Myeloid lineage enhancers drive oncogene synergy in CEBPA/CSF3R mutant acute myeloid leukemia

Abstract

Acute Myeloid Leukemia (AML) develops due to the acquisition of mutations from multiple functional classes. Here, we demonstrate that activating mutations in the granulocyte colony stimulating factor receptor (CSF3R), cooperate with loss of function mutations in the transcription factor CEBPA to promote acute leukemia development. The interaction between these distinct classes of mutations occurs at the level of myeloid lineage enhancers where mutant CEBPA prevents activation of a subset of differentiation associated enhancers. To confirm this enhancer-dependent mechanism, we demonstrate that CEBPA mutations must occur as the initial event in AML initiation. This improved mechanistic understanding will facilitate therapeutic development targeting the intersection of oncogene cooperativity.

Fig. 1

CSF3R^T618I and CEBPA mutations cooperate in vitro. a Colony assay from mouse bone marrow transduced with CSF3R^T618I and either CEBPA^WT (W), CEBPA^F82fs (F), or CEBPA^V314VW (V) (n = 3/group). b Representative images for a. c Serial replating of colony assay in a. d Colony assay from mouse bone marrow transduced with CSF3R^T618I and either empty vector or CEBPA^K313KR. e Example images from cytospins of conditions in a grown for 72 h in cytokine-free liquid culture (scale bar represents 10 μm). f Colony-forming assay from mouse bone marrow transduced with the combination of CSF3R^T618I (T), CEBPA^F82fs (F), CEBPA^V314VW (V), two oncogene combinations and empty vector controls. In all, 1700 cells were plated per condition. g Representative images for f. h Colony assay from mouse bone marrow transduced with Empty Vector (E) JAK3^WT (W), JAK3^M511I (M), or CEBPA^V314VW. i Representative images of colony assay for h. j Colony assay from mouse bone marrow from CEBPA KO or Cre- littermate control bone marrow transduced with Empty Vector or CSF3R^T618I. k Representative colony assay images for j. In all cases n = 3/group and values are represented as mean with error bars representing SEM. Significance of comparisons assessed by ANOVA with Sidak post-test *p < 0.05, ***p = 0.001, ****p < 0.0001). Source data are provided as a Source Data file.

Fig. 2

Mutant CEBPA blocks myeloid differentiation in response to CSF3R. a Venn diagram of differentially expressed genes from RNA-seq on lineage-negative mouse bone marrow transduced with empty vector, CSF3R^T618I, CEBPA^V314VW or the combination of oncogenes. b Hierarchical clustering of interacting genes. c, d Select enriched gene sets in CSF3R^T618I vs. other categories NES = normalized enrichment score (q < 0.05). GSEA p value calculated by empirical permutation test and FDR adjusted. e Motif enrichment at promoters of differentially expressed (DE) genes. Top five motifs per category with q < 0.05 shown. P values generated via comparison to binomal distribution and FDR adjusted. f Expression of differentially expressed HOX genes in pediatric AML patients harboring CEBPA mutations. g Expression of genes differentially expressed in both murine and pediatric human CSF3R/CEBPA AML as compared with CEBPA-mutant CSF3R-WT AML. h PCA analysis of pediatric CEBPA mutant AML using convergent human-mouse gene set. Cases with biallelic CEBPA mutations are indicated by CEBPA-Bi. i Expression of genes differentially expressed in both murine and adult human CSF3R/CEBPA AML as compared with CEBPA-mutant CSF3R-WT AML. j PCA analysis of adult CEBPA mutant AML using convergent human–mouse gene set. Specific CEBPA mutations are indicated by the text. Source data are provided as a Source Data file.

Fig. 3

JAK/STAT activation and CEBPA mutations cooperate in vivo. a Survival of mice transplanted with CEBPA wild type or CEBPA^K/L fetal liver cells transduced with empty vector or CSF3R^T618I (n = 7–8/group). b WBC counts from mice in A. c Spleen weights from moribund mice in a (n = 3–5/group). d Example bone marrow smears for mice in a (scale bar represents 10 μm). e Survival of mice transplanted with 100,000 cells containing Empty Vector, JAK3^M511I + Empty Vector, CEBPA^V314VW + Empty Vector, JAK3^M511I + CEBPA^V314VW, CSF3R^T618I + Empty vector, CSF3R^T618I + CEBPA^F82fs, or CSF3R^T618I + CEBPA^V314VW (n = 5–6/group). f WBC counts of mice harboring mutation combinations in e (n = 5–6/group at experiment start). g Example bone marrow smears from mice in e (scale bar represents 10 μm). Values in bar graphs are represented as mean with error bars representing SEM. Source data are provided as a Source Data file.

Fig. 4

CEBPA mutations block differentiation in response to mutant CSF3R. a Experimental strategy to directly compare the differentiation of cells harboring CSF3R^T618I and CSF3R^T618I/CEBPA^V314VW. b Bone marrow flow cytometry comparing myeloid differentiation markers in mice transplanted under conditions in a. c Quantification of populations in b. d Example bone marrow histology from mice transplanted in a (scale bar represents 10 μm). e Spleens from mice in a. f Spleen weights from mice in a. Values are represented as mean with error bars representing SEM. Significance of comparisons assessed by Students T-test or ANOVA with Sidak post-test as appropriate **p < 0.01, ****p < 0.0001. Source data are provided as a Source Data file.

Fig. 5

CEBPA mutations disrupt activation of myeloid lineage enhancers. a Number of condition-specific enhancers identified in this study. b ChIP-seq heat maps for H3K4me1 and H3K27ac at +1.5 kb around the center of CSF3R^T618I-specific enhancers, across treatment conditions. ChIP tracks display fold enrichment relative to corresponding input. c Gene ontology analysis for condition-specific enhancers. P values generated via comparison to binomal distribution and FDR adjusted. d De novo motif enrichment in condition-specific enhancers. Values in parenthesis correspond to motif match score. P values generated via comparison to binomal distribution and FDR adjusted. e The epigenetic landscape at the Nos2 locus demonstrates condition-specific promoter and enhancer activation. f Assessment of CEBPA ChIP-seq peak overlap with condition specific enhancers by χ² test with Pearson residual values plotted. Adjusted p values (by the method of Holm–Bonferroni) are displayed above bars. g Gene ontology analysis for CEBPA-independent CSF3R^T618I active enhancers. h The epigenetic landscape at the E2f2 locus demonstrates enhancer activation in response to CSF3R^T618I. Source data are provided as a Source Data file.

Fig. 6

CEBPA mutations must precede mutations in CSF3R to block differentiation. a Diagram of order of acquisition system. Oncogenes in the MIG (GFP+) vector tagged with GFP are constitutively expressed, and oncogenes in the idsRed vector are expressed only after Cre-mediated recombination. Expression of the second oncogene can be induced by culture in 4-OHT. b Colony assay from mouse bone marrow transduced with CEBPA First (MIG-CEBPA^V314VW + idsRed-CSF3R^T618I) CSF3R First (MIG-CSF3R^T618I + idsRed-CEBPA^V314VW) and CSF3R only (MIG-CSF3R^T618I + idsRed Empty) and plated in 4-OHT (n = 3/group). c Representative images of colony assay. d Human CEBPA and CSF3R expression in mouse bone marrow expressing MIG-CEBPA^V314VW, idsRed-CEBPA^V314VW, MIG-CSF3R^T618I or idsRed-CSF3R^T618I measured by TaqMan quantitative PCR (n = 3/group). e Clustering by Euclidian distance for RNA sequencing performed on lineage negative mouse bone marrow expressing CEBPA-first, CSF3R-first, or CSF3R-only (n = 3–4/group). f K-means clustering of top 750 differentially expressed genes per pairwise comparison (q < 0.05, log₂Fold Change < −1 or >1) (n = 3–4/group). g Representative enriched gene sets from GSEA performed on samples from e (n = 3–4/group). GSEA p value calculated by empirical permutation test and FDR adjusted. In all cases, values are represented as mean with error bars representing SEM. **p < 0.01, ****p < 0.0001, as measured by ANOVA with Sidak’s post-test. Source data are provided as a Source Data file.

Fig. 7

CEBPA mutations must precede mutations in CSF3R for AML initiation. a Diagram of order of acquisition system. Oncogenes in the MIG vector tagged with GFP are constitutively expressed and oncogenes in the idsRed vector are expressed only after Cre mediated recombination. Expression of the second oncogene can be induced by administration of tamoxifen. Mice were administered 75,000 GFP/RFP-positive cells. b Survival reported as time from day 1 of tamoxifen induction (n = 5/group). Statistical significance calculated by Log Rank test, **p < 0.01. c Expression of CD11b and GR-1 in bone marrow at 5 weeks in moribund CEBPA^V314VW-first mice, healthy CSF3R^T618I-first mice or at 7-weeks in leukemic CSF3R^T618I-first mice (only one animal represented by CSF3R-first leukemia, others are representative of 3/group). d Manual differentials of peripheral blood from mice 5 weeks after tamoxifen induction (n = 3/group), **p < 0.01, ****p < 0.0001. e Representative bone marrow histology (scale bar represents 10 μm). f Spleen weight of mice sacrificed at 5 weeks post tamoxifen treatment, *p < 0.05 as calculated by Student’s T-test g Representative spleen histology (scale bar represents 100 μm for upper panels and 20 μm for lower panels). In all cases, values are represented as mean with error bars representing SEM. Source data are provided as a Source Data file.

Fig. 8

CEBPA mutations initiate AML through disruption of myeloid lineage enhancers. During normal myeloid differentiation, GCSF/CSF3R signaling drives both differentiation and proliferation of myeloid precursors. The differentiative program downstream of GCSF/CSF3R signaling is dependent on CEBPA acting at myeloid lineage enhancers. Mutant CEBPA blocks activation of these enhancers preventing transcription of differentiation-associated genes. When CSF3R mutations are introduced early, they activate transcription using the native enhancer repertoire initiating differentiation. Importantly this differentiation is insensitive to subsequent introduction of mutant CEBPA.