The CSF3R T618I mutation causes a lethal neutrophilic neoplasia in mice that is responsive to therapeutic JAK inhibition

We have recently identified activating mutations in the colony-stimulating factor 3 receptor (CSF3R; GCSFR) as targetable genetic drivers in ∼60% of chronic neutrophilic leukemia (CNL) and atypical (BCR-ABL–negative) chronic myeloid leukemia (aCML),¹  2 related chronic leukemias characterized by increased numbers of mature neutrophils and the absence of BCR-ABL. Subsequently, the prevalence of CSF3R mutations in cases meeting strict diagnostic criteria for CNL was found to be as high as 83%, with a lower frequency observed in cases meeting strict aCML diagnostic criteria.²  CSF3R mutations are found in approximately 1% of de novo AML^1,3  and can be acquired in patients with severe congenital neutropenia (SCN), which is correlated to an increased risk for development of AML.⁴ 

There are 2 classes of CSF3R mutations: truncations of the cytoplasmic domain and membrane proximal point mutations, including T618I.^1,3  CSF3R truncation mutations are the mutation type nearly always observed in SCN and lead to enhanced cell surface expression and signaling of the receptor.⁵  In contrast, membrane proximal mutations (particularly T618I) are the predominant mutation type observed in CNL/aCML and confer ligand-independent growth.^1,3  CSF3R mutations can activate downstream SRC- or JAK-family tyrosine kinase pathways, making these kinase pathways promising therapeutic targets for the treatment of leukemia patients with CSF3R mutations.¹ 

Transgenic mice harboring CSF3R truncation mutations do not develop leukemia,⁶  but the truncation mutations can enhance leukemia progression in the context of another genetic driver.⁷  The T618I mutation has greater cell transformation capacity than the truncation mutations in vitro,¹  but it is not known whether the T618I mutation alone is sufficient to drive CNL or aCML. In this study we created a CSF3R^T618I bone marrow transplant mouse model that results in expansion of neutrophils in the peripheral blood and bone marrow, neutrophil infiltration in the spleen and liver, and eventual death, demonstrating that the T618I mutation alone is capable of driving neutrophil expansion. This neutrophilic expansion is dependent on JAK kinase signaling, because therapeutic JAK inhibition lowers white blood cell (WBC) count and reduces spleen size.

Human CSF3R transcript variant 1 (NM_000760.2) pDONR vector was purchased from GeneCopoeia. CSF3R^T618I mutation was made using the QuikChange II XL site-directed mutagenesis kit (Agilent Technologies). The Gateway Cloning System (Invitrogen) was used to clone CSF3R^WT and CSF3R^T618I into the MSCV-IRES-green fluorescent protein (GFP) plasmid.

Wild-type BALB/C mice (000651) were purchased from Jackson Labs. Retroviral infection and transplantation was performed as previously described.⁸  All mouse work was performed with approval from the Oregon Health & Science University Institutional Animal Care and Use Committee.

Mice were administered 90 mg/kg ruxolitinib phosphate (ChemScene) dissolved in 5% dimethyl acetamide, 0.5% methylcellulose, or vehicle alone by oral gavage twice daily as previously described.⁹ 

After red blood cell lysis, cells were stained with the following antibodies for 20 minutes at 4°C: PE-CD3 clone 145-2C11 (eBioscience), PerCP Cy5.5-CD19 clone HIB1g (BD PharMingen), APC-CD11b clone M1/70 (eBioscience), and E450-Gr-1 clone RB6-8C5 (eBioscience). All flow cytometry was performed on an Aria III (BD Biosciences). Data were analyzed using FlowJo software (TreeStar).

Peripheral blood was collected from live animals immediately into fix/lyse buffer (BD Biosciences) for 15 minutes at 37°C. Cells were then permeabilized with methanol and stained with PE-pSTAT3 (pY705) (BD Biosciences) and analyzed by flow cytometry.

Spleens, livers, and femurs were removed at necropsy and fixed in 10% zinc formalin. Fixed tissues were sectioned and stained using hematoxylin and eosin by the Histopathology Shared Resource at Oregon Health & Science University.

To determine whether CSF3R^T618I is sufficient to drive neoplastic expansion of neutrophils, we transplanted bone marrow expressing CSF3R^T618I or CSF3R^WT into irradiated mice. CSF3R^WT was chosen for comparison to control for any effects of ectopic CSF3R expression. Blood counts were monitored 1 to 2 times per week (Figure 1A). The CSF3R^T618I mice had an initial transient leukocytosis predominantly comprised of granulocytes (Figure 1A-C), trending back to normal by day 33 post transplant. The initial leukocytosis was specific to CSF3R^T618I mice, indicating that it is a direct effect of the mutation rather than simply overexpression of CSF3R. At day 47, the CSF3R^T618I mice had a dramatic rise in WBCs, again comprised of predominantly mature granulocytes (Figure 1A-C; supplemental Figure 1, available on the Blood Web site). By day 90, this leukocytosis was persistent and uniformly fatal in mice transplanted with hematopoietic cells expressing CSF3R^T618I (Figure 1D). Expression of CSF3R^WT did not lead to significant leukocytosis or morbidity in any mice. CSF3R^T618I activated signaling downstream of the receptor, leading to an increase in phospho-STAT3 relative to mice expressing CSF3R^WT (Figure 1E).

Bone marrow of both CSF3R^T618I and CSF3R^WT mice exhibited abnormal hypercellularity; this was most pronounced in the CSF3R^T618I mice (supplemental Figure 2). The marrow was composed mostly of mature granulocytes (supplemental Figure 2). CSF3R^T618I and CSF3R^WT mice also had infiltration of mature granulocytes into the spleen and liver (supplemental Figure 2), which was more pronounced in the CSF3R^T618I than the CSF3R^WT mice (supplemental Figure 2). The ability of CSF3R^T618I as well as CSF3R^WT to enhance production of mature myeloid cells is consistent with the normal physiological role of CSF3R in promoting neutrophil production and maturation.^10,11  The T618I mutation in CSF3R further augments this signaling—possibly by conferring ligand independence to the receptor³ —resulting in a more exaggerated proliferation of granulocytes. The dramatic granulocytic expansion in mice transplanted with cells expressing CSF3R^T618I led to death in all cases, whereas the granulocytic expansion in the CSF3R^WT mice was not sufficient to cause death.

We previously demonstrated that activating CSF3R mutations lead to preferential downstream signaling via JAK kinases, and a CNL patient carrying a JAK activating CSF3R^T618I mutation showed marked clinical improvement after administration of the JAK1/2 inhibitor ruxolitinib.¹  To determine whether the granulocytic expansion seen in CSF3R^T618I mice is dependent upon the JAK kinase pathway, we tested the effect of ruxolitinib in a second cohort of CSF3R^T618I mice. Oral administration of ruxolitinib (90 mg/kg 2×/d) or vehicle was started at day 12 post transplant, at which time mice already exhibited leukocytosis. Ruxolitinib treatment resulted in a prompt reduction in WBC count and a decrease in spleen weight (Figure 2A-C). Consistent with its ability to improve constitutional symptoms such as fatigue and early satiety in myelofibrosis,^12,13  ruxolitinib-treated mice had increased body weight compared with vehicle-treated mice (Figure 2D). This demonstrates that the pathologic expansion of granulocytes in the CSF3R^T618I mouse model is sensitive to JAK inhibition and warrants further investigation into the therapeutic use of JAK inhibitors in patients with CNL harboring the CSF3R^T618I mutation.

Here we demonstrate that expression of CSF3R^T618I in hematopoietic cells is sufficient to drive a lethal granulocytic expansion in mice. These data, in addition to the high prevalence of the CSF3R^T618I mutation in patients with CNL,^1,2  implicate CSF3R^T618I as an important genetic driver of neutrophilic leukemias. Moreover, disease in the mouse model responds to therapeutic JAK inhibition, consistent with our previously observed disease response in a CNL patient treated with ruxolitinib.¹  The CSF3R^T618I bone marrow transplant provides a tractable mouse model of neutrophilic leukemia that will be useful for further studying the underlying biology of and therapeutic interventions in neutrophilic myeloproliferative neoplasms.

The authors thank the histopathology shared resource at the Oregon Health & Science University for preparing the histology slides and performing hematoxylin and eosin staining.

This work was supported in part by The Leukemia & Lymphoma Society, the Training Program in Molecular Hematology (5T32HL007781), a Leukemia & Lymphoma Society Fellow Award, and a Medical Research Foundation Early Clinical Investigator Award (J.M.); and grants from the V Foundation for Cancer Research, the William Lawrence and Blanche Hughes Fund, the Gabrielle’s Angel Foundation for Cancer Research, and the National Cancer Institute, National Institutes of Health (5 R00CA151457-04) (J.W.T.).

Contribution: A.G.F., S.B.L., L.R.R., A.A., M.M.L., M.L.A., J.D.M., and J.E.M. performed experiments and analyzed data; A.G.F., J.E.M., B.J.D., and J.W.T. designed experiments; and A.G.F., J.E.M., and J.W.T. wrote the manuscript.

Conflict-of-interest disclosure: J.W.T. and A.G.F. both receive funding for administration of clinical trials from Incyte, manufacturer of ruxolitinib. The other authors declare no competing financial interests

The current affiliation for A.G.F. is Division of Hematology/Oncology, University of California-Irvine, Irvine, CA.

Correspondence: Jeffrey W. Tyner, Oregon Health & Science University, BRB 511, Mailcode L592, 3181 SW Sam Jackson Park Rd, Portland, OR 97239; e-mail: tynerj@ohsu.edu.