Page 188 - 2021_06-Haematologica-web

M. Martin-Izquierdo et al.
nosed with MDS eventually transform into sAML.4 Disease progression is associated with a dismal prognosis, partly because most of these patients are resistant to currently available treatments, and the long-term survival rate of treated patients is less than 10% after a couple of years.5-7
In recent years, new high-throughput genomic technolo- gies, such as next-generation sequencing (NGS), have enabled a large number of studies to elucidate some of the mechanisms involved in MDS pathogenesis such as epige- netic regulation, transcription, signaling pathways, splicing, cohesin complex, apoptosis and angiogenesis. However, MDS exhibit great genetic and clinical heterogeneity, so the nature of their pathogenesis is still not fully understood.8-10
The mutational dynamics and clonal evolution underly- ing disease progression have just begun to become clear. Previous studies have identified multiple genes recurrently mutated in MDS and sAML and these have provided insight into the great intratumoral heterogeneity typical of progression from MDS to sAML.8-13 These studies have shown that the evolution of the disease is a complex process involving new additional alterations co-existing with the MDS founder clone.14 Moreover, recent studies have described the association of mutations in genes such as TET2, RUNX1, ASXL1, and STAG2 with high-risk MDS, as well as the presence of mutations in genes activating sig- naling pathways, such as FLT3, PTPN11, NPM1, and NRAS, which are newly acquired in sAML and associated with faster progression.15-17 However, this complexity and the lack of large cohorts of serial samples means that molecular mechanisms of disease progression are only partly under- stood. Thus, longitudinal sequencing genomic studies are still required to determine which mutations or combina- tions of them are important in leukemic transformation.
In this study, we performed whole-exome sequencing (WES) and/or targeted deep sequencing (TDS) on serial samples from MDS patients who evolved to sAML (discov- ery cohort) before and after progression, as well as TDS on additional MDS patients who did not progress to sAML during follow-up (control cohort). The results were validat- ed in an independent series of MDS patients (validation cohort). Interestingly, we undertook an integrative analysis to determine the mutational dynamics of the pathways and genes and to identify how mutations, alone or in combina- tion, contribute to leukemic transformation. The study showed involvement of co-occurrence of alterations in the cohesin and Ras pathways in the MDS transformation to sAML, as well as a high proportion of newly acquired or increased clonal selection of mutations in the chromatin- modifier genes in MDS patients who received a disease- modifying therapy before their progression to sAML.
Methods
Study design
In order to study the mutational changes occurring during the evolution to sAML from a previous myelodysplastic phase, 486 samples from 437 patients were included in the study. The patient series was divided into three cohorts (Online Supplementary Figure S1): i) discovery cohort: a cohort of MDS → sAML progressing patients that included 42 patients diagnosed with MDS who pro- gressed to sAML; according to the study design, 84 BM serial patient-matched samples were collected and sequenced on two occasions with the first sampling, at initial presentation of the dis- ease (diagnosis, MDS stage), and the second sampling, after pro-
gression to sAML (disease evolution, leukemic phase); all samples were analyzed by a TDS strategy; furthermore, 16 of those pro- gressing patients (32 samples) were initially studied by WES; infor- mation about the treatment received before progression was avail- able for all 42 patients: azacytidine (n=16), lenalidomide (n=4) and no treatment or supportive care (n=22); ii) control cohort: a cohort of MDS non-progressing patients consisted of 14 BM paired sam- ples from seven MDS patients who did not progress to sAML after a minimum of 3-year follow-up for low-risk MDS (LR-MDS) and 1 year for high-risk MDS (HR-MDS) (median fol- low-up of 52 months; range, 20-89 months); according to the study design, the second sampling in this control cohort corresponded to a time when the disease was stable and TDS was performed on all these samples; iii) validation cohort: a cohort of 388 BM or PB sam- ples from patients suffering MDS at diagnosis and for which only one time-point (sample) was studied by TDS;. notably, 63 of these patients eventually evolved to sAML, while 325 had not pro- gressed to sAML after a median follow-up of 19.6 months. The main patient clinical characteristics are summarized in the Online Supplementary Table S1.
This research was performed in accordance with the Declaration of Helsinki guidelines, and was approved by the Local Ethics Committee (“Comité Ético de Investigación Clínica, Hospital Universitario de Salamanca”). All patients provided writ- ten informed consent.
Sequencing analysis
Whole-exome sequencing
WES was performed on matched diagnosis-progression samples from 16 patients of the discovery cohort. The mean coverage of WES was 77.6x (range, 36-124) and at least 73% of the captured regions had a coverage of 30x or more for all 32 samples (Online Supplementary Table S2). See the Online Supplementary Appendix for full details.
Targeted-deep sequencing
All genomic DNA samples underwent TDS using an in-house custom capture-enrichment panel of 117 genes previously related to the pathogenesis of myeloid malignancies (Online Supplementary Table S3). The mean coverage of TDS was 665x (range, 251-1,198) where 99.5% of target regions were captured at a level greater than 100x. See the Online Supplementary Appendix for full details.
Analysis of mutational dynamics
The main aim of this study was to analyze the mutational changes occurring between the first sampling (MDS stage) and the second sampling (stable disease/sAML stage) in the discovery and control cohorts. To this end, variant allele frequency (VAF) at these two stages were compared using two approaches: i) VAF ratio between second and first sampling, where thresholds of >1.2 and <0.8 were used to classify mutations as increasing or decreasing, respectively, while ratios between these thresholds were consid- ered to be stable; and ii) Fisher´s exact test where values of P<0.05 were taken to indicate statistically significant changes during pro- gression.
Results
Molecular landscape of the progression from myelodysplastic syndromes to secondary acute myeloid leukemia
In order to characterize the main cohort of the study, the discovery cohort, 16 patients (patients #27 - #45) were ana- lyzed by WES at the time of diagnosis and at leukemic
2216
haematologica | 2021; 106(8)