Mutant IDH1 cooperates with NPM1c or FLT3ITD to drive distinct myeloid diseases and molecular outcomes

Abstract

In human acute myeloid leukemia (AML), mutations of isocitrate dehydrogenase-1 (IDH1) often co-occur with NPM1 mutations, and less frequently with FLT3 mutations. To investigate whether the effects of IDH1 mutation differ according to the specific co-occurring mutation, we generated two strains of double knock-in mutant mice. Idh1^R132H combined with Npm1c induced overt AML, whereas Idh1^R132H plus Flt3^ITD resulted in Flt3^ITD-driven myelo- or lymphoproliferation that was minimally affected by Idh1^R132H and rarely generated AML. Gene expression profiling revealed differences between Idh1^R132H;Npm1c cells and Idh1^R132H;Flt3^ITD cells and suggested altered heme metabolism and immune responses in the former. The profile of Idh1^R132H;Npm1c cells corresponded to that of human IDH-mutated AML cells, particularly those resistant to inhibitors of mutant IDH. Compared to treatment with a menin inhibitor, IDH1-targeted therapy of Idh1^R132H;Npm1c AML-bearing mice was less efficacious in improving cell differentiation and extending survival. The differential cooperation of Idh1^R132H with Npm1c vs. Flt3^ITD may have implications for the devising of subtype-specific treatments for human AML.

Keywords: FLT3; IDH1; NPM1; acute myeloid leukemia; preclinical mouse model.

Fig. 1.

Differing gross hematopoietic phenotypes of Idh1^R132;Npm1c vs. Idh1^R132;Flt3^ITD mice. (A) Kaplan–Meier survival curves of WT, Idh1^R132, Npm1c, Idh1^R132;Npm1c, Flt3^ITD, and Idh1^R132;Flt3^ITD mice. For survival curves in all Figures, P values were calculated by the log-rank test. *P < 0.05, **P < 0.01 and ***P < 0.001. ns, not significant. (B, Top) Quantitation of weights of spleens from 7 to 13 mo old WT (n = 23), Idh1^R132 (n = 7), Npm1c (n = 12), and Idh1^R132;Npm1c (n = 25) mice, and from 11 to 18 mo old WT (n = 8), Idh1^R132 (n = 5), Flt3^ITD (n = 6), and Idh1^R132;Flt3^ITD (n = 9) mice. (Bottom) Representative macroscopic images of spleens from the mice in the Top panels. For all panels of this type, values represent individual mice, with the mean ± SD also shown. For all Figures, unless otherwise indicated, data are the mean ± SD, with *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 as assessed by unpaired Student’s t test. (C and D) White blood cell (WBC) counts (C) and hemoglobin (Hb) levels (D) in 7 to 13 mo old cohorts of WT (n = 22), Idh1^R132 (n = 7), Npm1c (n = 12), and Idh1^R132;Npm1c (n = 25) mice, and in 11 to 18 mo old cohorts of WT (n = 8), Idh1^R132 (n = 5), Flt3^ITD (n = 6), and Idh1^R132;Flt3^ITD (n = 9) mice. (E) Representative images of H&E-stained bone marrow (BM) sections at two magnifications (Top and Middle), and May-Gründwald-Giemsa (MGG)-stained BM smears (Bottom), from aged mice of the indicated genotypes. Red arrowheads indicate blasts. (Scale bars, 20 µm.) (F) Representative image of an MGG-stained peripheral blood (PB) smear from an aged Idh1^R132;Npm1c mouse with leukemic disease. (Scale bar, 20 µm.) (G) Representative flow cytometric analyses of CD11b and Gr1 expression among viable BM cells from (Left) 7 to 13 mo old WT (n = 22), Idh1^R132 (n = 7), Npm1c (n = 12), and Idh1^R132;Npm1c (n = 23) mice, and (Right) 11 to 18 mo old WT (n = 8), Idh1^R132 (n = 5), Flt3^ITD (n = 6), and Idh1^R132;Flt3^ITD (n = 9) mice. For all flow cytometric data, numbers in specific compartments are the mean percentage. (H) Quantitation of frequencies of CD11b⁺Gr1⁺ and CD11b⁺Gr1⁻ cells from the data in (G).

Fig. 2.

Distinct stem/progenitor cell phenotypes exhibited by Idh1^R132;Npm1c vs. Idh1^R132;Flt3^ITD mutants. (A) Frequency of cells negative for any lineage-differentiated cell surface marker (Lin⁻ cells) among viable total BM cells from young mice of the indicated genotypes (n = 7 to 9/group). See also SI Appendix, Fig. S2A. (B) Frequency of Lin⁻ cells among viable total BM cells from (Left) 7 to 13 mo old WT (n = 23), Idh1^R132 (n = 7), Npm1c (n = 12), and Idh1^R132;Npm1c (n = 23) mice, and from (Right) 11 to 18 mo old WT (n = 7), Idh1^R132 (n = 5), Flt3^ITD (n = 5), and Idh1^R132;Flt3^ITD (n = 9) mice. (C) Frequency of LK (Lin⁻IL-7R⁻Sca1⁻cKit⁺) cells among the Lin⁻ BM cells isolated from the young mice in A. See also SI Appendix, Fig. S2A. (D) Frequencies of LK cells among the Lin^- BM cells isolated from the aged mice in B. (E) Frequencies of CMP (FcγR⁻CD34⁺), GMP (FcγR⁺CD34⁺), and MEP (FcγR⁻CD34⁻) progenitors in the LK fraction from young mice of the indicated genotypes (n = 6 to 10/group). See also SI Appendix, Fig. S2B. (F) Frequency of LSK (Lin⁻IL-7R⁻Sca1⁺cKit⁺) cells among the Lin⁻ BM cells isolated from the young mice in A. See also SI Appendix, Fig. S2A. (G) Frequencies of LSK cells among the Lin⁻ BM cells isolated from the aged mice in B. (H) Frequencies of CD150⁺CD48⁻ (LT-HSC) cells and the other indicated cell populations within LSK fractions isolated from young mice of the indicated genotypes (n = 4 to 9/group). See also SI Appendix, Fig. S2C. (I) Frequencies of CD11b⁻cKit⁺ cells and CD11b⁺cKit⁺ cells among viable BM cells from young mice of the indicated genotypes (n = 7 to 10/group). (J) Representative FCA of cKit and CD11b expression among viable total BM cells from (Left) 7 to 13 mo old WT (n = 23), Idh1^R132 (n = 7), Npm1c (n = 12), and Idh1^R132;Npm1c (n = 23) mice, and (Right) 11 to 18 mo old WT (n = 8), Idh1^R132 (n = 5), Flt3^ITD (n = 6), and Idh1^R132;Flt3^ITD (n = 9) mice. (K) Frequencies of CD11b⁻cKit⁺ and CD11b⁺cKit⁺ cells among the cells in J. (L) Colony formation assays of BM cells that were isolated from young mice of the indicated genotypes and serially plated five times in methylcellulose medium (CFU 1-5) (n = 3). (M) Schematic diagram summarizing the differing effects of the Npm1c and Flt3^ITD mutations on hematopoietic cell differentiation toward the myeloid lineage.

Fig. 3.

Distinct transcriptomic signatures of Idh1^R132;Npm1c vs. Idh1^R132;Flt3^ITD cKit⁺ cells. (A) GSEA of RNA-seq data of cKit⁺ cells isolated from mice of the indicated genotypes at 3 or 6 mo of age. Significant enrichment (green) and significant depletion (red) are based on NES for the indicated comparison (q < 0.05). Blank columns, not significant. (B) Venn diagrams showing numbers of overlapping up- or down-regulated genes in cKit⁺ cells from 3-mo-old (Left) and 6-mo-old (Right) mice of the indicated genotypes (compared to WT). Differential gene expression analysis was performed using DESeq2, and genes with fold-change >1.5 and adjusted P-value <0.01 were regarded as differentially expressed. Genes of interest are indicated.

Fig. 4.

Idh1^R132 cooperates with Npm1c to induce AML in mice. (A) Kaplan–Meier survival curves of CD45.1⁺ recipient mice transplanted with BM cells from moribund mice of the indicated genotypes. (B) Pie charts depicting the relative frequencies of diseases developing in mice of the indicated genotypes. AML was deemed to be present if disease developed rapidly and was fatal to transplant recipients within 8 wk (red). “Myeloid” was defined as disease featuring myeloid cell proliferation but not judged to be AML (orange). “B” indicates a disease featuring B cell proliferation (green). “T” indicates a disease featuring T cell proliferation (blue). (C) Identification of the indicated somatic mutations in 10 lines of Idh1^R132;Npm1c mice exhibiting AML. VAF, variant allele frequency, with three different ranges shown. (D) Kaplan–Meier survival curves of recipient mice transplanted with cKit⁻ cells, CD11b⁺cKit⁺ cells, GMP, or LSK cells from Idh1^R132;Npm1c AML mice. Two mouse lines of transplantable Idh1^R132;Npm1c AML (#5, Top; #10, Bottom) were used. (E) Similarities in gene expression signatures between human primary AML cells bearing the indicated mutations and cKit⁺ cells from 3- or 6-mo-old Idh1^R132;Npm1c mice. Shown are gene set variation analysis (GSVA) scores of human AML samples from a published Oregon Health & Science University (OHSU) cohort (Nature, 2018) as determined using the murine differential gene expression signature of Idh1^R132;Npm1c vs. WT. Human samples are combinatorially partitioned by their NPM1, IDH1, and FLT3 genotypes. (F) GSEA based on transcriptomic data from human AML samples (OHSU, Nature 2018) bearing the indicated gene mutations vs. human AML with no FLT3, NPM1, or IDH1 mutations. Significant depletion (red) is based on NES for the indicated comparison (q < 0.05). Blank columns, not significant.

Fig. 5.

Idh1^R132 has a modest effect on Npm1c-induced phenotypes through epigenetic modifications. Plot depicting loci with hyper- or hypomethylated CpGs, as determined by bisulfite-sequencing, in cKit⁺ cells from 6-mo-old mice of the indicated genotypes.

Fig. 6.

Effects of combined Idh1^R132 and Npm1c mutations on responses to chemotherapy and IDH inhibitor treatments, and potential therapeutic alternatives for mice bearing AML with Idh plus Npm1c mutations and showing IDHi resistance. (A and B) Colony formation by WT, Idh1^R132;Npm1c and Idh1^R132;Flt3^ITD BM cells that were plated in methylcellulose medium containing the indicated concentrations of (A) daunorubicin or (B) Ara-C. Data are expressed relative to untreated controls. (C) Quantitation of AG-120 concentrations in plasma, BM, and spleens of recipients that were transplanted with Idh1^R132;Npm1c AML cells and treated with either vehicle or 150 mg/kg AG-120-SDD by oral gavage twice daily from day 2 to 18 (n = 3 to 4/group). (D) D-2HG levels in PB from the mice in C (n = 3 to 4/group), and untreated Vav-Cre WT mice (control; n = 2). (E) Percentage of CD45.2⁺ (donor-derived) cells among total CD45⁺ cells [CD45.1⁺ (recipient-derived) cells plus CD45.2⁺ cells] in BM from the mice in C (n = 4/group). (F) Representative FCA of CD11b and Gr1 expression among CD45.2⁺ (donor-derived) cells in BM of the mice in C (n = 4/group). (G) Frequencies of CD11b⁻Gr1⁻ cells (Left) and CD11b⁺Gr1⁺ cells (Right) from the data in F. (H) Frequency of cKit⁺ cells among CD45.2⁺ (donor-derived) cells in BM of the mice in C (n = 4/group). (I–K) CBC parameters in PB from the mice in C (n = 4/group). WBC count (I), hemoglobin (Hb) levels (J), and platelet (PLT) counts (K) are shown. (L) Kaplan–Meier survival curves of CD45.1 recipient mice transplanted with Idh1^R132;Npm1c AML cells and treated with either vehicle or AG-120-SDD as indicated from day 2 to 18 (n = 6/group). (M) Schematic diagram of experiment in which CD45.1 mice were transplanted with Idh1^R132;Npm1c AML cells (CD45.2) and treated with either vehicle or the indicated drugs in separate cohorts. (N) Kaplan–Meier survival curves of the treated AML mice in M (n = 6/group). (O) WBC count on day 18 in PB from transplanted AML mice treated with the indicated drugs as in M (n = 4/group). (P) Percentage of CD45.2⁺ (donor-derived) cells among total CD45⁺ cells [CD45.1⁺ (recipient-derived) cells plus CD45.2⁺ cells] in BM from the mice in M (n = 4/group). (Q) Frequency of cKit⁺ cells among viable cells in BM from the mice in M (n = 4/group). (R) D-2HG levels in PB from the mice in M (n = 4/group). (S) Frequencies of CD11b⁻Gr1⁻ cells (Left) and CD11b⁺Gr1⁺ cells (Right) among CD45.2⁺ (donor-derived) cells in BM from the mice in M (n = 4/group). (T) Hb levels in the mice in M (n = 4/group). (U) Body weights of the mice in M (n = 4/group) over 18 d.