Comprehensive Characterization of Evolution of Genomic Complexity By Structural Variant and Mutational Profiling in Myelodysplastic Syndrome Patients with Hypomethylating Agent Failure

Introduction

MDS patients (pts) who develop resistance to hypomethylating agent therapy (HMA-F) have dismal outcomes. Genomic aberrations underlying progression to HMA-F remains poorly understood. By sequencing, studies have demonstrated acquisition of new mutations affecting transcription factors such as FLT3 and RAS-MAPK pathway genes in a subset of cases. Until now, evolution of genomic complexity in HMA-F by acquisition of new structural variants (SVs) has not been systematically investigated. This is due to lack of genome-wide technologies to quantitatively identify all types of SVs at a gene/exon-level resolution (deletions/insertions/inversions/fusions etc..). In this study, we optical genome mapping (OGM), a novel, non-sequencing-based approach for accurate high-resolution SV detection, combined with ultra-deep coverage targeted NGS to comprehensively analyze genome-wide changes affecting both SVs and somatic mutations using paired baseline and progression samples from 21 MDS pts who developed HMA-F.

Methods

We identified all MDS pts with sufficient (1.5 million) cryopreserved, viable bone marrow (BM) mononuclear cells at both diagnosis (baseline) and HMA-F. All pts underwent BM morphologic exam, traditional banding analysis (TBA), and 81-gene targeted NGS panel (2500X coverage; LOD: 2% VAF; myeloid genes). We excluded pts with AML-defining abnormalities by standard-of-care work-up (per WHO 2022 criteria) and CMML pts. OGM was performed using standard procedures: extraction of ultra-long DNA molecules, sequence-specific fluorophore labeling and stain, followed by loading on to nanochannels and imaging (Saphyr; Bionano Genomics). Rare Variant Pipeline was used for detection of clinically significant SVs [minimum 300X coverage; 5000 bp size cut-off). We used bidirectional paired genome-wide analysis for identifying unique SVs seen exclusively at either baseline or HMA-F.

Results

There were 21 pts [11 men; 10 women; median age: 73 (43-89)] with available cryopreserved cells for comprehensive mutational and SV profiling. Median BM blast% was 5 (1-18). The median time to HMA failure was 11 (6-21) months. All pts had at least 1 genomic aberration: 14 (66%) by TBA and 20 (95%) by NGS (median: 3 mutations per pt.). Seven (35%) pts had normal karyotype (NK), 6 (30%) pts had complex karyotype (CK). Most frequent gene mutations included TET2, ASXL1 TP53 [each seen in 7 pts, 33%], SRSF2 [6 (30%)], DNMT3A, RUNX1 and STAG2 (5 pts each, 25%).

Baseline OGM identified multiple clinically significant SVs, 6 of which, seen across 4 (19%) pts, were cryptic by TBA. These included CBFB::MYH1 and KMT2A partial tandem duplication. OGM at HMA-F showed increased genomic complexity compared to baseline in 9 (43%) pts. Paired analysis comparing genome-wide SV burden at HMA-F with matched baseline sample revealed 306 new unique SVs at HMA-F not seen at baseline ( Fig 1). These included 30 deletions (8 pts), duplications (6 pts), inversions, intrachromosomal and intrachromosomal translocations (5 pts each). Recurrent aberrations included segmental deletions of chr 17, 1, 7, 9, 12 and 21, affecting ETV6, CDKN1B, IKZF1, RUNX1, FHIT and HOXA11 genes, exon-spanning duplication in RUNX1, and FANCA::NF1 and YTHDF3-SETBP1 fusions. Three (14%) pts (2 with CK, 1 with NK), showed a cataclysmic increase in the numbers of SVs (translocations/ inversions) indicative of chromoanagenesis ( Fig 2). In contrast, there were 29 SVs seen exclusively at baseline and not detected at HMA-F [intrachromosomal translocations affecting chr 3 ( FOXP1) and chr 7 ( CUX1), low-level IGH::BCL2, deletions and insertions].

Sequential NGS and TBA was available in 20 pts. At HMA-F, 13 (62%) pts developed additional mutations, most frequent in NRAS, followed by TP53, TET2, CSF3R, KIT, second mutation in CEBPA, DDX41 and WT1 genes. The median number of mutations per pt. was 4. The highest increase in SV burden by OGM was noted in pts who did not acquire new mutations by NGS at HMA-F (5/7; 71%). Seven (33%) pts showed cytogenetic clonal evolution at HMA-F by TBA. OGM identified new SVs at HMA-F in 2 (50%) pts who did not show clonal evolution by NGS or TBA.

Conclusions

HMA-F MDS is characterized by increased genomic complexity by acquisition of new somatic mutations (NGS) and SVs (OGM), highlighting their complementary role in MDS progression. OGM further highlighted cataclysmic genomic changes by chromoanagenesis in a subset of MDS pts at HMA-F.