|Home | About | Journals | Submit | Contact Us | Français|
Chronic and juvenile myelomonocytic leukemias (CMML and JMML) are aggressive myeloproliferative neoplasms that are incurable with conventional chemotherapy. Mutations that deregulate Ras signaling play a central pathogenic role in both disorders, and Mx1-Cre, KrasLSL-G12D mice that express the Kras oncogene develop a fatal disease that closely mimics these two leukemias in humans. Activated Ras controls multiple downstream effectors, but the specific pathways that mediate the leukemogenic effects of hyperactive Ras are unknown. We used PD0325901, a highly selective pharmacological inhibitor of mitogen-activated protein kinase kinase (MEK), a downstream component of the Ras signaling network, to address how deregulated Raf/MEK/ERK signaling drives neoplasm formation in Mx1-Cre, KrasLSL-G12D mice. PD0325901 treatment induced a rapid and sustained reduction in leukocyte counts, enhanced erythropoiesis, prolonged mouse survival, and corrected the aberrant proliferation and differentiation of bone marrow progenitor cells. These responses were due to direct effects of PD0325901 on Kras mutant cells rather than to stimulation of normal hematopoietic cell proliferation. Consistent with the in vivo response, inhibition of MEK reversed the cytokine hypersensitivity characteristic of KrasG12D hematopoietic progenitor cells in vitro. Our data demonstrate that deregulated Raf/MEK/ERK signaling is integral to the growth of Kras-mediated myeloproliferative neoplasias, and further suggest that MEK inhibition could be a useful way to ameliorate functional hematologic abnormalities in patients with CMML and JMML.
Somatic mutations in the NRAS or KRAS oncogenes are identified in 20–40% of patients with chronic or juvenile myelomonocytic leukemias (CMML or JMML). Mutations in related molecules collectively implicate hyperactive Ras signaling in ~50% of CMML and 90% of JMML (1). The lack of compounds that directly inhibit oncogenic Ras has led to widespread efforts to find alternative therapeutic targets. We previously developed a genetically engineered mouse model that recapitulates many features of human myeloproliferative neoplasms. In this animal model, KrasLSL-G12D mice carry a conditional Kras allele that expresses oncogenic K-RasG12D after an upstream “stop” cassette is removed by Cre recombinase (2). Mx1-Cre transgenic mice express Cre in response to polyinosinic-polycytidylic acid (pIpC). Therefore, Mx1-Cre, KrasLSL-G12D mice express KrasG12D from the endogenous locus after treatment with pIpC. These mice (hereafter designated Mx1-Cre, KrasG12D) rapidly develop a progressive myeloproliferative neoplasm that is characterized by leukocytosis, splenomegaly, and severe anemia (3, 4). The similarities of this model to human myeloproliferative neoplasms suggest that it might be useful for studying disease mechanisms and for testing potential therapeutic strategies.
We previously showed that the Raf/MEK/ERK signaling pathway is modestly hyperactive in primary hematopoietic progenitor cells from the bone marrow of Mx1-Cre, KrasG12D mice (5). Because numerous pathways are potentially deregulated by oncogenic Ras, the importance of deregulated Raf/MEK/ERK signaling in KrasG12D-driven myeloproliferative neoplasia remains unclear (6). To address this question and to evaluate alternative therapeutic strategies for CMML and JMML, we treated Mx1-Cre, KrasG12D mice with PD0325901, a potent and highly specific inhibitor that binds to an allosteric site on mitogen-activated protein kinase kinase (MEK) that is not conserved in other protein kinases (7–9). We show that PD0325901 treatment improves multiple hematologic abnormalities in Mx1-Cre, KrasG12D mice by direct effects on bone marrow progenitor cells that express oncogenic Kras. This demonstrates that deregulated Raf/MEK/ERK signaling is integral to Kras-mediated myeloproliferative neoplasia, and suggests that MEK inhibition may be a useful approach for treating patients with CMML and JMML.
The ability of PD0325901 to inhibit MEK in vivo was validated by measuring ERK phosphorylation induced by GM-CSF stimulation of primary hematopoietic progenitor cells (Fig. 1A). In a flow cytometry based assay, Lin−/lo c-kit+ CD34+ CD105− bone marrow cells were enriched for myeloid progenitors that responded to GM-CSF. We treated Mx1-Cre, KrasG12D mice with PD0325901 and measured the ability of GM-CSF to evoke protein phosphorylation in bone marrow harvested at various times after administration. An oral dose of 5 mg/kg suppressed the ability of GM-CSF to phosphorylate ERK in mouse bone marrow cells for 18–24 h (Fig. 1B), which is consistent with previous data in this mouse strain (10). Phosphorylation of STAT5, which is independent of Raf/MEK/ERK activity, was unimpaired (Fig. 1C), consistent with the expected specificity of PD0325901.
To investigate whether PD0325901 reduces the severity of disease in Mx1-Cre, KrasG12D mice, we induced KrasG12D expression in 3–4 week old pups and allowed the myeloproliferative neoplasia to progress until the age of 8 weeks. The disease was well-established by this time, as indicated by high blood leukocyte counts (41,000/μL ± 25,000 s.d.) (Fig. 2A) and low hemoglobin concentrations (10.6 g/dL ± 3.8 s.d.) (Fig. 2B), compared with wild-type control mice. Mx1-Cre, KrasG12D mice and wild-type littermates were then randomized to receive PD0325901 at a dose of 5 mg/kg/day or vehicle treatment. Mx1-Cre, KrasG12D mice that received the PD0325901 MEK inhibitor demonstrated rapid improvements in composition of the peripheral blood, with reduced leukocyte counts (Fig. 2A), disappearance of anemia (Fig. 2B) and reticulocytosis (Fig. 2C), and reduced splenomegaly (Fig. 2D). Daily treatment with PD0325901 also prolonged dramatically the survival of Mx1-Cre, KrasG12D mice compared with vehicle-treated mice (8.1 vs. 2.0 weeks on trial; p<0.0001 by log rank test) (Fig. 2E). Two of three Mx1-Cre, KrasG12D mice treated for 12 weeks died with KrasG12D T-lineage leukemia/lymphoma (T-ALL), suggesting that some hematopoietic malignancies are not susceptible to MEK inhibition (10). There were no adverse effects of PD0325901 administration observed in wild-type mice.
To investigate the mechanism of this response to MEK inhibition, we first explored whether PD0325901 treatment eliminated Kras mutant cells from the bone marrow. We used PCR genotyping to determine the configuration of the conditional Kras allele in myeloid colonies obtained from Mx1-Cre, KrasG12D mice treated with PD0325901. Because each colony is derived from a single progenitor cell, termed a granulocyte/macrophage colony forming unit (CFU-GM), this analysis estimates the frequency of progenitor cells that express oncogenic KrasG12D. This assay demonstrated that the majority of CFU-GM contained the mutant Kras allele in the expressed configuration that lacks the “stop” cassette (Fig. 2F), even after 5 weeks of PD0325901 treatment. The improvement in blood counts without a change in the genotype of progenitor cells suggests that MEK inhibition alters the behavior of hematopoietic cells expressing oncogenic K-RasG12D.
Bone marrow contains discrete populations of multi-potent progenitors with the potential to generate granulocytes and macrophages, megakaryocytes and erythrocytes, or all four myeloid lineages. We used flow cytometry to ascertain whether improved blood composition reflected changes in the abundance of these populations(Fig. 3A) (11). Vehicle-treated Mx1-Cre, KrasG12D mice demonstrated a bias toward granulocyte and macrophage progenitors at the expense of megakaryocyte and erythroid progenitors and progenitors of the four myeloid lineages; conversely, PD0325901 treatment restored the distribution seen in wild-type control mice, in which these populations are approximately equal in size (Fig. 3A,B). Analysis after only three daily doses of PD0325901 showed that changes in progenitor cell distributions precede any improvement in the peripheral blood (Fig. 3B,C). We did not find a significant difference in the numbers of cells expressing hematopoietic stem cell markers in the bone marrow of treated versus control mice (Fig. S1). Together, these data suggest that MEK inhibition reduces myeloproliferative neoplasia severity by modulating the behavior of hematopoietic progenitor cells, but not stem cells, in the bone marrow.
Treatment with PD0325901 also normalized the distributions of differentiating myeloid cells. Mx1-Cre, KrasG12D mice demonstrated a relative increase in the percentage of monocytes and a decrease in the percentage of intermediate granulocytes in the bone marrow and the spleen, but treatment with PD0325901 restored percentages of these populations to near normal levels (Fig. 3D, 3E) (4–10% monocytes and 20–30% intermediate granulocytes in bone marrow, and <1% of each in spleen). Similarly, untreated Mx1-Cre, KrasG12D mice progressively lose the TER119+ erythroid population in the marrow, and accumulate immature erythroid cells in the spleen (12). Treatment with PD0325901 partially corrected the distribution of differentiating cells in the bone marrow (Fig. 3F), and eliminated the massive accumulation of splenic erythroblasts (Fig. 3G). Collectively, these studies demonstrate that MEK inhibition alleviates aberrant myelopoiesis and ineffective erythropoiesis in Mx1-Cre, KrasG12D mice. PD0325901 treatment had no adverse hematological effects, such as anemia or leukopenia, in wild-type mice, although we observed an increase in splenic erythroid cells, suggesting that MEK inhibition may impose some stress on normal erythropoiesis. (Fig. 3G).
Bone marrow cells from patients with JMML and CMML and from Mx1-Cre, KrasG12D mice display a characteristic hypersensitive growth pattern, resulting in greater production of granulocyte-macrophage colonies in response to GM-CSF (3, 13). CFU-GM progenitors from Mx1-Cre, KrasG12D mice treated with PD0325901 remained hypersensitive to GM-CSF in vitro (Fig. 4A), which is consistent with retention of the activated KrasG12D allele (Fig. 2F). Similarly, we did not observe selective inhibition of Kras mutant CFU-GM progenitor colony growth over a range of PD0325901 concentrations in the presence of a saturating concentration of GM-CSF (Fig 4B). Nevertheless, some specificity was observed in cultures containing a low concentration of PD0325901 (0.1 μM). Whereas this had only modest effects on total number of colonies when GM-CSF was present at saturating levels (Fig. 4C), PD0325901 0.1 μM both abrogated growth of cytokine-independent myeloid colonies and restored a normal response to increasing concentrations of GM-CSF (Fig. 4D). At saturating doses of GM-CSF, a low concentration of PD0325901 (0.1 μM) was sufficient to normalize the numbers and types of cells within the colonies (Fig. 4E,F). Together, these data show that a low concentration of PD0325901 is sufficient to overcome the dominant effects of oncogenic KrasG12D expression and impart a normal program of proliferation and differentiation in primary myeloid progenitors. A structurally distinct MEK inhibitor, PD98059, similarly inhibited the growth of cytokine-independent colonies, confirming that MEK is the physiological target of PD0325901 in this assay (Fig. S2) (14).
Patients with well-differentiated forms of cancer may derive substantial clinical benefit from therapies that modify the aberrant behavior of the malignant clone, without damaging normal tissues, even if the tumor is not eradicated. In this case, the value of therapy lies in reducing the symptoms caused by the neoplasm for a meaningful time period. A classic example of this paradigm is the use of anti-proliferative agents, such as hydroxyurea, in the treatment of chronic myelogenous leukemia (CML), a myeloproliferative neoplasm driven by the aberrant fusion kinase Bcr-Abl (15). Even as highly effective small-molecule-targeted inhibitors of the Bcr-Abl kinase have markedly prolonged the survival of patients with CML, the causative mutation often remains detectable at low levels, and clinical experience supports the idea that a highly beneficial treatment does not always eliminate all mutant stem cells (16). In our studies, Mx1-Cre, KrasG12D mice treated with PD0325901 demonstrated a rapid and sustained improvement in hematopoietic abnormalities, even though the KrasG12D oncogene persisted at high levels in the bone marrow. This is similar to patients with myeloproliferative neoplasms that undergo clinical, but not molecular, remissions. An implication of this idea is that a reduction in oncogene ‘allele burden,’ which is frequently used as a surrogate of disease response in clinical trials, may not always be required for a therapeutic agent to demonstrate clinical utility.
Our data indicate that the dramatic hematological responses observed in Mx1-Cre, KrasG12D mice treated with the MEK inhibitor PD0325901 are not due to a purely anti-proliferative effect, as observed with conventional chemotherapy. The effects of MEK inhibition in Mx1-Cre, KrasG12D mice also contrast with those of Abl kinase inhibitors in patients with CML, where treatment provides a selective advantage to normal myeloid progenitor populations (17). Instead, the resolution of both leukocytosis and anemia indicates that PD0325901 rebalances the output of the hematopoietic system, despite continued KrasG12D expression. This “rebalancing” might indicate that MEK activity regulates the expression of genes controlling lineage choice during hematopoietic differentiation. Such a model would be consistent with studies showing that cytokine stimulation instructs lineage choice in multipotent progenitors (18), and that Ras activation promotes monocytic over granulocytic differentiation (19). Further experiments to identify genes that respond to MEK activity in multipotent cells, and to test their influence on cell fate, are required to address this hypothesis.
Curative therapy for myeloproliferative neoplasia will need to eliminate neoplastic stem cells. Whether or not Kras mutant alleles could be eliminated by MEK inhibition was not fully addressed in these experiments. Because recombination in the bone marrow of Mx1-Cre mice is usually incomplete, Mx1-Cre, KrasG12D mice can be expected to harbor a small pool of stem cells in which the conditional KrasLSL-G12D allele remains in the germline, non-expressed configuration. We have demonstrated such genetic chimerism in the bone marrow of young Mx1-Cre, KrasG12D mice (20). However, we did not determine if this persisted until the initiation of PD0325901 treatment. Interestingly, a similar question applies to patients with myeloproliferative neoplasia; some patients might lose their normal hematopoietic elements over the course of their disease. Although we did not observe emergence of cells lacking oncogenic Kras in these studies, it is possible that longer or more intense MEK inhibition might be more effective. Alternatively, inhibition of other signals evoked by mutant Kras may be required for elimination of mutant stem cells. The retention of KrasG12D stem cells in the bone marrow likely contributed to the development of T-ALL in some mice after prolonged drug treatment, because KrasG12D stem cells can create a reservoir of lymphoid progenitors that are susceptible to undergoing leukemic transformation (20). The observation that these acute leukemias arise during ongoing treatment suggests they are less dependent than myeloproliferative neoplasia on hyperactive Raf/MEK/ERK signaling.
In this study, we have provided direct evidence in vivo that aberrant hematopoiesis caused by hyperactive Ras signaling is mediated by the Raf/MEK/ERK pathway. These results are consistent with a previous investigation, which showed that constitutive activation of MEK could block erythroid differentiation in vitro, (21). However, the requirement for MEK in myeloproliferative neoplasia has been called into question by our previous biochemical studies, which revealed that endogenous levels of KrasG12D do not strongly activate Raf/MEK/ERK signaling (3, 5). This implied that aberrant hematopoiesis might be mediated by other effector pathways regulated by Ras (22). Therefore, the importance of Raf/MEK/ERK signaling in mediating the myeloproliferative phenotype provides insight into the mechanism by which excessive Ras activity causes hematologic malignancies.
Our preclinical data imply that MEK inhibitors might be useful for treating patients with JMML and CMML, by alleviating the burdens of anemia, splenomegaly, and complications from myeloid cell overproduction such as tissue infiltration. In particular, treating children with JMML prior to hematopoietic stem cell transplantation might improve their clinical status by reducing the morbidity caused by infiltration of organs with leukemic cells (1). Many adults with CMML suffer substantial morbidity owing to chronic anemia (23), and a treatment that both reduces myeloproliferation and enhances erythropoiesis, as seen with MEK inhibitors, could prove clinically beneficial.
All procedures were approved by the institutional animal research committee. F1 (129Sv/Jae x C57BL/6) Mx1-Cre, KrasLSL-G12D mice and littermates were injected intraperitoneally with 250 μg of pIpC (Sigma-Aldrich) at 4 weeks of age. Mice were randomly assigned to cohorts that received either 5 mg/kg PD0325901 (Pfizer) or hydroxypropylmethylcellulose (Sigma-Aldrich) vehicle by gavage once daily. Blood cell and reticulocyte counts were measured weekly using a Hemavet blood analyzer (Drew Scientific) and Retic-Count reagent (BD Biosciences). Mice were removed from study after 5 weeks of PD0325901 treatment, or when Hb <6 g/dL.
Nucleated bone marrow and spleen cells were stained with the following antibodies (eBiosciences, Biolegend, AbCam and BD Biosciences): lineage markers (PE-Cy7–conjugated CD3, CD4, CD5, CD8, B220, TER119, CD11b, or Gr1), FITC-CD34, PE-CD16/32, Pacific Blue-Sca1, APC-750–C-kit for stem cells and progenitors, PE-Cy7-CD11b, FITC-Gr1, PE-Ly6C for myeloid and monocytic precursor maturation (24) and FITC-CD71, PE-Ter119, PE-Cy7-C-kit, Pacific Blue-CD105 for erythroid precursor maturation (25, 26). Intracellular phosphoproteins were analyzed as described (5), with the addition of staining for CD34 and CD105. Antibodies were titrated individually. Data were collected on a LSRII flow cytometer (BD Biosciences) and analyzed with FlowJo (TreeStar).
Nucleated bone marrow and spleen cells (5 × 104 cells per culture) were suspended in M3231 methylcellulose medium (StemCell) supplemented with recombinant murine GM-CSF (PeproTech) 0.01–10 ng/ml, and/or PD0325901 0.01–25 μM, PD98059 (Cell Signaling Technology) or DMSO 0.2%. CFU-GM colonies were counted on day 7. To evaluate recombination of the KrasLSL-G12D allele, colonies were individually harvested into 10 μL distilled water, and 2 μL was analyzed by PCR (2).
Genotyping was performed using primers SD5′ (AGCTAGCCACCATGGCTTGAGTAAGTCTGCA), DT5′ IO (GTCGACAAGCTCATGCGGGTG), and LJ3′ (CCTTTACAAGCGCACGCAGACTGTAGA), with 35 cycles of 94°C (30 s), 65°C (90 s), and 72°C (60 s). This reaction yields amplicons of 500 bp for the wild-type Kras allele and 550 bp for the KrasLSL-G12D allele. Recombination of the KrasLSL-G12D allele was determined by PCR with the Advantage-GC kit (Clontech) using primers 5′-1 (GGGTAGGTGTTGGGATAGCTG) and 3′-3 (TCCGAATTCAGTGACTACAGATGTACAGAG), with 35 cycles of 95°C (30 s), 58°C (30 s), and 72°C (30 s); the wild-type Kras allele yields a 285-bp amplicon, and the recombined KrasLSL-G12D allele yields a 315-bp amplicon. The non-recombined KrasLSL-G12D allele does not amplify under these conditions.
Data were analyzed using Prism 4.0 software (GraphPad). Student’s t tests, 2-tailed, unpaired, were used to compare blood cell counts and spleen weights after treatment. WBC data were log transformed to correct heteroscedasticity; otherwise, Welsh’s correction was applied when variances were unequal. For bone marrow and spleen population frequencies, and for colony counts, effects of genotype/treatment cohort and populations were analyzed by 2-way ANOVA, and Bonferroni post-test comparisons were performed within each population against the wild-type/vehicle cohort. Kaplan Meier survival analysis was analyzed by the log rank test.
Inhibiting the Raf/MEK/ERK pathway reverses the harmful effects of oncogenic Kras on hematopoietic differentiation, suggesting a strategy for treating myeloproliferative neoplasms.
We are indebted to D. Tuveson and T. Jacks for KrasLSL-G12D mice, and to J. Sebolt-Leopold and Pfizer, Inc. for PD0325901.
Funding: This work was supported by grants from the National Institutes of Health (K08CA103868, K08CA119105, R37CA72614, U01CA84221, and T32CA128583); the Leukemia and Lymphoma Society (LLS7019-04); the V Foundation for Cancer Research; the Aplastic Anemia and Myelodysplastic Syndrome Foundation; the MPD Foundation; the Ronald McDonald House Charities of Southern California/Couples Against Leukemia; the Jeffrey and Karen Peterson Family Foundation; and the Frank A. Campini Foundation. N.L. was supported by the UCSF Helene and Charles Linker Fellowship.
Author contributions: N.L. participated in the design and execution of all experiments. M.F.G. designed and performed initial experiments to evaluate the phenotypic effects of PD325901 in Mx1-Cre, KrasG12D mice. J.O.L. helped define pharmacokinetic and pharmacodynamic properties of PD0325901 and its toxicity in this mouse strain. W.X.H. designed and performed colony assay experiments. J.K.A. maintained mouse strains, and designed and performed pharmacodynamic studies of PD0325901 using flow cytometry. K.S. was instrumental in obtaining PD0325901, aided in the design and interpretation of experiments, and provided ongoing suggestions and discussions throughout the study. B.S.B. conceived and supervised this study, was involved in the design and evaluation of all experiments, and wrote the manuscript with assistance from N.L. and K.S.
Competing interests: None of the authors have competing interests. J.O.L. was not affiliated with Genentech while contributing to this work.