated with a myeloproliferative CMML phenotype. In CMML, particularly in MD-CMML, spliceosome mutations (SM) and epigenetic regulatory mutations (ERM) have been iteratively confirmed as the most prevalent mutations [10], [11], [12].

The recurrent mutations of MDS are similar to those in MD-CMML. The clonal architecture of MDS at diagnosis is characterized by the presence of 2–4 causative driver mutations. In MDS, as in CMML, the main mutations belong to the groups of spliceosome mutations and epigenetic regulator mutations. Previous studies have indicated mutations in the functional groups of SM and ERM are early events in disease initiation, while mutations in miscellaneous gene categories are responsible for disease progression. Co-mutations in other cellular pathways are limited to less than 10 % in MDS and predominately involve transcription factors, cohesin, cellular signaling pathways, DNA repair, or tumor suppressor genes [13], [14], [15], [16].

Materials and methods

A retrospective database analysis was performed to assess the impact of myeloid mutations on transfusion requirements in patients with MDS/MD-CMML. A homogenous study cohort was ensured by including only dysplastic hematological neoplasms with comparable hematological phenotypes, such as cytopenia and myelodysplasia.

The genes included in the ERM group were TET2 (NM_017628.4), DNMT3A (NM_022552.5), ASXL1 (NM_015338.6, Exon 10, 12, 13), EZH2 (NM_004456.5), IDH1 (NM_005896.4, Exon 4), and IDH2 (NM_002168.4, Exon 4) [11], 14], 18]. The SM group genes were SRSF2 (NM_003016.4, Exon 1), U2AF1 (NM_006758.3, Exon 2, 6), SF3B1 (NM_012433.4, Exon 10–16), and ZRSR2 (NM_005089.4) [11], 15], 16]. Additionally, the association between higher transfusion requirements and an increase transformation rate of MDS/MD-CMML into acute myeloid leukemia (AML) was analyzed. AML was defined as the presence of >20 % myeloblasts in peripheral blood and/or bone marrow. Genetic exceptions that allow AML classification below this myeloblast threshold, as per WHO criteria, were not present in our cohort.

Molecular genetic analysis (SM and ERM or NON-SM/ERM mutations) was carried out using the NGS Sophia Genetics^® myeloid solution panel, including 30 genes and gene sections of ABL1, ASXL1, BRAF, CALR, CBL, CEBPA, CSF3R, DNMT3A, ETV6, EZH2, FLT3, HRAS, IDH1, IDH2, JAK2, KIT, KRAS, MPL, NPM1, NRAS, PTPN11, RUNX1, SETBP1, SF3B1, SRSF2, TET2, TP53, U2AF1, WT1, and ZRSR2. NGS was conducted on the Illumina MiSeq platform. DNA was first fragmented and ligated to platform-specific adapters during library preparation. Adapter-ligated fragments were then enriched for the target regions using a hybridization-based capture system to ensure comprehensive coverage of the regions of interest. The prepared libraries were loaded onto the MiSeq flow cell, where cluster generation was achieved through bridge amplification. Sequencing was performed using Illumina’s sequencing-by-synthesis chemistry, enabling high-fidelity base calling. Primary sequence data processing, including base calling and demultiplexing, was carried out with the MiSeq control software. The resultant sequence reads were aligned to the hg19 human reference genome using the Data-Driven Medicine (DDM) software from Sophia Genetics, which also facilitated variant calling and initial classification. Variants were annotated and interpreted using tertiary analysis tools, including ClinVar, Alamut Visual, and VarSome. Variants were classified according to the guidelines of the American College of Medical Genetics and Genomics (ACMG). Variants classified as benign or likely benign were excluded.

Cytogenetics in this study encompassed all structural and numerical chromosomal abnormalities, which were detected through conventional karyotyping and/or fluorescence in situ hybridization (FISH) analyses.

Basic patient data and the administered RBC and PC were assessed by analyzing electronic patient records. The initial diagnoses of MDS/MD-CMML were established according to the 2017 WHO diagnostic criteria. During the retrospective database analysis, the updated 2022 WHO diagnostic criteria were carefully considered to ensure consistency with the revised definitions and classifications (updated 2022 WHO diagnostic criteria).

The electronic medical records of the hospital, Klinikum Wels-Grieskirchen, were analyzed from 2018 (the time of the establishment of Next Generation Sequencing of myeloid mutations in our laboratory institute) until 2023. Patients with MDS or MD-CMML in whom at least one somatic mutation was detected were included in the study. Exclusion criteria were defined to reflect potential confounding transfusion triggers. Thus, patients with concomitant neoplastic disease, chronic renal impairment (higher than CKD stage 3b), and hemolytic or hemorrhagic anemia were excluded a priori. It was unnecessary to add iron, folate, or vitamin B12 deficiencies or alcohol abuse to our exclusion criteria as these conditions are routinely ruled out in clinical practice by our standard hematological diagnostic workup as the most common differential diagnoses of dysplastic neoplasms in Middle Europe.

Results

The primary aim was to investigate whether the number of myeloid mutations influences transfusion dependency and if an increasing number of myeloid mutations increases the need for transfusion. Secondly, the different transfusion requirements between mutations in the main functional genetic groups, such as SM and ERM, were scrutinized. The mutational status of patients was assessed at the time of diagnosis. Transfusion requirements were recorded starting from the time of diagnosis and tracked over the subsequent observation period. A total of 173 patients with MD-CMML or MDS were diagnosed at Klinikum Wels-Grieskirchen during the observation period, and after applying the aforementioned restrictive inclusion and exclusion criteria, 119 were included in the study (Table 1).

The study cohort exhibited demographic and clinical characteristics typical for patients with MD-CMML/MDS, with a mean age of 67.76 years and a slight male predominance (M:F ratio 72:47). The cohort consisted of 30 CMML (25.2 %) and 89 MDS (74.8 %) patients. Anemia was more frequently observed than thrombocytopenia in the cohort, with a higher proportion of MDS cases compared to MD-CMML cases. This is reflected by a median hemoglobin level of 10.8 g/dL, indicative of anemia, and a platelet count of 124 × 10³/μL, which highlights the presence of thrombocytopenia in a subset of patients.

On a molecular level, the most frequently observed mutations were in epigenetic regulators (ERM; median: 2 mutations per patient), followed by spliceosome mutations (SM; median: 1 mutation) and non-ERM/SM mutations (NON-ERM/SM; median: 0 mutations). RBC requirements were notable, with a mean of 9.45 RBC units per month (median: 4.0 RBC units). PC transfusions were infrequent, with a mean of 1.50 PC units per month (median: 0).

First, the impact of the number of mutations on transfusion dependency in MDS/MD-CMML was investigated. Patients with ≥3 myeloid mutations (55 patients) were compared to those with <3 mutations (64 patients). In the group with ≥3 myeloid mutations, the mean hemoglobin value at diagnosis was 11.32 g/dL compared to 9.95 g/dL in patients with <3 mutations. Anemia was more prevalent in patients with a higher number of mutations (93 %) compared to those with fewer mutations (72 %). Transfusion dependency was higher in the cohort with ≥3 mutations, with a mean of 2.60 RBC transfusions per month compared to 0.54 per month in the group with fewer mutations (p<0.001, r=0.6). In addition, a lower mean platelet count (147,000/µL) and a higher rate of thrombocytopenia (83 %) were observed in patients with ≥3 mutations compared to those with fewer mutations (mean platelet count of 173,000/µL and thrombocytopenia rate of 59 %). The PC transfusion dependency was higher in patients with more mutations (mean 0.34 per month) compared to those with fewer mutations (0.01 per month) (Figure 1).

Figure 1:

The number of mutations is consistently shown on the x-axis (A–D). Panel (A) illustrates the difference in RBC levels between individuals with high and low mutation counts. Panels (B–D) depict how transfusion requirements for PC and RBC increase with the number of mutations.

Subsequently, the impact of important confounding factors (RBC transfusion triggers), such as cytogenetic aberrations and NON-SM/ERM mutations, was assessed. A negative binomial regression analysis showed no significant effect of NON-SM/ERM (p=0.424) and cytogenetic aberrations (p=0.873) on RBC transfusion dependency in the study cohort. These results were confirmed by a simple linear regression which considered the observational period: NON-SM/ERM (p=0.6348) and cytogenetics (p=0.390) did not significantly alter the analysis of transfusion dependency of RBC per month in MDS/MD-CMML. Therefore, cytogenetic aberrations and NON-SM/ERM were not significant confounders in our study of the impact of SM and ERM and the total number of mutations on RBC transfusion dependency in MDS/MD-CMML. A negative binomial model, assumed due to strong overdispersion, showed a highly significant correlation between RBC transfusion rate and the total number of mutations (β1=0.53), resulting in an odds ratio of =1.69. Thus, the number of RBC transfusions per month increased by 69 % per mutation.

After analyzing the effect of mutation burden on RBC and PC transfusion dependency, we were interested in whether there were differences between functional mutational groups. The impact of SM and ERM mutations on transfusion requirements in MDS/MD-CMML was analyzed. Cases with at least one ERM or SM, but not both, were included. Significantly higher RBC transfusion requirements per month were observed In the SM-dominant sub-cohort (p<0.001), however the analysis of the PC transfusion dependency did not show a significant difference between the SM and ERM groups (p=0.16).

Finally, analysis revealed a higher transfusion dependency in the cohort of MDS/MD-CMML patients who progressed to AML (15 patients) than those who did not progress (p<0.0001).

Discussion

MDS/MD-CMML are characterized by cytopenia, including anemia and thrombocytopenia. Consequently, the requirements for RBC and PC transfusions are typically high for these patients [19], 20]. Morphological screening of those entities can now be supported by digital pathology systems [21], [22], [23]. Large cohort studies have demonstrated that incorporating molecular genetic data can increase diagnostic and prognostic precision when assessing critical primary endpoints, such as overall survival, progression-free survival, and leukemia-free survival [24], 25]. Given that molecular genetics improve the precision of patient care in MDS and CMML and that transfusion dependency is a frequent requirement of these patients, it was of interest to determine whether molecular genetics can impact transfusion dependency. Therefore, we analyzed the association between mutational burden and transfusion burden in MDS/MD-CMML patients.

Our study investigated whether a high myeloid mutational burden (defined as three or more mutations) is associated with a higher transfusion requirement compared to a low myeloid mutational burden (defined as 1–2 mutations). It was found that patients with a higher number of mutations had a higher prevalence of anemia and thrombocytopenia and required more RBC and PC transfusions. Potential transfusion triggers were excluded via the patient recruitment strategy and statistical analysis using linear regression and negative binomial regression analysis. These new clinical findings expand the spectrum of the genetic-to-phenotype correlation in MDS/MD-CMML. Previous studies have reported adverse patient outcomes in MDS and CMML patients with a higher number of myeloid mutations [24], 25]. The cumulative effect of these mutations results in a higher severity of disease, which is compatible with our findings and indicates higher PC and RBC requirements and a higher AML transformation rate with and increasing mutational burden.

Additionally, the effect of functional molecular genetic groups, beyond the number of mutations, was investigated. Specifically, we compared the ubiquitous molecular genetic groups of SM (including SRSF2, SF3B1, U2AF1, and ZRSR2) and ERM (including TET2, DNMT3A, ASXL1, EZH2, IDH1, and IDH2) in MDS/MD-CMML [12], 26], 27]. To exclude the quantity effect of mutations on transfusion dependency, patients with exclusively one SM or one ERM were compared. In this scenario, it was found that patients with SM had a higher RBC requirement but not a higher PC requirement. Previous in vitro studies have increased our understanding of SM and ERM and their contribution to myelodysplastic changes in hematopoiesis and their role in leukemogenesis. SM occurs in the splice machinery apparatus, where mature mRNA is received after removing intronic sequences from pre-mRNA and ligating the remaining exons. Different compositions and utilization of exons result in different protein isoforms with varying cell functions. One leading concept is that SM alters the removal of introns from the protein-coding transcripts [28], 29]. In two hotspot SM mutations affecting the U2AF1 gene, it has been reported that these mutations result in intron retention with sequence-dependent mis-splicing of genes involved in hematopoiesis [30]. U2AF1 is highly expressed in erythropoietic cells, and U2AF1 knockdown studies have demonstrated reduced erythropoietic cell growth, increased apoptosis, and delayed erythropoietic cell differentiation [31]. Many studies support the concept that in SRSF2 and U2AF1 mutations, alternative exon usage is primarily responsible for deregulated splicing processes, whereas SF3B1 mutations result in aberrant 3`splice site usage in three genes involved in heme biosynthesis and iron metabolism in erythropoietic cells. This dysregulated splicing pathway might be responsible for ring sideroblast formation in erythropoietic cells [32], 33].

The mechanisms of ERM in altering hematopoiesis differ from those of SM. Epigenetic processes predominately modify gene expression (primarily by methylation, posttranslational histone modification, and RNA-based mechanisms) and do not directly alter the DNA nucleotide sequence. Epigenetics in neoplastic processes often report on promotor hypermethylation with consecutive tumor suppressor gene silencing or hypomethylation with consecutive genomic instability. In dysplastic neoplasms, disrupted methylation patterns have been shown to primarily affect the granulopoiesis with insufficient maturation and dysgranulopoietic changes, such as hypogranulation and chromatin condensation [34], [35], [36]. Clinical studies have supported the basic research findings via a higher anemia tendency of SM mutations than in ERM [27], 37], 38]. In addition, ERM are considered earlier mutational events than SM in dysplastic neoplasms. Thus, ERM-dominant dysplastic neoplasms correspond to a lower severity of disease with lower tendency toward cytopenia. Based on this data, this study expanded the clinical characterization of SM vs. ERM and highlighted the clinical significance of the SM-related anemia dimension via increased RBC requirements.

Conclusions

A high number of myeloid mutations are related to a high need for RBC and PC transfusions. A high transfusion dependency is associated with a higher rate of AML transformation. Beyond the quantitative effect of myeloid mutations on transfusion dependency, SM is associated with higher RBC requirements (but not PC requirements) compared to ERM. Incorporating molecular genetics into patient care can improve the precision of transfusion requirements in dysplastic neoplasms, such as MDS/MD-CMML. A precise a priori assessment including molecular genetics at diagnosis, can improve the estimation of transfusion requirements, which is beneficial for patient follow-up. An added value is conceivable because, based on molecular genetics, some patients may need more transfusions and closer monitoring. Therefore, incorporating molecular genetics for estimating transfusion dependency is expedient for clinical practice. This is the first study to specifically analyze the differences in transfusion dependency caused by various functional gene groups such as spliceosome mutations and epigenetic regulator mutations. Further studies with larger sample sizes are needed to confirm the data from this exploratory study.