Myelodysplastic syndrome in the omics era : from genomic prognostication to mechanisms of resistance

Abstract

MDS is a myeloid neoplasm characterized by an ineffective hematopoiesis and it poses a predisposition to progression to acute myeloid leukemia. The frontline regimen for MDS includes hypomethylating agents such as azacitidine and decitabine but a sizable proportion of patients are either irresponsive to HMA or develop resistances during treatment. The classification system devised by World Health Organization (WHO) in 2016 is conventionally used and has been revised in 2022. It categorizes MDS based on cell morphology, bone marrow blast percentage, cytogenetic abnormalities, and the presence or absence of specific mutations. Similarly, the revised International Prognostic Scoring System (IPSS-R) is used as prognostic scoring tools based on age, basic blood counts, and number of cytogenetic abnormalities. However, small mutations are underemphasized in both tools. The mutation pattern of Asian MDS is also unclear with limited work focusing on the Asian population. In view of this, the current work investigated the mutation signature of Asian MDS using targeted sequencing and the molecular mechanism underlying HMA resistance using various omic technologies. Including Hong Kong and multiple centers across Asia, 1225 MDS specimens were collected and sequenced using a 54-gene panel. We unveiled unique signatures in different MDS subtypes under the WHO classifications and showed several hotspot regions in some of these genes. Harnessing the sequencing data generated from these patients, Asian MDS dataset was also evaluated against a molecular-based scoring system for MDS published recently by another research group, but risk group segregation was poor. Therefore, this study produced a new prognostic model using the Asian MDS dataset to predict overall survival, leukemia free survival and time to progression. A web calculator is also produced for public use. Additionally, germline mutations of seven additional genes were sequenced in a subset of 554 patients and searched for germline predispositions to MDS conventionally known to associate with MDS. As expected, many reported germline mutations were also detected in the Asian cohort including DDX41, ETV6 and GATA1/2¸ and they accounted for approximately half of the variants detected across the study cohort. In the context of studying HMA resistance, single cell sequencing was performed on serial time points of two patients who were refractory to HMA and other treatment regimens. In these patients, clonal SNP and INDELs were detected, and dynamical evolutions were observed longitudinally. Copy number gain of DNMT3A was also discovered. Interestingly, this mutation can be clonal concomitant with IDH2 and FLT3-ITD mutations or as an independent event from EZH2 clonal mutations. In the final multi-omic study using in-house derived resistant cells lines, PTEN was found silenced attributed to DNA hypermethylation during resistance in vitro and in patient. Unexpectedly, MDM2 hypomethylation while downregulation in vitro was observed. Using long read sequencing, this study unprecedentedly identified alternative RNA splicing of MDM2 in association with dysregulated RNA methylation. In conclusion, this study has developed a prognostic model for Asian MDS by profiling 1225 MDS patients. Multiple germline mutations have also been detected concordant with current knowledge. Finally, this study has identified multi-omic dysregulations during HMA resistance.