The changing mutational landscape of acute myeloid leukemia and myelodysplastic syndrome

Abstract

Over the past few years, large-scale genomic studies of patients with myelodysplastic syndrome (MDS) and acute myelogenous leukemia (AML) have unveiled recurrent somatic mutations in genes involved in epigenetic regulation (DNMT3A, IDH1/2, TET2, ASXL1, EZH2 and MLL) and the spliceosomal machinery (SF3B1, U2AF1, SRSF2, ZRSR2, SF3A1, PRPF40B, U2AF2, and SF1). The identification of these mutations and their impact on prognostication has led to improvements in risk-stratification strategies and has also provided new potential targets for the treatment of these myeloid malignancies. In this review, we discuss the most recently identified genetic abnormalities described in MDS and AML and appraise the current status quo of the dynamics of acquisition of mutant alleles in the pathogenesis of AML, during the transformation from MDS to AML, and in the context of relapse after conventional chemotherapy.

Implications: Identification of somatic mutations in AML and MDS suggests new targets for therapeutic development.

Figure 1

Role of proteins encoded by genes mutated in MDS or AML in DNA methylation. DNMTs catalyze the addition of a methyl group to cytosine. IDH enzymes are responsible for catalyzing the reversible conversion of isocitrate to α-KG and CO₂ in a two-step reaction. In turn, α-KG is required by the TET proteins to oxidize 5-mC to 5-hmC, an intermediate in DNA demethylation. Mutant IDH proteins metabolize α-KG to 2-HG, an oncometabolite that competitively inhibits the enzymatic activity of TET proteins.

Figure 2

MDS and AML mutations target spliceosomal genes associated with initial 3′ splice site recognition. Somatic mutations have been found in all major genes involved in 3′ splice site recognition and spliceosome commitment. This includes SF1, U2AF2, and U2AF1, which are critical for commitment complex E formation in U2-dependent splicing, as well as SF3A1, SF3B1, and ZRSR2, proteins that are essential for stabilization of splicing complex A. It is important to note that SRSF2 serves as a generalized enhancer of 3′ splice recognition facilitating splice site recognition through stabilization of U2AF1 interaction with the AG dinucleotide. Because SRSF2 functions through binding of exonic splicing enhancer (ESE) sequences it acts on only a subset of exons. ZRSR2, which is the only spliceosomal gene predominately targeted by inactivating nonsense mutations, is essential for U2AF-independent recognition of U12-mediated RNA splicing.

Figure 3

Schema representing a potential path of acquisition of mutant disease alleles leading to MDS, transformation to AML, and relapse after failure to respond to standard chemotherapy. Normal HSCs randomly acquire hundreds of nonpathogenetic mutations in a time-dependent fashion. At some point, a “driver” mutation is acquired (e.g., TET2) that confers a growth advantage to a “founding clone.” Previously acquired background mutations (i.e., “passenger”) migrate with the driver mutation and are present in all its progeny. Driver mutations disturb normal myeloid maturation and induce dysplastic features clinically compatible with MDS. Subsequent acquisition of mutant alleles in the founding clone (e.g., FLT3) may result in leukemic transformation (i.e., secondary AML). Patients with AML receive induction to remission cytotoxic chemotherapy followed by consolidation chemotherapy. A fraction of patients are cured upon eradication of the founding clone as well as “subclones” with a more complex mutational make-up than the former. However, a significant fraction of patients relapse after chemotherapy due to resistance and subsequent reexpansion of resistant clones. This sequence of events is driven by the acquisition of new mutations (e.g., RUNX1, ASXL1, and TP53) either in the founding clone or in a subclone of the founding clone as a consequence of therapy-induced selection pressure or as a result of DNA damage directly induced by cytotoxic chemotherapy. TP53 mutations have been associated with a phenomenon termed chromothripsis, which may occur at any point during the pathogenesis of AML. The latter induces local chromosome fragmentation resulting from DNA double-strand breaks, likely repaired by NHEJ. The net result is a conglomerate of complex chromosomal rearrangements that adds a new layer of genomic complexity to that provided by the numerous point mutations present in AML genomes and provides the genetic basis for tumor growth and resistance to therapy.