TET2 lesions enhance the aggressiveness of CEBPA-mutant acute myeloid leukemia by rebalancing GATA2 expression

Abstract

The myeloid transcription factor CEBPA is recurrently biallelically mutated (i.e., double mutated; CEBPA^DM) in acute myeloid leukemia (AML) with a combination of hypermorphic N-terminal mutations (CEBPA^NT), promoting expression of the leukemia-associated p30 isoform, and amorphic C-terminal mutations. The most frequently co-mutated genes in CEBPA^DM AML are GATA2 and TET2, however the molecular mechanisms underlying this co-mutational spectrum are incomplete. By combining transcriptomic and epigenomic analyses of CEBPA-TET2 co-mutated patients with models thereof, we identify GATA2 as a conserved target of the CEBPA-TET2 mutational axis, providing a rationale for the mutational spectra in CEBPA^DM AML. Elevated CEBPA levels, driven by CEBPA^NT, mediate recruitment of TET2 to the Gata2 distal hematopoietic enhancer thereby increasing Gata2 expression. Concurrent loss of TET2 in CEBPA^DM AML induces a competitive advantage by increasing Gata2 promoter methylation, thereby rebalancing GATA2 levels. Of clinical relevance, demethylating treatment of Cebpa-Tet2 co-mutated AML restores Gata2 levels and prolongs disease latency.

Fig. 1. TET2 mutations impair outcome for patients with CEBPA-mutant AML.

a Frequency of co-occurring mutations in CEBPA^DM AML cases, data aggregated from published cohorts^–, (321–557 cases; detailed in Supplemental Table 1). b Overall survival of CEBPA^DM patients with wild-type (TET2^WT; 84 patients) or mutated TET2 (TET2^MUT; 35 patients). The data were analyzed by Mantel-Cox Log-rank test. c Volcano plot depicting differentially expressed genes dependent on TET2 mutational status in the cohort of CEBPA-mutant patients in the Beat AML dataset (TET2^WT 11 and TET2^MUT 5 patients). Differential analysis was performed with DESeq2 (P < 0.05). Source data are provided as a Source Data file.

Fig. 2. TET2 deficiency accelerates Cebpa-mutant AML.

a Schematic representation of generation of Tet2-knockout clones with CRISPR/Cas9. The illustration was created with BioRender.com. b Proliferative outgrowth of Cebpa^p30/p30 cells with Tet2 indels. c Volcano plot depicting differentially expressed genes dependent on the Tet2 mutational status in Cebpa^p30/p30 cells (Tet2^WT 7 and Tet2^MUT 5 clones). Differential analysis was performed with DESeq2 (P < 0.05). d Experimental setup for evaluating the effect of Tet2-deficiency (Tet2^Δ/Δ) in Cebpa^DM AML initiation in vivo. The illustration was created with BioRender.com. e Myeloid (Mac1⁺) contribution of donor-derived blood and bone marrow (BM) cells evaluated after BM transplantation and Cre-LoxP recombination. (Blood samples: Week 12; 6 mice per group. Week 24; Cebpa^Δ/p30Tet2^+/+ and Cebpa^+/ΔTet2^Δ/Δ 6 mice per group and Cebpa^+/ΔTet2^Δ/Δ 3 mice. Week 36; 3 mice per group. BM samples: 3 mice per group.) Data are presented as mean±SEM and analyzed by one-way-ANOVA followed by Dunnett’s multiple comparisons correction. f Survival of lethally irradiated recipient mice after BM transplantation and Cre-LoxP recombination (Cebpa^Δ/p30Tet2^+/+ 12 mice, Cebpa^Δ/p30Tet2^Δ/Δ 14 mice, and Cebpa^+/-Tet2^Δ/Δ 14 mice). The data were analyzed by Mantel-Cox Log-rank test. g Volcano plot depicting differentially expressed genes dependent on Tet2 deficiency status in Cebpa^Δ/p30 leukemic blasts (samples from 3 mice per group). Differential analysis was performed with DESeq2 (P < 0.05). h Frequency of proliferating (Ki67⁺) cells in BM of moribund recipient mice (specimens from 3 mice per group). Data are presented as mean±SEM and analyzed by a two-tailed unpaired t-test. Source data are provided as a Source Data file.

Fig. 3. Loss of TET2 leads to reduced Gata2 levels in Cebpa-mutant AML.

a Conserved targets of the CEBPA-TET2 axis visualized in a Venn-diagram of downregulated genes in CEBPA-TET2 co-mutated AML overlaid with corresponding data from Tet2-deficient in vivo and in vitro models of Cebpa^DM AML (P = 0.0264 vs. number of overlapping genes expected by random distribution assessed by Wilson/Brown binominal test). b Heatmap of differentially accessible regions assessed by assay for transposase-accessible chromatin sequencing (ATAC-seq; FDR < 0.05), and c genomic distribution of downregulated peaks (FDR < 0.05, Log2FC < 0) upon Tet2 mutation (4 clones per group). Differential analysis was performed with DiffBind and region enrichment analysis with GREAT. d Representative genome wide DNA-methylation status in leukemic blasts from the in vivo model assessed by whole genome bisulfite sequencing (WGBS) showing frequency of methyl-cytosine (mC) across the transcription start site (TSS) ±1000 base pairs, gene body scaled to 4000 base pairs, and transcription termination site (TES) ±1000 base pairs. Methylation analysis was performed with Bismark and visualized using deepTools. e Median and interquartile range of percent mC at promoters of down- (n = 172), not expressed (n = 6539), not differentially expressed (not DE; n = 14759) and up-regulated (n = 57) genes (averaged data generated from 2 Cebpa^Δ/p30Tet2^+/+ and 3 Cebpa^Δ/p30Tet2^Δ/Δ mice). Whiskers indicates max–min and data were analyzed by two-tailed unpaired t-test. f Promoter DNA methylation of conserved target genes in leukemic blast (samples from 2 Cebpa^Δ/p30Tet2^+/+ and 3 Cebpa^Δ/p30Tet2^Δ/Δ mice). Data are presented as mean±SEM. The data were log-transformed and analyzed by two-tailed unpaired t-test. g Gata2 variant mRNA expression in Cebpa^Δ/p30Tet2^Δ/Δ and Cebpa^Δ/p30Tet2^+/+ leukemic blasts (samples from 3 mice per group). Data are presented as mean ± SEM. The data were log-transformed and analyzed by a two-tailed unpaired t-test. h Schematic genomic view of the Gata2 locus, including representative examples of assay for transposase-accessible chromatin using sequencing (ATAC-seq) in Cebpa^p30/p30 cells. Source data are provided as a Source Data file.

Fig. 4. Moderate Gata2 reduction increases competitiveness of Cebpa-mutant AML.

a Experimental setup for evaluating the effect of Gata2 knockdown, via short hairpin RNA (shRNA) mediated silencing, on Cebpa^p30/p30 leukemic cells in a competitive in vivo assay. The illustration was created with BioRender.com. b Gata2 mRNA in Cebpa^p30/p30 leukemic cells prior to transplantation. c Representative flow cytometry profiles of input and output of shControl (no knockdown), shGata2A (low knockdown), and shGata2D (high knockdown). d Competitive advantage of targeting shRNA (GFP⁺) vs. non-targeting shRNA (YFP⁺) cells in vivo assessed as by flow cytometry (Control 4, shGata2A 4, shGata2B 4, shGata2C 4, and shGata2D 3 mice). Data are presented as mean±SEM. Data were log-transformed and analyzed by one-way-ANOVA followed by Dunnett’s multiple comparisons correction. e Experimental setup for Gata2 CRISPR/Cas9 mutagenesis in Cebpa^p30/p30 cells, and outgrowth of heterozygous mutated clones. Percentages of Gata2 mutated clones are indicated. f Growth curve of Cebpa^p30/p30 clones with Gata2 mutation (Cebpa^p30/p30Gata2^+/MUT, n = 10) or wild type Gata2 (Cebpa^p30/p30Gata2^+/+, n = 3). Data are presented as mean±SEM and analyzed by two-tailed unpaired t-test. Red lines mark individual mutated clones. g Presence or absence of GATA2 mutations (GATA2^MUT) in CEBPA double mutated (CEBPA^DM) AML cases with or without TET2 mutations (TET2^MUT) in aggregated data from published cohorts^–,,, (detailed in Supplemental Table 2a). Data were analyzed by Wilson/Brown binominal test. Source data are provided as a Source Data file.

Fig. 5. Increased CEBPA p30 binding to the Gata2 distal hematopoietic enhancer drives expression of Gata2 via TET2.

a Gata2 mRNA expression in mouse Cebpa^p30/p30 leukemic granulocyte/monocyte progenitors (GMPs) vs normal GMPs (samples from 4 Cebpa^+/+ and 2 Cebpa^p30/p30 mice) and, b CEBPA binding to the Gata2 distal hematopoietic enhancer (G2DHE; −77 kb) region (samples from 2 mice per group), data from Jakobsen et al.. Data are presented as mean ± SEM. Differential analysis was performed with DESeq2 (P < 0.05). c Schematic genomic view of the Gata2 distal hematopoietic enhancer (G2DHE), including normalized chromatin immunoprecipitation sequencing (ChIP-seq) signal of CEBPA (data from Heyes et al.), TET2 and H3K27Ac (data from Heyes et al.), as well as assay for transposase-accessible chromatin using sequencing (ATAC-seq) in Cebpa^p30/p30 cells without (light blue) and with (green) mutation in Tet2. d Gata2 mRNA levels in response to targeting of the G2DHE by CRISPR-Cas9 in Cebpa^p30/p30 cells in vitro using indicated sgRNAs and e the averaged change in Gata2 mRNA levels of the 12 deletions (averaged data from 2 separate experiments). Data are presented as median ± range and analyzed by two-tailed Wilcoxon signed-rank test. f Experimental setup for evaluating the effects of Cebpa knockout on Gata2 V2 mRNA expression and DNA methylation of the CpG island at the promoter of Gata2 V2 in MLL-fusion driven AML (iMLL-AF9). The illustration was created with BioRender.com. g Gata2 V2 mRNA expression (leukemic cell lines generated from 2 separate mice were assayed on 2 separate days in 2–3 technical replicates each). Data are presented as mean ± SEM and the individual cell lines are indicated by circles or squares. Data were log-transformed and analyzed by two-tailed unpaired t-test. h DNA methylation of the Gata2 V2 promoter CpG-island (2 separate leukemic cell lines). Data are presented as median±range and the individual cell lines are indicated by circles or squares. i Experimental setup for evaluating the effects of Cebpa knockdown on TET2 binding to the G2DHE in Cebpa^p30/p30 cells with inducible expression of shRNA targeting Cebpa and control (Renilla), respectively. The illustration was created with BioRender.com. j TET2 binding to the G2DHE assessed by ChIP-qPCR (3 replicates per condition). Data are presented as mean±SEM and analyzed by two-tailed unpaired t-test. k Frequency of GATA2 and/or TET2 mutations (GATA2^MUT and TET2^MUT, respectively) in CEBPA high expressing (CEBPA^HIGH; 45 cases) vs. CEBPA low expressing (CEBPA^LOW; 61 cases) AML cases, data from Beat AML cohort. The distributions of GATA2^WTTET2^WT vs. GATA2^MUT and/or TET2^MUT cases were analyzed by Wilson/Brown binominal test. Source data are provided as a Source Data file.

Fig. 6. Demethylating treatment restores Gata2 expression and prolongs survival in TET2-deficient Cebpa-mutant AML.

a Experimental setup for evaluating the effect of short-term 5-azacytidine (5-AZA) treatment in vivo. Recipient mice were sub-lethally irradiated and transplanted with leukemic BM from moribund secondary recipient mice. Three individual Cebpa^Δ/p30Tet2^Δ/Δ clones (A–C) and two Cebpa^Δ/p30Tet2^+/+ clones (A–B) were used, respectively. The illustration was created with BioRender.com. b Expansion of myeloid (Mac1⁺) donor-derived cells in bone marrow (BM) assessed by flow cytometry, and c Gata2 mRNA expression in sorted leukemic blasts by qPCR assessed 24 hours after the last of three injections of 5-AZA or vehicle (samples from 3 mice per clone and 6 and 9 mice per group, for Cebpa^Δ/p30Tet2^{+/+ and} Cebpa^Δ/p30Tet2^Δ/Δ, respectively). Dot plots showing individual mice for separate clones and bar graphs shows mean ± SEM for each group. Data were analyzed by Kruskal–Wallis test followed by Dunn’s correction for multiple comparisons. d Experimental setup for evaluating the effect of 5-AZA treatment on AML progression in vivo. The illustration was created with BioRender.com. e Survival of sub-lethally irradiated tertiary recipient mice after transplantation of leukemic BM from moribund secondary recipient mice (clone A from both genotypes) in response to intermittent 5-AZA treatment (5-AZA treated groups 8 mice and vehicle-treated groups 4 mice). The data were analyzed by Mantel–Cox Log-rank test. Source data are provided as a Source Data file.

Fig. 7. TET2 lesions enhance the aggressiveness of CEBPA-mutant AML by rebalancing GATA2 expression.

a Model of Gata2 differential expression as a consequence of (I) elevated CEBPA p30 due to the hypermorphic effect of the CEBPA^NT, (II) TET2 deficiency and, (III) CEBPA deficiency. b Schematic illustration of two strategies for CEBPA^DM AML to rebalance GATA2 levels by (I) loss-of-function mutations in TET2 and (II) loss-of-function mutations in one GATA2 allele. The illustrations were created with BioRender.com.