In 1987 a mutation within codon 12 of the NRAS gene was reported for the first time by Janssen et al. in a patient with CMML in a study investigating molecular alterations of RAS genes in a variety of preleukemic disorders and leukemias of myeloid origin (13). Subsequently is has been shown that RAS mutations are rare events in BCR/ABL negative chronic myeloid leukemia (CML) but are prevalent in CMML (14). In this study mutations in the RAS oncogene were found in 17 of 30 (57%) CMML patients. In the last years the molecular landscape in patients with CMML has been described by several groups. Molecular abnormalities can be seen in >90% of patients with CMML (15) with a marked heterogeneity among CMML patients. A large number of gene mutations in genes encoding epigenetic regulators (TET2, ASXL1, EZH2, UTX, IDH1, IDH2, DNMT3A) (9, 16–22) splicing factors (SF3B1, SRSF2, ZRSF2, U2AF1) (23, 24), and cytokine signaling molecules (NRAS, KRAS, CBL, JAK2, FLT3) have been reported (9, 25–29). Mutations in the transcription regulators RUNX1, NPM1, and TP53 have also been found in CMML (9, 30, 31). Table 1 shows the frequencies of gene mutations in 3 different CMML cohorts in which comprehensive molecular analyses has been reported (9, 11, 32). Considering all molecular data reported mutations in TET2 (~60%), SRSF2 (~50%), ASXL1 (~40%) and RAS pathway (~30%) are most common (15) but no molecular aberration is specific of this entity, as they can be detected with different frequencies in other myeloid neoplasms (33).

In 1988 Geissler et al. have originally reported extensive in vitro formation of myelomonocytic colony forming units (CFU-GM) without addition of exogenous growth factors in a subset of patients with CMML ( Table 2 ) (34). This spontaneous CFU-GM colony formation in CMML was markedly reduced by addition of anti-granulocyte/macrophage colony-stimulating factor (GM-CSF) antibodies, but not by antibodies against other growth factors, suggesting that this is a GM-CSF-dependent in vitro phenomenon (35) Figure 1 . The biologic basis for this observation was later provided by Padron when he reported hypersensitivity of CMML progenitors using phospho-STAT5 flow cytometry (36). Moreover, the group in Vienna could show in a small retrospective study that CMML patients with high spontaneous CFU-GM growth (>100/10⁵ PB mononuclear cells) have an inferior prognosis as compared to patients with low myeloid colony formation suggesting a clinical significance of the original observation (37). These results have been recently extended in a much larger CMML patient cohort indicating that spontaneous myeloid colony-formation was, compared to other single established prognostic factors, the strongest predictor regarding overall survival (OS) (38). This may indicate that in vitro cultures using unmanipulated mononuclear cells (MNC) may be a more global test that covers different aspects of malignancy better than any of the single parameters that are currently used to characterize the behavior of a tumor.

There is also another in vitro phenomenon that seems to be characteristic for CMML patients. Semisolid in vitro cultures from PBMNC of normal individuals usually contain a higher concentration of erythroid colonies (BFU-E) as compared to myeloid colonies (CFU-GM). Skewed differentiation toward the myelomonocytic over erythroid commitment in patients, as indicated by an inverse BFU-E/CFU-GM ratio, is a common finding in CMML patients (39). Interestingly, the lack of myelomonocytic skewing separated patients with a particularly favorable prognosis and a minimal risk of transformation.

Myelomonocytic leukemias can be recapitulated by transplantation of mouse BM cells carrying an oncogenic mutation in the Nras locus (40–42). Interestingly, alterations of the other RASopathy genes including Kras, Cbl, Ptpn11 and Nf1 may also result in a similar phenotype in preclinical mouse models ( Table 3 ) (44–47). In all these in vivo-models animals develop a myeloproliferative disorder with clonal expansion of the granulomonopoiesis.

The effects of molecular aberrations in genes of the epigenetig machinery have been also studied in preclinical animal models (48–50). In a mouse model with complete functional deletion of Tet2 resulted in a progressive enlargement of the hematopoietic stem cell compartment and eventual myeloproliferation in vivo. Tet2 +/- mice displayed increased stem cell self-renewal and extramedullary hematopoiesis, indicating that Tet2 haploinsufficiency contributes to hematopoietic transformation in vivo (48). Importantly, one third of Tet2 -/- and 8% of Tet2 +/- mice died within 1 year of age because of the development of myeloid malignancies reminiscent of CMML indicating that Tet2 loss may represent a predisposition for the development of this malignancy. Moreover, it was shown that transplantation of Tet2 -/-, but not wild-type (WT) or Tet2 +/- BM cells, was associated with elevated white blood cell (WBC) counts, monocytosis and splenomegaly in WT recipient mice (49).

Recent evidence suggests that considering cancer only as a consequence of genetic aberrations is too simple (77). There is growing evidence that the complex nature of transformation from a normal to a cancer cell within different tissues is a result of the interplay among genetic and epigenetic events, tissue structure, exposure and the tissue microenvironment. Thus, molecular analysis of a tumor by NGS alone may be not sufficient to cover the biology of a tumor and emphasize the need for more comprehensive methods to characterize the biology of a tumour. By combining structural data, functional in vitro findings, an in vivo preclinical models a comprehensive view of pathogenesis of CMML is possible.

Similar to the in vitro phenomenon of spontaneous erythroid colony (78) and megakaryocyte colony formation (79) due to hypersensitivity to growth factors in patients with BCR/ABL negative MPN spontaneous myeloid colony formation seems to be an in vitro feature in a subset of patients with CMML. Molecular aberrations of RASopathy genes in murine hematopoietic cells induce growth-factor-independent CFU-GM formation in vitro due to hypersensitivity of granulomonocytic precursors to GM-CSF (40, 41, 43, 44, 46, 47). Moreover, in juvenile myelomonocytic leukemia (JMML) in which molecular aberrations are mainly restricted to RASopathy genes autonomous myeloid colony formation due to GM-CSF-specific hypersensitivity is a hallmark feature of disease, which has been included in the diagnostic criteria (80). In a small series of CMML patients who had in vitro cultures and molecular analyses Geissler et al. observed a close correlation between high spontaneous myeloid colony growth and the presence of RAS pathway mutations as shown in Figure 2 (81). This initial observation was later confirmed in a larger patient cohort including 100 CMML patients (82). The median number of spontaneously formed CFU-GM/10⁵ MNC was 147.5 in RAS-mutated patients as compared with 2 in RAS-wildtype patients (p<0.00001). Unstimulated myeloid colony formation in RAS-mutated patients was also much higher than spontaneous formation of CFU-GM in normal individuals (median 4.8/10⁵ PBMNC) which has been reported by this group previously (83). There was no significant difference regarding spontaneous CFU-GM formation in CMML patients with molecular aberrations in genes of epigenetic regulation and RNA-splicing, respectively. High spontaneous myeloid colony formation was also never observed in CMML patients with the JAK2 V617F mutation as the only molecular aberration in signaling pathways [Geissler et al., unpublished data]. All these findings, in mouse and human, indicate that hypersensitivity to GM-CSF, as manifested by growth factor-independent CFU-GM growth in vitro, is caused by molecular aberrations of the RAS-pathway which may be a major driver in CMML pathogenesis, in particular in MP-CMML. Moreover, it reveals high autonomous CFU-GM growth as a functional surrogate parameter of RAS pathway hyperactivation in CMML.

Myelomonocytic skewing has been proposed as a key phenomenon in the pathophysiology of CMML. In a seminal paper using mutation-specific discrimination analysis of single-cell-derived colonies in 28 patients with CMML, Itzykson et al. could show that the main features of this disease are early clonal dominance, arising at the CD34+/CD34- stage of hematopoiesis, and granulomonocytic differentiation skewing of multipotent and common myeloid progenitors (84). Geissler et al. could demonstrate that myelomonocytic skewing as determined by semisolid cultures can separate subgroups of CMML patients with a different phenotype, a different genotype and a different prognosis (39). The definitive link of this phenomenon to the pathophysiology of CMML comes from animal studies in which hematopoietic cells are genetically manipulated with molecular aberrations that are commonly found in CMML patients. Functional inactivation of TET2 in cord blood CD34+ cells skews progenitor differentiation toward the granulomonocytic lineage at the expense of lymphoid and erythroid lineages (85). In mice, deletion of Tet2 results in an increased hematopoietic repopulating capacity with an altered differentiation skewing towards the granulomonocytic lineage (49). Other epigenetic modifiers such as ASXL1 have also been shown to impact skewing of hematopoiesis. Asxl1 -/- mice had a reduced hematopoietic stem cell (HSC) pool, and Asxl1 -/- HSCs exhibited decreased hematopoietic repopulating capacity, with skewed cell differentiation favoring granulocytic lineage (50). Furthermore the splicing factors SRSF2 and U2AF1 seem to affect skewing. Mutations in both SRSF2 and U2AF1 are associated with abnormal differentiation by skewing granulomonocytic differentiation towards monocytes (86). Collectively, many molecular aberrations that can be found in CMML, induce myelomonocytic skewing in the preclinical mouse model providing the genetic basis for this key finding in patients.

Recent molecular analyses of large populations have indicated that somatic mutations in hematopoietic cells leading to clonal expansion are commonly acquired during human aging (87). Clonally restricted hematopiesis is associated with an increased risk of subsequent diagnosis of myeloid neoplasia. As some of the genes frequently mutated in age-related clonal hematopoesis such as TET2 and ASXL1 are also commonly mutated in CMML and aged hematopoiesis is characterized by a myelomonocytic differentiation bias it was recently hypothesized that CMML and aged hematopoiesis may be closely related (88). Analyses of the somatic mutation landscape of CMML by whole exome sequencing followed by gene-targeted validation indicated that most CMML patients (71%) had mutations in >2 age-related clonal hematopoiesis (ARCH) genes and 52% had >7 mutations overall (89). A higher mutation burden was associated with inferior survival. Considering age-adjusted population incidence and ARCH mutation rates one may speculate that CMML represents the leukemic conversion of the myelomonocytic-lineage-biased aged hematopoietic system.

There is now increasing evidence that mutations in TET2 are in fact an initial clonal driver in CMML (88, 90). This view is based on the high frequency (60%) of these mutations in CMML patients (9, 19, 21, 22), the fact that TET2 mutated clones can be detected in a small fraction of older subjects with clonal, but non-leukemic hematopoiesis (90–93), the competitive advantage of murine and human HSC invalidated for TET2 (48, 85) and the results of single-cell clonal tracking experiments indicating that a TET2 mutation, when present, is often the earliest recurrent genetic event in CMML (84). According to data from this study the preferred order of mutational accumulation is TET2 (or IDH1/2) or ASXL1 (EZH2) first, followed by molecular aberrations in spliceosome components (SRSF2, SF3B1, U2AF1, or ZRSR2). Mutations in the RAS-signaling pathway seem to be rather late events which induce GM-CSF hypersensitivity and myeloproliferation.

Around 20% of CMML patients progress and transform to AML. Although the mechanisms behind are not known in detail, available data suggest that molecular aberrations in chromatin modelling as well in cell signaling may contribute to this process. Among genes of the epigenetic machinery, ASXL1 may have the most important impact on transformation. The ASXL1 gene regulates chromatin by interacting with the polycob-group repressive complex proteins (PRC1 and PRC2). Abdel-Wahab et al. reported that ASXL1 mutations resulted in loss of PRC2-mediated H3K27 trimethylation (94). In a study by Itzykson et al. in which the prognostic impact of different molecular aberrations in CMML patients was studied, only ASXL1 mutations retained their significant impact on AML-free survival in the multivariate analysis indicating the major role of this molecular aberration in the transformation process (9). Of these, only nonsense and frameshift ASXL1 mutations have been shown to negatively impact OS. The impact of mutations of RAS-pathway components on progression/transformation of CMML is more complex. The first indication of a potential role of NRAS aberrations in CMML evolution has been reported, at the molecular level, by Ricci et al. (95). In this study molecular analyses have been performed in 22 MD-CMML patients and in 18 MP-CMML patients. MP-CMML patients had a higher frequency of RAS mutations compared with MD-CMML. In two patients who progressed from MDS-CMML to MP-CMML, allele specific PCR showed low levels of the RAS mutations at the time of myelodysplastic disease which became predominant in the myeloproliferative phase, documenting for the first time the expansion of a RAS mutated clone in concomitance with CMML evolution. Other studies have confirmed that the MPN phenotye of CMML is a disease phase significantly associated with hyperactivation of the RAS-pathway. In a study reported by the Austrian study group MP-CMML as compared to MD-CMML patients had higher circulating blasts, LDH, RAS-pathway mutations, more often splenomegaly and higher growth-factor-independent myeloid colony growth in vitro (12). Recently, genetic differences were assessed between subtypes in 973 molecularly annotated Mayo Clinic-GFM-Austrian CMML patients. In this analysis NRAS mutations alone did not reach statistical significance as an independent factor impacting AML-free survival, however, the combined oncogenic RAS-pathway category including NRAS, KRAS and CBL was statistically significant in a model that only included genetic factors (74). Considering the fact, that spontaneous colony formation in CMML functionally covers the most frequent RASopathy gene mutations (38) these data are in line with findings in a small study which have been previously reported. In this study patients with CMML growth-factor-independent colony formation after transformation was significantly increased as compared to CFU-GM growth before transformation (37). Furthermore, a correlation of RAS-pathway mutations and spontaneous myeloid colony growth with progression and transformation could be demonstrated in a retrospective analysis of 337 CMML patients (96). Moreover, recent preclinical models also suggest that activating Nras mutations and somatic loss-of-function mutations in Tet2 exert cooperating effects and accelerate disease progression (97, 98). Altogether, these findings suggest that oncogenic RAS-pathway activation is a phenomenon associated with the MP-CMML phenotype, progressive disease and with transformation to AML.

Etoposide (VP16) is a DNA-damaging molecule by inhibition of topoisomerase. Preliminary reports suggested that etposide could give good results in CMML, with true complete responses in some cases and in improvement rather than worsening of cytopenias (51) ( Table 4 ). Hydroxyurea (HU), a potent ribonucleotide reductase inhibitor, acts as an S-phase-specific agent with inhibition of DNA synthesis. In a randomized phase III trial in patients with proliferative CMML, HU was more effective and achieved faster response than cytotoxic chemotherapy with VP16 (52). Interestingly, this study remains up to now the only randomized trial in a pure CMML patient population which demonstrated a survival benefit. Allogeneic stem cell transplantation which is the only curative therapy is rarely feasible because of age and/or comorbidities. While unresponsiveness to aggressive chemotherapy is a characteristic for most CMML patients, there may be subgroups that might benefit from more intensive chemotherapy. Although the presence of an NPM1 mutation, in contrast to AML patients, is an inferior prognostic parameter in CMML, CMML patients with this molecular aberration have shown relatively high response rates in a retrospective analysis (112).

More targeted treatment with cytotoxic drugs can be expected by more detailed immunophenotypical characterization of surface proteins on CMML stem cells which could be used as potential target structures for antibody-drug conjugates (113). One example is the use of the IL-3 receptor as target structure for tagraxofusp, a CD123-directed cytotoxin consisting of human IL-3 fused to truncated diphtheria toxin. This antibody-drug conjugate has shown impressive activity in blastic plasmacytoid dendritic-cell neoplasm (BPDCN) that overexpresses CD123 (114). In an early clinical trial in patients with relapsed/refractory CMML 80% (8/10) of the patients receiving tagraxofusp showed ≥50% reduction in splenomegaly and three patients achieved bone marrow CR (53).

It is important to note that the approval of the hypomethylating agents (HMA) azacitidine and decitabine (DEC), respectively, was originally based on MDS studies which included only few patients with CMML. In a phase III clinical multicenter trial of 358 MDS patients including 11 patients with dysplastic CMML the median overall survival was 24.5 months in the azacitidine (AZA) group as compared to 15.0 months in the conventional care group leading to the FDA approval of AZA for this subtype of CMML (54). The approval of decitabine (DEC) for CMML was also based on a phase III clinical trial of 170 patients with MDS, 14 of them with CMML (55). The ORR was significantly higher in the DEC group versus supportive care (17% vs. 0%, p < 0.001), but the median OS was not significantly different between the two arms. Additional phase II studies confirmed the efficacy of hypomethylating agents in all subtypes of CMML and, therefore, these agents are considered commonly as standard of care for higher risk CMML ( Table 3 ) (56–63). In the largest retrospective study with a pure CMML cohort patients were treated with AZA (n = 56) and DEC (n = 65) (115). The ORRs were 41% by the IWG MDS/MPN response criteria (AZA-56%, DEC-58%), with CR rates of <20% for both agents. No significant differences in response rates were seen between MP-CMML and MD-CMML. Similar findings were reported in a smaller, prospective phase II Italian study, with 43 CMML patients receiving DEC (63). The ORR after 6 cycles was 47.6%, with seven CRs (16.6%), eight marrow responses (19%), one partial response (2.4%) and four hematological improvements (9.5%). After a median follow-up of 51.5 months, median OS was 17 months, with responders having a significantly longer survival than non-responders. Despite some efficacy of HMA in CMML patients one has to keep in mind that this treatment does not alter mutational allele burden and disease biology (116).

Proof of efficacy but greatly variable response with HMA provide the rationale for searching biomarkers that predict response. Differentially methylated regions of DNA have been shown to separate DEC responders from non-responders by Meldi (117). Other predictors for response to HMA treatment were reported by Duchmann et al. (118). In a retrospective analysis of 174 CMML patients treated with HMA multivariate analysis mutations in ASXL1 predicted lower ORR, and RUNX1 mutations and CBL mutations predicted inferior OS. The combination of TET2 mutation and ASXL1 wildtype predicted higher CR and better OS. A multicenter retrospective study including 949 non-selected, consecutive CMML patients investigated whether HMA provide a benefit in subgroups of CMML patients (119). Adjusted median OS for patients treated with HU versus HMA was 15.6 months as compared to 20.7 months; (p=0.0002). In patients with MP-CMML, median OS was 12.6 months as compared to 17.6 months; (p=0.0027) for patients treated with HU versus HMA. HMA were not associated with an OS advantage for patients classified as having lower-risk disease (ie, MD-CMML with <10% blasts, CMML-0, or lower-risk CPSS). Considering all the caveats of a retrospective nonrandomized trial these data suggest HMA as the preferred treatment for patients with higher-risk CMML and those with MP-CMML. A recent European multicenter randomized phase III trial evaluating DEC +/- HU versus HU in advanced MP-CMML, however, did not show significant differences in outcome. Although HMA definitively play an important role in the management of CMML patients the need for newer, rationally derived therapies is apparent (120).

TET2 enzymes have been shown to provide a homeostatic link between intracellular metabolism and epigenetic gene regulation (121). These evolutionary conserved dioxygenases play a key role in the conversion of 5-methyl-cytosine (5-mC) to 5-hydoxymethyl-cytosine (5-hmC). TET dioxygenases require alpha-ketoglutarate, oxygen, Fe(II), and ascorbate for optimal activity (122). Isocitrate dehydrogenase (IDH) is a key enzyme for cellular respiration in the tricarboxylic acid (TCA) cycle. IDH mutations found in malignancies block normal cellular differentiation and promote tumorigenesis via the abnormal production of the oncometabolite 2-hydroxy-glutarate (2-hG). Recently, two inhibitors targeting IDH2 and IDH1 gene mutations, have become important components in AML management since molecular aberrations of IDH genes can be found in 20% of patients AML (123, 124). Although mutations involving IDH1 and IDH2 are uncommon in CMML (1% and 5-10%, respectively) IDH1/2 inhibitors are likely to present therapeutic options for these patients.

Loss-of-function mutations in TET2 occur in around 60% of CMML patients and are considered mutually exclusive with IDH1/2 mutations. Recently there has been accumulated significant preclinical evidence suggesting that ascorbate can restore dysfunctional TET2 activity. Agatocleous et al. generated mice lacking Gulo, the enzyme responsible for ascorbate synthesis. The resulting phenotype resembled mice carrying a homozygous Tet2 deletion (48). Indeed, ascorbate-depleted stem and progenitor cells showed decreased levels of 5-hmC, predominantly mediated by reduction of Tet2 function (125). On the other hand treatment with vitamin C mimicked Tet2 restoration in a reversible transgenic RNAi mouse model as described by Cimmino (64). Low ascorbate levels have been demonstrated in a subgroup of patients with hematologic malignancies (126). Although no beneficial effects of vitamin C intake regarding leukemia development have been seen in previous reports, these new preclinical data show that the possible impact of supra-physiological concentrations of vitamin C on leukemogenesis remains an interesting treatment concept, particular in CMML-patients harboring a partial or complete loss of TET2 function. In fact the is a current phase II trial which studies the effect of ascorbate in combination with AZA in patients with newly diagnosed AML, MDS, MDS/MPN with TET2 mutations (NCT03397173).

Geissler et al. have shown that growth factor-independent in vitro myeloid colony formation by CMML cells can be inhibited by the addition of anti-GM-CSF antibodies, but not by addition of antibodies against IL-6, Il-3, or G-CSF indicating that GM-CSF signaling may play an important role in the pathophysiology of CMML (35). Because of its cytokine synthesis-inhibiting effects IL-10 was studied on CMML cell growth in vitro. The addition of IL-10 revealed a profound and dose dependent inhibitory effect on spontaneous in vitro growth of CMML cells (35). It was shown that IL-10 induced suppression of CMML cell proliferation was associated with reduced GM-CSF production by leukemic cells, both at the mRNA and protein level. Therefore it was concluded that the inhibitory effect of IL-10 in vitro is most likely through suppression of endogenous GM-CSF release. Based on these findings a small pilot trial was initiated in which five patients with CMML were treated with 4µg/kg/day recombinant human IL-10 sc for 1 month and with 8 µg/kg/day for another month (65). Although no meaningful effects of IL-10 treatment was seen on the WBC counts in any of the five patients, one out of three patients with histologically confirmed skin infiltration markedly improved during IL-10 therapy. IL-10 administration was associated with a decline in lysozyme serum levels, a biomarker of the monocytic cell lineage, and downregulation of CD86 which has been shown to be upregulated by GM-CSF and downregulated by IL-10 in vitro. Interestingly, the clinical impact of IL-10 war recently supported by a study in which cytokine profiles were analyzed using cryopreserved PB plasma samples from 215 CMML patients (127). CMML patients with decreased IL-10 expression were found to have a poor OS when compared to CMML patients with increased IL-10 expression (P = 0.017), even when adjusted for other prognostic features including ASXL1.

Lenzilumab is a monoclonal antibody with high affinity for human GM-CSF. Based on data showing that anti-GM-CSF antibodies significantly inhibited the growth factor independent myeloid in vitro colony formation from primary CMML patient samples (35) and a study reporting that 90% of primary CMML samples demonstrated GM-CSF-dependent STAT5 hypersensitivity (36) lenzilumab was studied in CMML patients. In this early clinical trial of 15 CMML patients the antibody was well tolerated and effective with a durable ORR of 33.3% (66).

The clinical management of FLT3-mutated AML has been changed by the development of FLT3 inhibitors such as midostaurin and gilteritinib which are now in use in the frontline and relapsed/refractory settings in patients with a FLT3 mutation (128, 129). FLT3 aberrations have been reported in 1-3% of CMML patients (9, 32). Although these aberrations are uncommon in CMML FLT3 signaling may also occur in wildtype FLT3 malignancies. Thus, mice carrying a mutation in the RING finger domain of c-CBL develop a myeloproliferative disease involving hematopoietic progenitors that show increased FLT3 signaling (67). The incidence of molecular aberrations of the CBL gene has been reported from 10-14% (9, 11, 32) and thus is more common than that of the FLT3 gene. Therefore, CMML patients with mutations in the CBL gene could be potential candidates for studies with FLT3 inhibitors. In an ongoing phase I/II trial the FLT3 inhibitor quizartinib in combination with AZA is investigated in patients with untreated or HMA-refractory MDS, MDS/MPN with FLT3 or CBL mutations (NCT04493138).

Mutated RAS proteins have been deemed “undruggable” for a long time due to their high affinity for GTP and lack of accessible binding pockets. However, the discovery by Ostrem et al. of compounds that covalently bind to the switch II pocket of KRAS G12C provided the rationale for the development of inhibitors suitable for clinical testing (130). At the moment this concept does not play an important role in the treatment concepts for CMML, since the KRAS G12C mutation is extremely rare in CMML.

RAS proteins require post-translational farnesylation by the enzyme farnesyltransferase to become functionally active. Therefore, inhibitors of this enzyme have been considered as potential candidates for RAS-pathway inhibition. In a clinical phase III trial 85 patients with newly diagnosed JMML, a RAS pathway driven disease, were enrolled between 2001 and 2006 (131). 47 patients received the FTI tipifarnib alone in a phase II window before proceeding to HSCT. Tipifarnib as a single agent was safe and achieved a response rate of 51%, but failed to reduce relapse rates or improve long-term overall survival in the phase III trial. In a preliminary report of a phase II trial in CMML patients tipifarnib was well tolerated, however, had only limited efficacy (132).

The elucidation of the RAS/MEK/ERK signaling pathway in regulating cell proliferation has stimulated the development of selective MEK inhibitors (MEKi). These molecules represent promising therapies for RAS-driven neoplasias and RASopathies associated with hyperactivated RAS signaling. Preclinically, the MEKi PD0325901 was highly effective in reversing the CMML-like phenotype in a Kras G12D and in a NF1 -/- mouse model (68, 69). In a phase II study in patients with in Neurofibromatosis 1 (NF1) which is a prototypic RASopathy the MEKi selumetinib resulted in at least 20% reduction in the size of plexiform neurofibromas (pNF) from baseline in 71% of patients and was associated with clinically meaningful improvements (133). On the basis of this clinical benefit, selumetinib received FDA approval for children 2 years of age and older with inoperable, symptomatic pNF. In another phase II trial trametinib, another MEKi, was studied in patients with relapsed/refractory leukemias (70). Cohort 1 included patients with relapsed/refractory AML or high-risk MDS with NRAS or KRAS mutations, cohort 2 patients with AML, MDS, or CMML with a RAS wild-type mutation or an unknown mutation status, and cohort 3 patients with CMML with an NRAS or KRAS mutation. The recommended dose for trametinib was 2 mg orally daily. The overall response rates for cohorts 1, 2, and 3 were 20%, 3%, and 27%, respectively, with a preferential activity among myeloid malignancies with RAS mutations. Repeated cycles of trametinib were well tolerated with manageable or reversible toxicities. Thus, some therapeutic potential of trametinib was demonstrated in myeloid malignancies, particularly in RAS-pathway mutated CMML.

There is some evidence of activity or JAK inhibitors in CMML patients. In a study by Geissler the specific JAK2 inhibitor TG101209 was found to either block or strongly inhibit spontaneous CFU-GM growth in vitro in all 10 CMML patients tested (71). Among these 10 patients 6 were tested by NGS and, in 5 of them, RAS-pathway hyperactivation was documented due to mutations in NRAS (n=3) or PTPN11 (n=2), respectively. In a NRAS-mutant CMML patient who was treated with the JAK1/2 inhibitor ruxolitinib off label, spleen response and the disappearance of constitutional symptoms was associated with a decrease of autonomous CFU-GM formation ex vivo. Thus, therapeutic potential of inhibition of the JAK2/STAT5 pathway by ruxolitinib in CMML is suggested. In a phase I/II clinical trial of ruxolitinib in 20 CMML patients the recommended dose of ruxolitinib was 20 mg twice daily and the ORR of 35%, 5/9 spleen responses, and 10/11 symptom responses were seen (72). Correlative analysis demonstrated a downregulation in inflammatory cytokines and GM-CSF-dependent STAT5 phosphorylation in responders. Further studies are required to demonstrate a potential disease modifying effect of ruxolitinib in CMML.

Biological crosstalk is a phenomenon in which one component of a signal transduction pathway can affects another pathway. Thus, the PI3 Kinase-pathway may be aberrantly activated in CMML without molecular aberration in it. Treatment with inhibitors of this aberrantly activated signaling could have the potential to impact malignant cell growth. Using the class I PIK3 inhibitor pictilisib this approach has been successfully applied in a Kras G12D and in a NF1 -/- mouse model with a CMML-like phenotype (73). In this models, pictilisib attenuated activation of both PI3K/AKT and RAS/MEK/ERK pathways in primary hematopoietic cells. Several PI3K inhibitors have now received regulatory approval for the treatment of breast cancer and B-cell malignancies suggesting that the treatment concept of PI3K-pathway inhibition comes into the clinic (134, 135). Thus, based on some crosstalk between the RAS-signaling and the PI3K/AKT-pathway PI3K inhibitors could be important molecules for the design of future therapeutic strategies for patients with CMML.

In MP-CMML RAS-pathway mutations are associated with a unique gene expression profile enriched in mitotic kinases including polo-like kinase 1 (PLK1) (74) as shown in a study using a multiomics platform and biochemical and molecular analyses. In this study unmutated MLL regulated PLK1 transcript levels via promoter monomethylation of lysine 4 of histone 3. In the preclinical mouse model pharmacologic inhibition of PLK1 was effective in RAS-mutant patient-derived xenografts providing the rationale for a new biomarker-driven therapeutic approach in patients with proliferative CMML. Currently the administration of the PLK4 inhibitor CFI-400945 with or without HMA is tested in a phase II trial in patients with relapsed/refractory or untreated AML, MDS, or CMML (NCT04730258).

The inflammasome is a multimeric protein complex including NLRP3, NLRC4, AIM2 and NLRP6 that initiates an inflammatory form of cell death (pyroptosis) and triggers the release of proinflammatory cytokines (136). Recently, a functional link between oncogenic Kras G12D and inflammasome activation was reported in a preclinical model (75). In this mouse model Kras driven myeloproliferation and cytopenia was reversed by functional inactivation of NLRP3. A similar phenotypic improvement was seen with therapeutic IL-1-receptor blockade. Importantly, Kras activation induced the production of reactive oxygen species suggesting that KRAS not only has an ongogenic driver function but also activates the proinflammatory machinery. These findings open a new therapeutic opportunity for Kras mediated malignancies including CMML. Interestingly, there is a current phase II study, in which the anti-IL1ß inhibitor canakinumab is studied in ESA or HMA-refractory low risk-MDS or CMML (NCT04239157).

Given the complexity of CMML one can expect, that combinations of molecules impacting different pathways may yield better efficacy. At least in preclinical models this seems to be true (76). In Nras hyperactive mouse models mimicking MP-CMML inhibition of the MEK-pathway alone was only partially effective to improve disease associated features. Despite MEK inhibitor treatment 60% of Nras G12D expressing mice died within 20 weeks and surviving animals continued to retain their MP-CMML phenotype. Combined inhibition of the MEK and JAK/STAT signaling, however, greatly inhibited human and mouse CMML cell growth in vitro, rescued mutant Nras G12D-expressing HSC function in vivo, and promoted long-term survival without evident disease manifestation in animals with RAS-pathway driven MP-CMML. Still much work has to be done to address optimal ways to target these pathways in patients with CMML to improve clinical outcome.