A germline point mutation in the MYC-FBW7 phosphodegron initiates hematopoietic malignancies

Abstract

Oncogenic activation of MYC in cancers predominantly involves increased transcription rather than coding region mutations. However, MYC-dependent lymphomas frequently acquire point mutations in the MYC phosphodegron, including at threonine 58 (T58), where phosphorylation permits binding via the FBW7 ubiquitin ligase triggering MYC degradation. To understand how T58 phosphorylation functions in normal cell physiology, we introduced an alanine mutation at T58 (T58A) into the endogenous c-Myc locus in the mouse germline. While MYC-T58A mice develop normally, lymphomas and myeloid leukemias emerge in ∼60% of adult homozygous T58A mice. We found that primitive hematopoietic progenitor cells from MYC-T58A mice exhibit aberrant self-renewal normally associated with hematopoietic stem cells (HSCs) and up-regulate a subset of MYC target genes important in maintaining stem/progenitor cell balance. In lymphocytes, genomic occupancy by MYC-T58A was increased at all promoters compared with WT MYC, while genes differentially expressed in a T58A-dependent manner were significantly more proximal to MYC-bound enhancers. MYC-T58A lymphocyte progenitors exhibited metabolic alterations and decreased activation of inflammatory and apoptotic pathways. Our data demonstrate that a single point mutation stabilizing MYC is sufficient to skew target gene expression, producing a profound gain of function in multipotential hematopoietic progenitors associated with self-renewal and initiation of lymphomas and leukemias.

Keywords: FBW7; MYC; hematopoiesis; leukemia; lymphoma; progenitor cells; protein stability; self-renewal.

Figure 1.

Myc^T58A/T58A mutant mice exhibit increased MYC stability and are sensitized to late-onset hematopoietic malignancies. (A) Protein isolated from whole-cell extracts of spleens or thymuses from Myc^T58A/T58A and littermate control mice were analyzed by Western blotting. Littermate pairs were loaded into adjacent lanes (indicated by brackets) by an investigator blind to the genotypes and blotted, and blots were detected using antibody to Myc, with Histone H3 used as a loading control. (B) Single-cell suspensions from Myc^T58A/T58A and littermate control (Myc^+/+) mouse thymuses were suspended in culture media in the presence of cycloheximide for the time indicated, and Western blots for Myc and β-tubulin (loading control) were performed. One of three independent experiments with similar results is shown. The calculated half-life (T_1/2) (see Supplemental Fig. S1D) from the three experiments is indicated. (C) Twenty-one Myc^T58A/T58A, 17 Myc^T58A/+, and 18 littermate control Myc^+/+ mice were allowed to age until the development of disease. Mice were necropsied, and malignancies were confirmed using flow cytometry and pathological analysis. The data shown are a Kaplan–Meier plot of tumor-free survival. The asterisk denotes significant P-values for Myc^T58A/T58A and Myc^+/T58A mice compared with Myc^+/+ mice. Seven of 21 Myc^T58A/T58A mice developed myeloid leukemias, and four of 21 mice developed B-cell lymphomas (with three of uncertain lineage). One of 17 Myc^+/T58A mice developed myeloid leukemias, and one developed B-cell lymphomas (with two of uncertain lineage). A summary of observed malignancies is shown in the table (see also Supplemental Fig. S1I,J). (D–F) Mice that exhibited a detectable tumor mass or were moribund were sacrificed. (D) Tissues and blood smears were prepared from one normal mouse and two malignant mice (one relatively undifferentiated) and stained (Romanowsky stain). Magnifications of 20× and 40× (boxed region from 20×) are shown. Single-cell suspensions from malignant Myc^T58A/T58A or control Myc^+/+ mice were prepared from blood, bone marrow, and spleens and analyzed by flow cytometry using antibodies specific for the indicated cell surface markers (Mac1 and Gr1) to identify differentiated myeloid cells. Cell populations were gated on live cells (DAPI-negative). Flow data are shown for Myc^+/+ and relatively differentiated and undifferentiated myeloid leukemias from Myc^T58A/T58A mice. Flow cytometry for myeloid markers is shown. (E) Representative flow cytometry for a B-cell lymphoma from a malignant Myc^T58A/T58A mouse. (F) Tissues from the malignant animal shown in E (liver, lung, and spleen) were stained (hematoxylin and eosin) for microscopic analysis.

Figure 2.

Aberrant self-renewal of hematopoietic progenitors from Myc^T58A/T58A mice. (A) Bone marrow cells were isolated from Myc^T58A/T58A and Myc^+/+ control littermate mice and labeled with fluorescent antibodies to enumerate hematopoietic stem cells (Lin⁻ckit⁺Sca1⁺CD150⁺). The data are expressed as total number of HSCs per mouse. Four mice per genotype are represented on the graph. (B) Bone marrow hematopoietic progenitors were cultured from either Myc^T58A/T58A or Myc^+/+ mice in clonogenic methylcellulose assays in the presence of cytokines to support the growth of myeloid progenitor colonies. After 10 d in culture, individual colonies were picked and replated as secondary cultures supplemented with cytokines. Secondary colony formation was scored at day 10 as replatable CFU, and the ratios of replatable colonies compared with total colonies plated are shown (y-axis). Asterisks depict significance as assessed by t-test. The data shown are representative of two independent experiments. (C–E) Hematopoietic stem and progenitor cells from the bone marrow of three pooled Myc^T58A/T58A mice and Myc^+/+ control littermates were sorted by flow cytometry. Hematopoietic stem cells (HSCs) and one population of multipotential progenitors (called MPP1) were sorted based on expression of the indicated cell surface markers (see Supplemental Fig. S2K). The sorted stem or progenitor cells (CD45.2) were then mixed with competitor cells (CD45.1) and transplanted into lethally irradiated recipient mice. Antibodies specific to the two CD45 isoforms were used to determine the levels of reconstitution from the indicated sorted cell populations. The data are shown as percent CD45.2 donor. Asterisks depict significance as assessed by t-test. The data shown are a representative of two independent experiments with similar results.

Figure 3.

Single-cell multiome analysis comparing Myc^+/+ and Myc^T58A/T58A hematopoietic stem and progenitor cells. (A) Hematopoietic stem and progenitors (LSK-sorted cells) from freshly isolated bone marrow of either Myc^+/+ or Myc^T58A/T58A mutant Myc mice were sorted and barcoded to identify single cells using the 10X Genomics platform for single-cell multiome (RNA-seq and ATAC-seq) analyses. Two mice per genotype were used in two independent experiments to generate four data sets (four mice per genotype, for a total of eight mice). After alignment and analysis using R packages Seurat and Signac (based on RNA content and removal of duplicate barcodes), 7929 cells were included in the analysis. The data sets were then integrated (to account for batch effects) and clustered based on single-cell RNA-seq of Myc^+/+ and Myc^T58A/T58A. Following dimension reduction, progenitor cell populations were identified and labeled as detailed in the Materials and Methods. HSCs were set at 0, and cells were ordered along either lymphoid or myeloid pseudotimes. (B) The number of differentially expressed (Diff.genes), up-regulated (Up.Genes), and down-regulated (Down.Genes) genes, comparing T58A mutant with WT cells, was determined for each population of progenitor cells. Differentially expressed genes were determined by Wilcoxon rank sum analysis (P < 0.05). (C) Venn diagram depicting differentially expressed genes in three different hematopoietic progenitor cell populations (MPP1_LPrimed, MPP2_Prolif, and MPP3_MPrimed). Eleven Myc target genes are shown that were up-regulated in Myc^T58A/T58A mutant cells in all analyzed progenitor cell populations. (D) Heat map of single cells showing the expression of selected genes found to be up-regulated in Myc^T58A/T58A mutant progenitor populations. (E) Consensus binding motif activity in single-cell ATAC-seq in progenitors in Myc^+/+ and Myc^T58A/T58A mutant mice. (F) Hematopoietic progenitor cells were purified from the bone marrow of Myc^+/+ or Myc^T58A/T58A mice (LSK sorting). Sorted cells were then cultured in wells coated with Delta1 ligand to activate the Notch pathway in the presence of cytokine cocktail (see the Materials and Methods). After 2 wk of culture, increasing amounts of GM-CSF were added to the cultures to promote myeloid differentiation. Differentiation was assessed by flow cytometry staining for Sca1/ckit (Sca1/ckit⁺, primitive), which is retained on primitive cells, and Mac1/Gr1/F480 (myeloid, differentiated) to identify differentiated cells. Asterisks depict significance as assessed by t-test. The data shown are representative of two independent experiments with similar results.

Figure 4.

RNA-seq analysis comparing Myc^+/+ with Myc^T58A/T58A mutant pre-B and mature B cells. (A–C) IL-7-stimulated B cells derived from bone marrow were sorted from paired littermate WT or T58A mutant mouse bone marrow. RNA was extracted from cells after stimulation with IL-7 for 48 h, and libraries were prepared and barcoded. After sequencing, analysis was performed using EdgeR. (A) Differentially expressed genes are depicted in the volcano plot. Genes with decreased expression in Myc^T58A/T58A cells are shown in blue, and genes with increased expression are shown in red, with each gene plotted as the log₂ fold change (x-axis) against the −log₁₀ transformation of the false discovery rate (FDR; y-axis). (B,C) Enrichment analysis (using EnrichR package) was performed on genes with increased expression in Myc^T58A/T58A (B) and genes with decreased expression (C). (B, top) A plot depicting enriched pathways from MSigDB hallmark and reactome gene sets. (Bottom) A volcano plot showing differentially expressed genes involved in glycolysis (blue symbols) and some genes labeled in red. (C, top) A plot showing enrichment of down-regulated genes for pathways from MSigDB hallmark and ENCODE/CHEA ChIP-bound transcription factors. (Bottom) A volcano plot showing differentially expressed genes involved in interferon response, which are represented by blue symbols. (D) Flow cytometry analysis of glucose uptake potential was assessed by incorporation into cells of fluorescent-labeled NBDG (right panel) and Mitotracker red (left panel) in Myc^+/+ and Myc^T58A/T58A mutant cells stimulated with IL-7. The fluorescent intensity (x-axis) is plotted against cell number (y-axis). The data shown are representative of three independent experiments with similar results. (E) IL-7-stimulated cells were untreated or treated with either mitochondrial inhibitors (metformin or oligomycin) or glucose uptake inhibitor (2-DG) (x-axis). Cells were counted and plotted as percent of control untreated cells (y-axis). Myc^+/+ cells are shown in blue, and Myc^T58A/T58A cells are shown in red. The results are representative of three independent experiments. (F–H) B cells were sorted from Myc^+/+ or Myc^T58A/T58A mutant mouse spleens using anti-B220-coated magnetic beads (MACS). RNA was extracted after stimulation with LPS for 48 h as described above. Paired analysis was performed using the EdgeR package. Genes with decreased expression in Myc^T58A/T58A cells are shown in blue, and genes with increased expression are shown in red, with each gene plotted as the log₂ fold change (x-axis) against the −log₁₀ transformation of the false discovery rate (FDR; y-axis). (G,H) Enrichment analysis (using EnrichR package) was performed on genes with increased expression in Myc^T58A/T58A (G) and genes with decreased expression (H). (G, top) A plot depicting enriched pathways (from GO biological process). (Bottom) A volcano plot showing differentially expressed genes involved in ribosomal protein biogenesis (blue symbols) and some genes labeled in red. (H, top) A plot depicting enrichment of genes with decreased expression (MSigDB hallmark gene sets). (Bottom) A volcano plot showing differentially expressed genes involved in the unfolded protein response (blue symbols).

Figure 5.

Genomic analysis of Myc at promoters and enhancers in Myc^+/+ versus Myc^T58A/T58A cells. Pre-B or mature B cells sorted from the bone marrow and spleens of Myc^+/+ and Myc^T58A/T58A mice were stimulated with IL-7 (for marrow-derived pre-B cells; top panels) and LPS (for mature B cells; bottom panels) for 48 h, and 1 million cells were used for auto CUT&RUN to detect Myc, H3K27ac, and H3K3me2, followed by barcoding, sequencing, and alignment. (A) MYC-bound genes are shown (dark-blue circles) on the volcano plots depicting the RNA-seq data sets (also shown in Fig. 4) comparing gene expression in Myc^T58A/T58A versus Myc^+/+ B cells stimulated with IL-7 (left panel) and LPS (right panel). Each gene is plotted as the log₂ fold change (x-axis) against the −log₁₀ transformation of the FDR (y-axis). Differentially expressed genes are scattered above the dotted horizontal line (Padj > 0.05). Genes labeled in red are a sample of differentially expressed genes. The results shown are representative of two biological replicates. (B) Genomic tracks showing MYC, H3K27ac, Pol2, and control IgG peaks enriched in CUT&RUN and ChIP-seq (for Pol2) for Hk2 and Pkm genomic loci that have increased expression in Myc^T58A/T58A mutant cells. Tracks for Myc^+/+ cells are shown in blue, and tracks for Myc^T58A/T58A cells are shown in red. (C) Plots of normalized read counts (y-axis) for Myc were generated centered on the TSSs of all genes (left panel), genes with increased expression (middle panel), and genes with decreased expression (right panel). Plots are shown for WT cells (blue) and T58A cells (red). (D) Enhancers were identified in WT and T58A mutant IL-7-stimulated pre-B cells (IL-7) and LPS-stimulated mature B cells by peak calls for H3K27ac and H3K4me2 CUT&RUN (see Supplemental Fig. S5F). Enhancers were defined as sites enriched for both marks (and not promoter-bound). Myc CUT&RUN signal was plotted over all identified enhancer peaks for Myc^+/+ (blue lines) and Myc^T58A/T58A (red lines) cells. (E) Enhancers that were Myc-bound were identified by Myc peak calls that overlap enhancers. The distance to the nearest Myc-bound enhancer for every gene was calculated. The log₂ fold change for every gene (determined by RNA-seq comparing T58A with WT in IL-7-stimulated pre-B cells) was then plotted (x-axis) against the distance to the nearest Myc-bound enhancer (y-axis). Differentially expressed genes are represented by filled circles, and the −log₁₀ of the adjusted FDR determined for each gene by RNA-seq is shown by red color intensity. Statistical analysis was performed using two-sample Kolmogorov–Smirnov statistics (cumulative distribution analysis) comparing the distance to the nearest Myc-bound enhancer of each differentially expressed gene (comparing T58A with WT cells) versus the distance to the nearest enhancer of a same-sized random sample of genes activated in lymphocytes but not differentially expressed (see the Materials and Methods). (F) Same as E except this distance is plotted for the same number of enhancers in E that were not Myc-bound (no peaks for Myc).

Figure 6.

Diagram depicting alterations of hematopoiesis and transcription as a consequence of Myc-T58 phosphorylation. Multipotential progenitors from Myc^T58A/T58A mice have an aberrant self-renewal analogous to HSCs and sensitized to leukemias and lymphomas. This is associated with transcriptional changes in genes important in self-renewal of hematopoietic stem and progenitor cells and leukemic cells. Hematopoietic progenitors from Myc^T58A/T58A mice are poised away from inflammatory, differentiation, and apoptotic pathways despite increased glycolytic dependence, while genes involved in these processes are proximal to Myc-bound enhancers.