Background

Atypical Chronic Myeloid Leukemia (aCML) is an aggressive and genetically heterogeneous disease for which no standard of care exists. The classification of aCML, which is included in the group of Philadelphia-negative myeloid neoplasms, has been a matter of debate for years. While the identification of the translocation t(9;22)(q34;q11) in a patient with accumulation of mature granulocytes and their precursors is sufficient for the diagnosis of Chronic Myeloid Leukemia (CML) [1, 2], the absence of this translocation is pathognomonic of Philadelphia-negative Myeloproliferative Neoplasms. Given the concomitant presence of myeloproliferation and myelodysplasia, the 2002 World Health Organization (WHO) classification of myeloid neoplasms places aCML under the category called myelodysplastic/myeloproliferative neoplasms (MDS/MPN) [3] and the 2008 and 2016 revisions of WHO criteria did not change the classification [4, 5]. The MDS/MPN group includes chronic myelomonocytic leukemia (CMML), aCML, juvenile myelomonocytic leukemia (JMML), MDS/MPN with ring sideroblasts and thrombocytosis and MDS/MPN unclassifiable (MDS/MPN-U). According to the 2008 WHO classification of myeloid neoplasms and acute leukemia, the absence of BCR-ABL and PDGFRA, PDGFRB or FGFR1 rearrangements are minimal diagnostic criteria for aCML [4, 6]. However, the main feature characterizing aCML is the presence of neutrophilic leukocytosis and marked dysgranulopoiesis. Moreover, to fulfil the diagnostic criteria, the white blood count (WBC) should be ≥13 × 109/L with ≥10% of immature granulocytes and ≤20% blasts in the blood and the bone marrow [4, 6]. These diagnostic guidelines have been then applied in different studies that analyzed histopathological features and clinical data available for similar types of myeloid neoplasia like Chronic Neutrophilic Leukemia (CNL) and MDS/MPN-U. These reports confirmed that WHO criteria were really suitable to distinguish aCML from similar diseases [7,8,9,10,11]. For what concern patients’ treatment, no standard of care exists. Hematopoietic stem cell (HSC) transplantation is always the best option when a matching donor is available. Without this possibility, patients can be considered for treatment with general drugs like hypomethylating agents, pegylated-interferon-α, hydroxyurea, and/or erythropoiesis stimulating agents or for enrollment in clinical trials with specific inhibitors (the case of ruxolitinib and trametinib will be discussed later in this review) [12]. However, patients’ survival, which has been analyzed in different studies with some differences, remains dismal. In an Italian cohort of 55 aCML cases respecting the WHO criteria, the median overall survival was 25 months [13], while in an US study of 65 patients it was found to be 12.4 months [11].

Recurrent signaling pathways involved in myeloproliferation

A big effort has been made in the last decades to elucidate the molecular mechanisms leading to myeloproliferation. The identification of oncogenic mutations in signal transduction proteins pointed to the role of specific pathways in inducing excessive proliferation of myeloid lineages [14]. The subsequent development of mouse models carrying mutations found in patients and, conversely, the analysis of unexpected myeloproliferative phenotypes in genetically modified mice proved that the aberrant activation of these specific pathways plays a causal role in the onset of the pathology [15]. It came out that pathological myeloid proliferation is supported by few signaling pathways known to induce myelopoiesis by transducing signals from cytokines and growth factor receptors [16,17,18,19]. In this review we will primarily focus on three signal transduction pathways, the Janus kinase 2/signal transducers and activators of transcription (JAK2/STAT), the mitogen-activated protein kinase (MAPK) and the Rho associated coiled-coil containing protein kinase 1/2 (ROCK1/2) pathways. For all of them a role in myeloproliferation has been demonstrated by in vitro and in vivo studies and their involvement in human myeloproliferative diseases, including aCML, has been described [6, 14, 124]. It has been demonstrated that SEB impacts on AKT and MAPK pathways, responsible for cell proliferation and survival [125]. In particular, SEB binds to the SET nuclear oncoprotein protecting it from protease cleavage. In turn, SET represses PP2A activity [126, 127] that inhibits AKT and MAPK pathways. When SEB is mutated, it accumulates in the cells and, through SET, decreases PP2A activity, leading to increased cellular proliferation [117].

PTPN11 gene encodes for SHP2 (Src-homology-2 domain containing protein tyrosine phosphatase), a protein tyrosine phosphatase (PTPase) acting downstream to growth factor receptors. Mutations in the PTPN11 gene result in constitutively activated RAS. In fact, when SHP2 is mutated it activates guanine nucleotide exchange factors (GEFs), necessary for the conversion of GDP-RAS into GTP-RAS [14, 128]. Interestingly, SHP2 is phosphorylated by JAK1 and JAK2 and the phosphorylated form of SHP2 binds to GRB2 and activates RAS [129]. Moreover, JAK2, PTPN11 and RAS mutations were identified as mutually exclusive in MDS, suggesting their participation to the same pathway [130]. Given the central role of RAS mutation in MPN and the convergence of SETBP1, PTPN11 and JAK2 encoded proteins on MAPK pathway overactivation, patients carrying mutations in these genes could benefit from treatment with MEK inhibitors.

However, a number of genes mutated in aCML encodes for biosynthetic enzymes, transcription factors and epigenetic modifiers. These proteins are apparently unrelated with the signal transduction molecules previously discussed and their exact role in the onset of the pathology is still unclear.

ETNK1, for example, encodes an ethanolamine kinase (EKI 1) which phosphorylates ethanolamine to phosphoethanolamine in the phosphatidylethanolamine biosynthesis pathway. Two recurrent point mutations impairing the catalytic activity of the kinase have been described in ETNK1 gene in aCML [131]. The phosphatidylethanolamine biosynthesis pathway is involved in many biochemical processes like definition of membrane architecture, anchoring of proteins to the plasma membrane, mitochondria biogenesis, autophagy and progression to cytokinesis during cell division [116, 132, 133]. Due to the fact that EKI 1 contributes to different processes in the cell, the mechanisms by which the mutant protein induces myeloproliferation have not yet been clarified.

RUNX1 encodes the alpha subunit of the core binding factor (CBF) complex. This complex activates and represses transcription of genes involved in growth, survival and differentiation pathways in hematopoietic cells, maintaining the proper balance among different lineage progenitors [134]. This gene is recurrently mutated in a variety of hematological malignancies due to chromosomal translocations and somatic mutations. Mono- and biallelic RUNX1 mutations have been described in aCML [14]. Some mutations cause inactivation of the protein, while others induce a dominant negative activity [135]. However, the mechanism through which the mutant RUNX1 induces myeloid expansion is still to be understood.

The TET dioxygenases, TET1, TET2, and TET3, catalyze the transfer of an oxygen atom to the methyl group of 5-methylcytocine (5-mC), converting it to 5-hydroxymethylcytocine (5-hmC) [136, 137]. This modification, in turn, promotes locus-specific reversal of DNA methylation, impacting on DNA methylation landscape [138]. TET2 is frequently mutated in both myeloid and lymphoid malignancies [14, 122, 139,140,141,142] resulting in a wide hypermethylation phenotype [143], but, again, the precise pathways responsible for the phenotype downstream this global genome alteration have not been dissected. The hypomethylating agent decitabine, approved by FDA for the treatment of MDS and CMML, have been tested in aCML patients (regardless of TET2 mutational status) with some positive results, even if on small cohorts of patients, and deserves better investigations [144,145,146,147].

ASXL1 (Additional of sex combs-like 1) plays a role in the recruitment of the Polycomb Repressive Complex 2 (PRC2) to its target sequences and takes part in the complex involved in deubiquitination of histone H2A lysine 119 (H2AK119) [148, 149]. Mutations of the gene, identified in patients with AML, MPN and MDS, are associated with loss of ASXL1 expression [148]. Changes in the cell following ASXL1 mutations include: loss of PRC2-mediated gene repression, global loss of H3K27 trimethylation (H3K27me3) and derepression of the posterior HOXA cluster genes, including HOXA5–9, known to play a role in leukemogenesis [148].

All these proteins have in common a functional pleiotropy, since they can modify the expression of hundreds of genes or the functionality of many proteins in the cell. However, it is conceivable that, among the several deregulated events and pathways, few are responsible for leukemogenesis. In this view, it would be very useful to analyze the signaling pathways known to play a role in myeloproliferation in these mutational contexts in the final attempt to exploit targeted therapies with available inhibitors. Moreover, given that two or more mutations often occur simultaneously in aCML patients [119] combination therapies with different inhibitors seems, at least in theory, a promising approach.

Recently, two studies demonstrated that the percentage of healthy people showing clonal expansion of somatic mutations associated with hematologic diseases increases with age. The authors found that clonal haematopoiesis frequently involves DNMT3A, TET2, and ASXL1 mutant cells. Of note, somatic mutations were found to be associated with increased risk of hematological malignancies, as well as other adverse events [150, 151]. It will be tempting to envisage specific strategies for the prevention of the disease based on the mutations arising during the precancerous phases, however the predictive power of mutant hematopoiesis is low and additional biomarkers are needed to justify pharmacological intervention [150, 151].

Conclusions

aCML is a rare hematological disease for which no standard of care exists. NGS techniques have allowed in the past few years to highlight mutations in signal transduction proteins but also in proteins with pleiotropic functions, like transcription factors and chromatin-modifying enzymes [14]. These proteins may regulate the expression of thousands of genes simultaneously, deeply altering cell physiology. However, the precise mechanisms by which they induce and sustain tumorigenesis are still elusive. In particular, it is not known whether a single gene or a specific subgroup of genes controlled by these enzymes are responsible for cell transformation and through which mechanism. It is conceivable that wide alteration in gene expression could impact on the specific signal transduction pathways regulating proliferation and survival in haematopoietic cells. However, a wide analysis of signal transduction alterations in the different mutational contexts is still missing. This information will help to identify new therapeutic approaches in genetically defined subsets of diseases, but also to successfully repurposing existing drugs. As discussed in this review, JAK2, MEK and ROCK inhibitors might represent a treatment option for aCML patients. However, apart from encouraging preclinical studies and case reports, we still need multicenter randomized trials to test the potential benefits of these treatments in large cohorts of patients.