Search

Myelodysplastic syndromes (MDS) are a heterogeneous group of diseases with a high risk of transformation to acute myeloid leukemia (AML). One of the processes implicated in MDS pathogenesis is RNA splicing. Its alterations are caused by somatic mutations in splicing factor genes. Mutations in SF3B1(Splicing Factor 3b Subunit 1) gene are the most frequently found mutations in MDS. Circular RNAs (circRNAs) are covalently closed RNAs that are produced by back-splicing process. CircRNAs can regulate multiple biological processes through various molecular mechanisms, such as microRNA sponging. Their deregulation is frequently found in cancer. It is likely that they also contribute to the development of MDS, however, their role in MDS has not been researched yet. Therefore, our aim was to explore circRNA levels in MDS and analyze their association with patient prognosis. We further hypothesized that mutations in splicing factor genes can affect production of circRNAs and thus, we examined circRNA levels with respect to the mutational status. We explored transcriptome of 78 MDS patients, 7 AML patients, and 13 healthy donors using Illumina RNA-seq of total RNA isolated from CD34+ bone marrow cells. To associate circRNA levels with mutational status, Illumina TruSight Myeloid Sequencing Panel Kit examining 54 genes was applied. Of 8,620 circRNAs identified by RNA-seq, 204 circRNAs were deregulated in MDS (e.g., MENTRNL, EBF1, and PPM1L-derived circRNAs) and 246 circRNAs were altered between lower- and higher-risk patients (e.g., CHST15, TMTC2, and PDE3B-derived circRNAs). Most of the progression-related circRNAs (n = 234) showed elevated levels in higher-risk patients, suggesting that the back-splicing process might be stimulated during the disease progression. In MDS patients with SF3B1mutations, other 40 circRNAs were deregulated (e.g., ZNF91, ZEB1, and ZNF124-derived circRNAs). This circRNA profile was substantially different from the profiles associated with the rest of recurrently mutated splicing factor genes (SRSF2, U2AF1, and ZRSR2). To study alterations in forward- and back-splicing in SF3B1-mutated patients, we examined transcriptional differences on the levels of whole genes, transcript variants, and circRNAs and searched for specificities in transcription within individual gene loci. A set of circRNAs whose levels differed specifically without affecting expression of corresponding forward-spliced mRNAs included several oncology/hematopoiesis-relevant genes (e.g., ATM, CBL, ERCC5, ETV6, FLT3, and MAPK6). Gene loci with changed expression of both, alternative mRNA transcripts and circRNAs, but stable transcription on whole gene level included for example CDK14, KDM1A, and ZEB1. Because ZEB1(Zinc Finger E-Box Binding Homeobox 1) serves as an essential hematopoietic transcription factor, we focused on ZEB1-derived circRNAs (hsa_circ_0000228 and hsa_circ_0003793) in more detail. We demonstrated that upregulation of these circRNAs is SF3B1-specific and not related to any other clinical or molecular characteristics. Finally, using RNA-seq data from CRISPR/Cas9 edited K562 cells (Liberante FG, et al., Sci Rep. 2019; 9:2678), we confirmed that SF3B1K700E mutation leads to strong upregulation of ZEB1 circRNAs. To conclude, this is an early report showing for the first time that the levels of specific circRNAs are altered in MDS. We demonstrated that particular circRNAs may have potential to become markers that would contribute to more accurate prognosis of MDS patients. Further, we identified circRNAs with deregulated levels specifically in MDS patients with SF3B1mutation, suggesting that this mutation affects circRNA production.

Myelodysplastic syndromes (MDS) are a heterogeneous group of clonal hematopoietic stem cell (HSC) disorders characterized by dysplastic and ineffective blood cell production and a tendency for transformation to acute myeloid leukemia (AML). Because of the heterogeneity of MDS, the development of additional molecular tools able to refine the Revised International Prognosis Scoring System (IPSS-R), predict outcome and monitor the response to treatment would be highly beneficial.