Effective drug treatment identified by in vivo screening in a transplantable patient-derived xenograft model of chronic myelomonocytic leukemia

To establish novel and effective treatment combinations for chronic myelomonocytic leukemia (CMML) preclinically, we hypothesized that supplementation of CMML cells with the human oncogene Meningioma 1 (MN1) promotes expansion and serial transplantability in mice, while maintaining the functional dependencies of these cells on their original genetic profile. Using lentiviral expression of MN1 for oncogenic supplementation and transplanting transduced primary mononuclear CMML cells into immunocompromised mice, we established three serially transplantable CMML-PDX models with disease-related gene mutations that recapitulate the disease in vivo. Ectopic MN1 expression was confirmed to enhance the proliferation of CMML cells, which otherwise did not engraft upon secondary transplantation. Furthermore, MN1-supplemented CMML cells were serially transplantable into recipient mice up to 5 generations. This robust engraftment enabled an in vivo RNA interference screening targeting CMML-related mutated genes including NRAS, confirming that their functional relevance is preserved in the presence of MN1. The novel combination treatment with azacitidine and the MEK-inhibitor trametinib additively inhibited ERK-phosphorylation and thus depleted the signal from mutated NRAS. The combination treatment significantly prolonged survival of CMML mice compared to single-agent treatment. Thus, we identified the combination of azacitidine and trametinib as an effective treatment in NRAS-mutated CMML and propose its clinical development.

Treatment options are limited in CMML and include observation, cytoreduction with hydroxyurea, treatment with hypomethylating agents azacitidine or decitabine and allogeneic hematopoietic cell transplantation (alloHCT) in eligible patients. Only alloHCT has a curative intent and resulted in a 3-year overall survival of 31% in single-center experience [12]. Other treatment approaches aim at symptom relief in older patients. A matched-pair analysis of CMML patients treated with azacitidine (n = 22) compared to patients treated with hydroxyurea (n = 22) resulted in similar median overall survival (7.5 vs 6.2 months, respectively (p = 0.251) and thus the benefit of hypomethylating agents over hydroxyurea is unclear [13].
None of the mutations found in CMML patients is unique for the disease. However, ASXL1, TET2 and SRSF2 mutations occur more frequently in CMML patients (40-50% each) than in other myeloproliferative or myelodysplastic diseases [14]. Interestingly, mutations in signaling pathway genes like JAK2, NRAS and CBL are common in CMML and rare in MDS. CMML cells typically display hypersensitivity to granulocyte-macrophage colony-stimulating factor (GM-CSF) in vitro, which likely contributes to the monocytic phenotype of the disease and is mediated by mutations in signaling genes [15,16].
In order to better understand the pathophysiology of CMML and to evaluate novel treatment combinations, patient-derived xenograft (PDX) models were developed by transplanting primary patient cells in immunocompromised mice. NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ (NSG) mice transgenic for IL3, SCF and GM-CSF (NSGS) were shown to enhance engraftment of CMML cells likely through expression of GM-CSF [17]. Xenotransplantation of bone marrow or peripheral blood mononuclear cells from 18 CMML patients into NSGS mice gave rise to robust human hematopoietic cell engraftment in almost all cases. However, transplantation into secondary recipients succeeded only in 2 of 5 cases and further serial transplantation was not reported [18]. Similarly, cells from 15 of 16 CMML patients engrafted in primary NSG or NSGS recipient mice, but in secondary NSGS recipients only 3 of 10 samples engrafted with low levels of 2-9% human cells in bone marrow [17].
Overexpression of the transcriptional co-factor MN1 has been observed in a broad spectrum of AML patients [19][20][21][22][23] and has been identified as an independent prognostic marker for AML with normal karyotype. High expression of MN1 is associated with poor prognosis, shorter overall and relapse-free survival, and poor response to treatment [21]. In murine cells human MN1 overexpression induces aggressive, fully penetrant AML through the promotion of leukemic cell self-renewal [24][25][26], impairment of myeloid differentiation [24,25], and resistance to all trans retinoic acid-induced differentiation [25]. MN1-induced leukemias are also associated with upregulation of HOXA genes and MEIS1 [27]. Moreover, functional assays have revealed the critical dependence of the MEIS1/AbdB-like HOX protein complex for MN1-induced transformation [27].
We aimed to develop a xenograft model that allows expansion of CMML cells in vivo, serial transplantation for at least 3 generations of mice and functional manipulation of primary CMML cells for an in vivo shRNA screen. We hypothesized that supplementation of CMML cells with a human oncogene will promote expansion and serial transplantability while the functional dependencies of these cells on their genetic profile are maintained. We chose the MN1 oncogene for this approach as transduction of MN1 in CD34 + cord blood cells promotes self-renewal in vitro and engraftment in vivo, but does not induce AML in vivo nor does it enable serial transplantation [28].
To establish a serially transplantable PDX model for CMML we used lentiviral expression of MN1 in primary CMML cells for oncogenic supplementation and evaluated the growth potential of these cells in vivo by serial transplantations, a functional in vivo shRNA screen and a novel combination treatment in vivo.

Patient samples and xenotransplantation
Diagnostic bone marrow or peripheral blood samples were collected after obtaining written informed consent from CMML patients diagnosed and treated at Hannover Medical School [29,30]. Briefly, mononuclear cells were isolated from patient bone marrow aspirates or peripheral blood by density centrifugation using Biocoll Separating Solution (Biochrom, Berlin, Germany). The cell number was determined and depletion of CD3 positive cells was done following the manufacturer's protocol (CD3 MicroBeads kit, Miltenyi Biotec, Bergisch Gladbach, Germany). Patient cells were analyzed according to the principles of the Declaration of Helsinki, and the study was approved by the institutional review board of Hannover Medical School. Patient characteristics are listed in Supplementary Table S1. Detailed methods for cell culture, viral vectors and infection of cells, clonogenic progenitor assays, immunoblotting and ampliconbased next-generation sequencing [31,32] of treated CMML cells are described in the Supplementary Methods.
For xenotransplantation studies, 6-to 8-week-old female NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJmice transgenic for human interleukin-3, granulocyte-macrophage-colony-stimulating factor and stem cell factor (NSGS) were bred by the central animal laboratory of Hannover Medical School under pathogen-free conditions. All animal experiments were performed in accordance with the German animal protection law and approved by the Lower Saxony state office for consumer protection, Oldenburg, Germany. Xenotransplantation, engraftment monitoring, treatment and morphologic analysis of CMML cells are described in the Supplementary Methods. Technical details on serial transplantations are provided in Supplementary Table S2.

Quantitative RT-PCR
Total RNA was extracted from 1 million cells and reverse transcribed as previously described [25]. Quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) was done as previously described using SYBR green (Invitrogen, ThermoFischer Scientific, Schwerte, Germany) for quantification of double-stranded DNA on a StepOne Plus cycler (Applied Biosystems, Darmstadt, Germany). Relative expression was determined with the 2 −ΔΔCT method using ABL1 as the housekeeping gene, and expression values were normalized relative to control cells [33]. Used custom primers were shown in Supplementary Table S3.

Statistical analysis
Data are shown as mean ± standard error of mean (SEM). Student t test and Logrank test were used to assess differences between groups and survival curves, respectively. All statistical analysis were performed with GraphPad Prism Version 7 software. P < 0.05 was considered statistically significant. The size of the animal cohorts was based on our previous study [34,35]. All in vitro experiments were performed at least 3 times unless otherwise stated and all attempts of replication were successful.
The immunophenotype of engrafted human CD45 and EGFP positive CMML#1-MN1 cells showed a myelomonocytic differentiation expressing CD33, CD38 and CD14, including progenitor markers CD38 and in few cells CD34 (Fig. 1i). Morphologic analysis of bone marrow from CMML#1-MN1 mice showed myeloid progenitors and monocytic cells similarly to bone marrow smears from the CMML#1 patient at diagnosis and revealed MN1 expression by EGFP fluorescence (Fig. 1j and Supplementary  Fig. S1C). In summary, oncogenic supplementation of CMML cells with MN1 enabled engraftment of these cells that otherwise did not engraft in vivo.

CMML-MN1 cells engraft secondary recipient mice and are serially transplantable
To evaluate whether CMML#1-MN1 cells can be serially transplanted, CMML#1-CTL, CMML#1-EGFP and CMML#1-MN1 cells from bone marrow of primary recipients were isolated and intravenously transplanted into sublethally irradiated secondary NSGS recipient mice. Only CMML#1-MN1 cells engrafted in secondary recipients (Fig. 2a), and showed mean engraftment of 10%, 25% and 5% of human CD45+ cells in peripheral blood, bone marrow and spleen, respectively, at the time of sacrifice (Fig. 2b). Engrafted CMML#1-MN1 cells had a myeloid phenotype expressing CD33, CD38, and CD123, but lacked expression of lymphoid markers CD3 and CD19 ( Supplementary Fig. S1D), and developed splenomegaly (Fig. 2c). Cryopreserved CMML#1-MN1 cells from bone marrow and spleen of secondary transplanted mice were used as source to initiate serial transplantations in NSGS recipients up to 5 times (Fig. 2d). Serial transplantation resulted in higher engraftment levels in peripheral blood and shorter latency, suggesting that the cells underwent clonal selection.
Cells from two additional CMML patients were used to confirm MN1-enhanced serial engraftment. Detailed patient characteristics are provided in Supplementary Table S1. In the CMML#2 model MN1-transduced cells showed higher engraftment levels in several mice at 16 and 18 weeks after transplantation compared to CTL mice (Fig. 2e). Upon secondary transplantation only MN1-transduced cells engrafted (7% to 44% human CD45+ in peripheral blood, 11% to 98% human CD45+ in bone marrow and 18% to 36% human CD45+ in spleen), while CTL and EGFPtransduced cells did not engraft ( Fig. 2f and Supplementary  Fig. S2A). In the CMML#3 model we observed comparable robust engraftment of both CMML#3-MN1 and CMML#3-CTL cells at week 12 after transplantation in primary recipient NSGS mice (Fig. 2g). Upon secondary transplantation of spleen and bone marrow cells (65% to 85% human CD45+ cells in 1 st Tx CMML#3-MN1 and 26% to 47% human CD45+ cells in 1 st TxCMML#3-CTL mice) only MN1-transduced cells engrafted at high levels (59% to 82% human CD45+ cells in peripheral blood, 93% to 96% in bone marrow and 74% to 88% in spleen), while the CTL group did not engraft secondary recipient NSGS mice ( Fig. 2h  Clonal dominance of an NRAS/NOTCH1 mutated subclone during serial transplantation of CMML#1-MN1 cells We evaluated the biologic and genetic evolution of CMML#1-MN1 cells during serial transplantation from first to fifth generation. Up to the third transplantation CMML#1-MN1 cells showed preferential engraftment in bone marrow over peripheral blood and spleen (Fig. 3a). In the fourth transplantation engraftment in bone marrow and spleen was comparable and in the fifth transplantation engraftment was equally high in blood, bone marrow and spleen (Fig. 3a). Bone marrow morphology showed a similar monocyte dominated picture compared to the patients' bone marrow from first to third transplantation, while the blast count increased in the fourth and fifth transplantations above 20%, while the monocytes remained at high levels (Fig. 3b, c). This indicates that upon serial transplantation the most aggressive clones prevail and the disease is more related to AML, while it still reflects the same genetics that were found in the patients' cells. Spleen weight in sacrificed mice increased over time consistent with engraftment levels in the spleen (Fig. 3d). Genetic profiling of the patient's cells identified a founding clone characterized by DNMT3A, U2AF1 and BCOR mutations and two subclones with additional mutations. The first subclone was characterized by a NRAS G12D mutation and a NOTCH1 mutation, and the second subclone included a NRAS G12V mutation (Fig. 3e). Upon transduction with MN1 and transplantation in vivo only the NRAS G12D/ NOTCH1 subclone containing also the DNMT3A, U2AF1 g White blood cell count in peripheral blood of primary recipient NSGS mice at sacrifice (number of analyzed mice is indicated in the figure; mean ± SEM). h Hemoglobin level in peripheral blood of primary recipient NSGS mice at sacrifice (number of analyzed mice is indicated in the figure; mean ± SEM). i Immunophenotype of engrafted CMML#1 cells from bone marrow and spleen of primary recipient NSGS mice at sacrifice (gated on human CD45 + EGFP + CMML#1-MN1 cells; n = 3; mean ± SEM). j Morphology from bone marrow smears of the CMML#1 patient at diagnosis and from a transplanted mouse at sacrifice. The enlarged cell shows granules in the cytoplasm that can be identified as lipid granules in cells from later transplantations (see Fig. 3b). and BCOR mutations engrafted in PDX mice and persisted without further changes up to the fifth transplantation ( Fig. 3f and Supplementary Table S4). In summary, serial transplantation led to the selection of a more aggressive leukemic clone that was characterized by mutations recurrently found in CMML and AML patients. Only the NRAS G12D/NOTCH1 mutated subclone stably engrafted in vivo, and its consistent representation in all transplantations suggests that mutations in DNMT3A, U2AF1, BCOR, NRAS and NOTCH1 were required for in vivo proliferation despite the presence of MN1. We found that clonal selection of MN1-transduced CMML cells occurred during engraftment in vivo and not during the ex vivo transduction procedure as shown for the CMML#3-MN1 model ( Supplementary Fig. S4). The known GM-CSF/CSF2 dependency of CMML cells was evaluated for CMML#1-MN1 cells by transplanting cells from the fourth recipient generation into either NSG or NSGS mice. The latter mouse strain produces human GM-CSF besides IL3 and SCF. Upon transplantation of equal cell numbers CMML#1-MN1 cells engrafted much faster in NSGS compared to NSG recipient mice, suggesting that these MN1-transduced cells retained their GM-CSF sensitivity ( Supplementary Fig. S5).

Loss-of-function shRNA-screening identifies functional dependencies in CMML#1-MN1 cells
To evaluate the functional importance of the mutated genes and MN1 we performed a targeted shRNA screening in vivo. Bone marrow cells from the third generation of CMML#1-MN1 mice were harvested and transduced with a shRNA pool consisting of 2 independent shRNAs against 8 genes and 1 control shRNA against luciferase. Six shRNAs targeted the mutated genes in this patient or MN1, while 2 shRNAs targeted 2 control genes that are recurrently mutated in myeloid disease (GATA2 and NF1). 48 h after transduction the cells were either transplanted in NSGS mice or harvested for shRNA quantification. Eight weeks after transplantation the cells were harvested from blood, bone marrow and spleen and DNA was extracted from unsorted cells to quantify shRNA copies in total DNA by NGS (Fig. 4a). The shRNAs against the 8 genes were validated in vitro and induced knockdown usually by more than 80% (Supplementary Fig. S6 and Supplementary  Table S5). shRNAs against the control genes luciferase, GATA2 and NF1 were not depleted after transplantation in mice compared to the cells harvested after transduction in vitro (Fig. 4b). Interestingly, the shRNAs against the mutated genes and MN1 were strongly depleted in most mice and cell sources (blood, bone marrow, spleen) after in vivo transplantation except one of the two shRNAs against BCOR (Fig. 4b). This analysis suggests that at least four mutated genes are required for the survival of CMML-MN1 cells in vivo in addition to MN1 and shows that our model can be used to evaluate the specific contribution and biology of each mutated gene despite the ectopic expression of the MN1 oncogene.

The MEK-inhibitor trametinib prolongs survival of CMML-MN1 mice when combined with azacitidine
Hypomethylating agents are used in CMML patients with proliferative disease who are not eligible for alloHCT. As the NRAS-mutated subclone engrafted efficiently in NSGS mice and knockdown of NRAS depleted CMML#1 cells, we hypothesized that the combination of the MEK-inhibitor trametinib, which inhibits the signaling pathway downstream of NRAS, with azacitidine more efficiently inhibits CMML development than either drug alone. We transplanted CMML#1-MN1 cells from the fifth transplantation in NSGS mice and confirmed similar engraftment of human CD45+ EGFP+ cells in peripheral blood 2 weeks after transplantation (Supplementary Fig. S7A). Treatment was started 4 weeks after transplantation with azacitidine (1 mg/kg i.p. for 5 days) and/or trametinib (2 mg/kg p.o. for 18 days). At this time we assume similar engraftment levels between the indicated groups, as 4 weeks after the start of treatment engraftment was still comparable between the treatment groups ( Supplementary Fig. S7B). Trametinib treatment was interrupted until day 118 and then continued until the end of the experiment (Fig. 5a). Engraftment monitoring at week 15 after transplantation revealed diminished engraftment in peripheral blood of mice treated with azacitidine (6% CD45+ cells) and the combination of azacitidine+trametinib (3% CD45+ cells) as compared to trametinib treated (17% CD45+ cells) and vehicle-treated (34% CD45+ cells) mice ( Fig. 5b and Supplementary  Fig. S7B). Trametinib treated mice had similar survival as vehicle-treated mice, while azacitidine treated mice showed improved survival (Fig. 5c). Interestingly, mice treated with the combination of trametinib/azacitidine survived significantly longer than vehicle treated, azacitidine treated or trametinib treated mice alone (median survival for vehicle 125 days, azacitidine 166 days, trametinib 129 days and trametinib/azacitidine 194 days, Fig. 5c, P = 0.041).
Engraftment of CMML-MN1 cells was significantly lower in azacitidine and trametinib/azacitidine treated groups compared to vehicle control ( Supplementary Fig. S7C). The weight of the spleens was lower in the single and combination treatment groups compared to the vehicle group despite a longer latency to sacrifice (Fig. 5d). At the end of the treatment, WBC counts and hemoglobin counts were   NRas G12V Patient 1 st Tx 5 th Tx to decreased in the vehicle group but remained in the normal range for azacitidine and trametinib/azacitidine treated groups. Platelets were increased in azacitidine and trametinib/azacitidine treated groups when compared with vehicle or trametinib treated mice (Fig. 5e-g).
To evaluate whether the additive effect of trametinib and azacitidine was also found in primary cells from CMML patients we performed CFC assays from peripheral blood or bone marrow cells of five CMML patients (Supplementary  Table S1). The combination treatment reduced the CFC count significantly stronger compared to either azacitidine or trametinib monotherapy (Fig. 5h). In line with this finding, phosphorylated ERK (pERK) was most effectively inhibited in CMML#1-MN1 cells harvested from NSGS mice as well as primary CMML cells by the combination treatment compared to single-agent azacitidine, trametinib or vehicle treatment (Fig. 5i, j). In summary, the combination of the MEK-inhibitor trametinib and a hypomethylating agent prolonged survival in the CMML#1-MN1 model by additively inhibiting pERK signaling.

Discussion
We developed patient-derived CMML models that recapitulate the disease in vivo and allowed us to demonstrate that 1) ectopic MN1 expression enhances the engraftment and proliferation of cells that otherwise do not engraft in mice, 2) the functional relevance of CMML-related gene mutations is preserved in the presence of MN1, 3) MEK/ERK signaling is important for in vivo proliferation of CMML cells, 4) the serial transplantability of the MN1-CMML cells enables functional RNAi screening in vivo and 5) the combination of azacitidine with trametinib is an effective combination treatment in CMML#1-MN1 cells in vivo and primary CMML cells in vitro.
It was shown in recent years that CMML cells readily engraft in immunocompromised mice by intrafemoral or intravenous injection and are transplantable in secondary recipient mice in some cases [17,18]. However, transplantation in secondary recipients was inconsistent and no tertiary transplantation was reported. We chose the MN1 oncogene to enhance the engraftment of CMML cells as MN1 enhances self-renewal and blocks differentiation of CD34+ cord blood cells in vitro and induces a myeloproliferative disease, but not AML, in vivo. Only by combining MN1 with another oncogene, NUP98-HOXD13, could cord blood cells be transformed to AML and serially transplanted [28]. Here we confirm our hypothesis that the mutations present in CMML cells collaborate with MN1 and maintain their functional relevance. So far, only expression of the fusion genes MLL-AF9 and MLL-ENL were able to fully transform normal human hematopoietic stem cells [36][37][38][39]. The incomplete transformative capacity of MN1 may be ideally suited to enhance the proliferative capacity of malignant cells and enable their investigation in vivo.
The robust engraftment of MN1-CMML cells allowed us to perform in vivo RNA interference targeting the mutated Fig. 3 Clonal dominance of an NRAS-mutated subclone during serial transplantation of CMML#1-MN1 cells. a Engraftment of human hematopoietic cells in peripheral blood, bone marrow and spleen of NSGS recipient mice from first to fifth serial transplantation (number of analyzed mice is indicated in the figure, mean ± SEM). b Morphology from bone marrow smears of engrafted CMML#1-MN1 cells of NSGS recipient mice from first to fifth serial transplantation. c Differential blood count quantification of bone marrow smears from the patient (CMML#1) at diagnosis and engrafted CMML#1-MN1 cells in mice from first to fifth transplantation (n = 3; mean ± SEM). d Spleen weight of primary CMML#1-EGFP recipient mice and CMML#1-MN1 recipient mice from first to fifth serial transplantation at the time of sacrifice (each dot represents one mouse; mean ± SEM). e Variant allele frequencies of mutated genes in bone marrow cells of the CMML#1 patient and corresponding CMML#1-MN1 cells engrafted in the bone marrow and spleen of NSGS recipient mice at sacrifice from first to fifth serial transplantation (n = 3; mean ± SEM). f Graphical representation of clonal evolution of CMML#1 cells from the patient and its corresponding CMML#1-MN1 cells from first to fifth serial transplantation in NSGS recipient mice.  in 30% of CMML patients [7,9] and the NRAS/NOTCH1 mutated subclone in our patient was selected during the transplantation procedure and persisted throughout all 5 serial transplantations. Ricci et al. [40] underscored the importance of RAS pathway mutations by showing that they induce a CMML-like disease in mice. NOTCH1 mutations have not been reported in CMML, but mutations in NOTCH pathway genes were found in 5 of 47 CMML patients [41]. Klinakis et al. showed that deletion of at least NOTCH1 and NOTCH2 induced a CMML-like disease in mice [41]. It remains unclear whether the NOTCH1 mutation in our model was a loss-of-function or rather gain-offunction mutation, as inhibiting NOTCH1 in our experiments by shRNA depleted the CMML cells in vivo. U2AF1 mutations occur in approximately 7% of CMML patients [42]. U2AF1 mutated cells had a competitive disadvantage in mouse models compared to U2AF1 wildtype cells. The combination of mutated U2AF1 with RUNX1 induced AML in mice, but 2 of 3 established AMLs lost the U2AF1 mutation [43]. It is thus surprising that the U2AF1 mutation was maintained in our model over 5 generations, which represents a faithful model to study the impact of U2AF1 mutations in vivo. Knockdown of U2AF1 resulted in depletion of CMML cells in vivo in our study. A gain-offunction has been suspected for U2AF1 mutations, as they occur only at 2 hotspots in this gene, but catastrophic depletion of the splicing factor SRSF2 has also led to cell depletion [44], and thus the nature of our U2AF1 mutation remains unclear. DNMT3A mutations occur in approximately 5% of CMML patients [45]. The Arg882His mutation found in our patient indicates an origin of the disease in clonal hematopoiesis of indeterminate potential [46]. This mutation impairs the methyltransferase activity of DNMT3A [47] and loss of DNMT3A leads to enhanced self-renewal of hematopoietic stem cells and expansion of the DNMT3A null clone [48]. It was therefore unexpected that shRNA-mediated knockdown of DNMT3A depleted CMML cells in vivo. However, DNMT3A depletion in hepatocellular carcinoma cells resulted in reduced proliferation [49].
As RAS-pathway mutations are frequent in CMML and the MEK-inhibitor trametinib is approved for melanoma and shows single-agent clinical activity in RAS mutated patients with myeloid malignancies including CMML [50], we tested the combination treatment of trametinib and azacitidine in vivo. The reduced engraftment at 15 weeks, reduced spleen weight at death and the significantly prolonged survival compared to azacitidine monotherapy suggests that this combination is at least additive in NRASmutated CMML cells. The additive effect was confirmed in CFC assays with primary, non-transduced patient cells from five CMML patients and by the strong synergistic inhibition of phosphorylated ERK (p-ERK) in MN1-CMML and primary patient cells. All five patients included in the CFC assay lacked mutations in NRAS, but four of the five patients had a mutation in signaling genes that regulate the same pathway (2 patients with CBL mutations, one patient with a cKit mutation and one patient with a JAK2 mutation). This suggests that CMML cells may be sensitive to trametinib independent of NRAS mutations, as they have frequent comutations in other signaling genes. So far, the combination of azacitidine with trametinib has not been tested in CMML patients with or without RAS-pathway mutations, but our data suggest that this combination should be evaluated in a phase II clinical trial.
In summary, we established a transplantable CMML model with NRAS, NOTCH1, DNMT3A, U2AF1 and BCOR mutations and validated the functional importance of mutated NRAS. We identified the combination of azacitidine and trametinib as an effective treatment in NRASmutated CMML and suggest its clinical development to eventually improve the outcome of CMML patients. Fig. 5 The MEK-inhibitor trametinib prolongs survival of CMML#1-MN1 mice when combined with azacitidine. a Schematic representation of the treatment regimens (Aza, azacitidine; Tram, trametinib). b Monitoring of CMML burden in recipient mice at week 15 after transplantation (each dot represents one mouse; mean ± SEM). c Survival of NSGS recipient mice engrafted with CMML#1-MN1 cells and treated with vehicle (n = 10), trametinib (n = 8), azacitidine (n = 10) and the combination of trametinib+azacitidine (n = 7). d Spleen weight of NSGS recipient mice engrafted with CMML#1-MN1 cells and treated with vehicle, trametinib, azacitidine and the combination of trametinib+azacitidine at sacrifice (each dot represents one mouse; mean ± SEM). e White blood cell count in peripheral blood of CMML#1-MN1 bearing mice treated with vehicle, trametinib, azacitidine and the combination of trametinib+azacitidine 4 weeks after treatment start and at sacrifice (number of individual mice is indicated in the figure; mean ± SEM). f Hemoglobin levels in peripheral blood of CMML#1-MN1 bearing mice treated with vehicle, trametinib, azacitidine and the combination of trametinib+azacitidine 4 weeks after treatment start and at sacrifice (number of individual mice is indicated in the figure; mean ± SEM). g Platelet count in peripheral blood of CMML#1-MN1 bearing mice treated with vehicle, trametinib, azacitidine and the combination of trametinib+azacitidine 4 weeks after treatment start and at sacrifice (number of individual mice is indicated in the figure; mean ± SEM). h Effect of the combination treatment in five different CMML patients (unrelated to the CMML#1-MN1 cells). Primary CMML cells were treated with vehicle, azacitidine (500 nM), trametinib (20 nM) or the combination of azacitidine (500 nM) + trametinib (20 nM) and plated in duplicate in CFC media. After 20 days colonies were counted and expressed as percentage of the vehicle-treated cells. White dots within all groups represent cells from a patient without mutation in a signaling gene (n = 5, mean ± SEM). i Immunoblotting for p-ERK and ERK in azacitidine/trametinib-and vehicle-treated CMML patients (n = 3) and CMML#1-MN1 (n = 1) cells from third recipient mice at 6 h after treatment. j Average p-ERK-to-ERK ratio from immunoblots shown in Fig. 5i as percentage of vehicle-treated CMML cells (n = 4, mean ± SEM).
Facility of Hannover Medical School, and Martin Wichmann, Kerstin Görlich and Karin Battmer for their support. We acknowledge the assistance of the Cell Sorting Core Facility of Hannover Medical School supported in part by the Braukmann-Wittenberg-Herz-Stiftung and the Deutsche Forschungsgemeinschaft. This work was supported by an ERC grant under the European Union's Horizon 2020 research and innovation program (No. 638035), by grants 109714 and 70112697 from Deutsche Krebshilfe; and DFG grants HE 5240/5-1, HE 5240/5-2, HE 5240/6-1, and HE 5240/6-2.
Funding Open Access funding enabled and organized by Projekt DEAL.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.