Page 32 - Haematologica Vol. 110 - January 2025

REVIEW ARTICLE - Molecular pathogenesis and novel treatments for CMML L. Marando et al.
Epigenetic/splicing dysregulation in chronic myelomonocytic leukemia
Somatic mosaicisms form the structural background in >90% of CMML patients.16,17 Recurrently mutated genes include epi- genetic regulators (TET2 ~50% and ASXL1 ~40%), spliceosome components (SRSF2 ~50%) and cell signaling pathways (RAS ~30% and CBL ~15%)18-21 (Figure 2A). While these mutations are not exclusive to CMML and can be found across the spectrum of acute and chronic myeloid neoplasms as well as CHIP, specific mutations can define a dysplastic or a more aggressive proliferative phenotype (Figure 2B-D). In addition, studies in larger cohorts have highlighted that a pattern of TET2-SRSF2 (~46%) co-mutations,22 as well as biallelic TET2 mutations/alterations (biTET2, ~45%), are commonly identified in CMML, secondary to granulocyte-monocyte progenitor (GMP)-biased hematopoesis.23 These events oc- cur either when a secondary TET2 variant is subclonal to an ancestral TET2 variant or following loss of heterozygosity at 4q24.23 Notably, truncating TET2 variants are more likely to occur in the context of biTET2,23 highlighting that clonal se- lection for a complete loss of TET2 plays an important role in the evolution of CMML. At the protein level, TET2 plays a pleiotropic role in hematopoiesis, with mutations found in both myeloid and lymphoid malignancies. TET2 initiates DNA demethylation and regulates chromatin modifications at re- gions critical for lineage commitment and differentiation of HSC and progenitor cells.24 In murine models, loss of Tet2 (in particular, the catalytic domain25) induces a strong myeloid bias26 (Figure 3), with increased accessibility at enhancers of pro-myeloid differentiation genes thought to contribute to this phenotype.27 The strong myeloid bias induced by TET2 variants is also recapitulated in patients, as suggested by the observation that, at single-cell resolution, TET2 mutations mostly expand in the myeloid lineage.28
Variants in SRSF2, a component of the spliceosome ma- chinery, are the second most common ancestral variants in CMML, and also promote a myeloid bias (Figure 3). Mutations at the hotspot proline residue 95 (P95) alter messenger RNA (mRNA) splicing by changing the RNA binding affinity of SRSF2, leading to mis-regulation of exon inclusion.29 These variants, however, lead to only modest changes in global mRNA splicing. Even if a subset of mis-spliced transcripts has been proposed to be relevant for myelodysplastic syndromes,29-31 the molecular mechanisms whereby SRSF2 variants are implicated in MDS and CMML remain elusive. Using a condi- tional murine model of Srsf2P95H, it was shown that Srsf2P95H/+ native chimeras showed clonal expansion at the expense of wild-type HSC only when transplanted into lethally irradi- ated recipients without an external competitor, suggesting that the specific characteristics of a microenvironment that is aged matched to the HSC plays a crucial role in allowing this variant to establish clonal dominance.32 In addition, Srsf2P95H/+ mice were observed to develop monocytosis and dysplastic neutrophils on aging (~12 months after induction
of the recombinant allele), and eventually succumbed to a myeloproliferative disorder characterized by the presence of additional somatic mutations observed in CMML, including Ras pathway mutations.32
Among all the epigenetic drivers, truncating ASXL1 variants are associated with adverse outcomes in CMML, proliferative disease features (Figure 2B), and resistance to epigenetic therapies.5 Most ASXL1 variants are frameshift or nonsense mutations in exon 12 (last exon) resulting in truncation of the protein at the C-terminus and loss of the plant homeodomain (PHD).33 ASXL1 mutations have been shown to cause both loss of polycomb repressive complex 2 (PRC2)-mediated histone methylation (H3K27me3) at HOXA cluster genes,34 and loss of polycomb repressive complex 1 (PRC1)-mediated histone deubiquitination (H2AK119Ub).35,36
Given the complex ASXL1 interactions, our group attempted to further define the role of ASXL1 in CMML by interrogating the genome, transcriptome, and epigenome of wild-type ASXL1 versus truncating ASXL1-mutated primary CMML samples. We found that ASXL1-mutated patients gained accessibility at several enhancers enriched in ETS and BRD4 complex motifs, with upregulation in genes involved in cell cycle progression, DNA replication, and leukemogenesis.37
Signaling pathway mutations in chronic myelomonocytic leukemia
Signaling mutations are frequent in CMML and are usually associated with myeloproliferative features (so-called pro- liferative CMML).38 These are largely dominated by oncogenic RAS pathway mutations (>70%; NRAS, CBL, PTPN11, KRAS, NF1, BRAF) but also include JAK2V617F, FLT3 and CSF3R. The latter two are very infrequent, with the presence of an FLT3-internal tandem duplication usually heralding transformation to AML.39
RAS pathway mutations (NRAS, CBL, PTPN11, KRAS, NF1) Mutations in the epigenetic machinery are not specific to CMML, but are generally considered ancestral to signaling pathway mutations.38,40,41 Germline and somatic RAS pathway mutations are associated with juvenile myelomonocytic leu- kemia, an aggressive pediatric myeloproliferative neoplasm that resembles proliferative CMML.42 Features of proliferative CMML include constitutional symptoms, extramedullary he- matopoiesis and myeloproliferation. This contrasts with dys- plastic CMML, which is instead characterized by cytopenia(s) and a more indolent course (Figure 2B-D). Leveraging a large cohort of >1,000 CMML patients, our group demonstrated that oncogenic RAS mutations are more prevalent in, and occur at higher variant allele fractions in proliferative CMML than in dysplastic CMML.38 Interestingly, in murine models, NrasG12D has been shown to have a bimodal effect on HSC, both increasing and decreasing the rate at which some HSC divide; besides, NrasG12D can also increase reconstitution and self-renewal potential of HSC on serial transplantation.43
Haematologica | 110 January 2025
24