IRF2BP2 counteracts the ATF7/JDP2 AP-1 heterodimer to prevent inflammatory overactivation in acute myeloid leukemia (AML) cells

Abstract

Acute myeloid leukemia (AML) is a hematological malignancy characterized by abnormal proliferation and accumulation of immature myeloid cells in the bone marrow. Inflammation plays a crucial role in AML progression, but excessive activation of cell-intrinsic inflammatory pathways can also trigger cell death. IRF2BP2 is a chromatin regulator implicated in AML pathogenesis, although its precise role in this disease is not fully understood. In this study, we demonstrate that IRF2BP2 interacts with the AP-1 heterodimer ATF7/JDP2, which is involved in activating inflammatory pathways in AML cells. We show that IRF2BP2 is recruited by the ATF7/JDP2 dimer to chromatin and counteracts its gene-activating function. Loss of IRF2BP2 leads to overactivation of inflammatory pathways, resulting in strongly reduced proliferation. Our research indicates that a precise equilibrium between activating and repressive transcriptional mechanisms creates a pro-oncogenic inflammatory environment in AML cells. The ATF7/JDP2-IRF2BP2 regulatory axis is likely a key regulator of this process and may, therefore, represent a promising therapeutic vulnerability for AML. Thus, our study provides new insights into the molecular mechanisms underlying AML pathogenesis and identifies a potential therapeutic target for AML treatment.

Graphical Abstract

Figure 1.

IRF2BP2 is required for AML cell proliferation. (A)Distribution of CRISPR scores upon IRF2BP2 deletion in 357 human cancer cell lines (47), with AML cell lines marked in red. (B) Expression of IRF2BP2 in AML versus non-AML samples, using data from GePIA (38). (C) Kaplan–Meier survival curve (disease-free survival) of AML patients dependent on IRF2BP2 expression. Data were derived from TCGA (40) and were visualized using GePIA (38) using overall survival and median as cut-off. (D) Kaplan–Meier survival curve (overall survival) of AML patients dependent on IRF2BP2 expression. The data are based on multiple microarray datasets, visualized by the Kaplan–Meier–Plotter (42) using overall survival and ‘auto select best cutoff’. (E)Negative-selection competition assay showing the percentage of GFP+ control- or sgIRF2BP2-transduced cells over time (n = 2–3 each group). Data represent the mean ± s.d. of at least two biological replicates. P-values were evaluated via ANOVA. (F) Expression of IRF2BP2 and IRF2 after IRF2BP2 shRNA-mediated knockdown in THP1 cells. Data represent the mean ± s.d. of three biological replicates. P-values were evaluated via ANOVA. (G)Growth of THP1 and MV4-11 cells after knockdown of IRF2BP2. Data represent the mean ± s.d. of at least three biological replicates. Significance was evaluated via ANOVA. (H)Most enriched pahways found by GSEA when comparing AML samples from TCGA (40), with low versus high IRF2BP2 expression. (I)GSEA plots of the interferon-gamma response pathway, comparing low versus high IRF2BP2 expression in AML samples from TCGA. (J) GSEA plots of Myc target genes, comparing low versus high IRF2BP2 expression in AML samples from TCGA. n.s. = no significant; * P < 0.05; ** P < 0.01

Figure 2.

The ATF7/JDP2 dimer interacts with IRF2BP2 but has an opposing role in regulating inflammatory pathways. (A) Semi-endogenous co-immunoprecipitation showing the interaction of IRF2BP2 with itself and IRF2BP1. (B) Semi-quantitative mass-spectrometry after co-immunoprecipitation of Flag-IRF2BP2 in THP1 cells. See also Supplementary Table S7. (C) Immunofluorescence showing the colocalization of ectopically expressed IRF2BP2 with endogenous IRF2BP1, JDP2 and ATF7 in nuclear foci (white arrow) in HEK293 cells. (D) Confirmation of interaction of IRF2BP2 with ATF7 and JDP2 using co-immunoprecipitation experiments in HEK293 cells. (E) Domain structure of IRF2BP2, IRF2BP1, IRF2BPL, ATF7 and JDP2. NLS = Nuclear localization signal, ZF= zinc finger domain, RING = RING finger domain. (F) AlphaFold2-predicted dimer von IRF2BP1 middle domain. See also Supplementary Figure S2A–D. The prediction is available in the Supplementary Data File S1. (G) Co-immunoprecipitation of IRF2BP1 middle domain with itself. See also Supplementary Figure S2E-H. (H) Bubble plot showing the gene sets (hallmarks) correlating with ATF7 and JDP2 expression (low versus high) in TCGA AML samples. Inflammatory pathways negatively correlate with low ATF7 and JDP2 expression (lower left quadrant). NES=normalized enrichment score. (I) Bubble plot showing the correlation of gene sets that correlate with JDP2 and IRF2BP2 expression (low versus high) in TCGA samples. Inflammatory pathways anti-correlate (lower right quadrant). (J) As in (I), but the comparison between ATF7 and IRF2BP2 expression is shown. Inflammatory pathways anti-correlate (lower right quadrant). GSEA results from (H) to (J) are presented in Supplementary Table S6.

Figure 3.

IRF2BP2 associates with ATF7 and JDP2 via multiple interaction sites. (A) Western blot of a co-immunoprecipitation experiment of IRF2BP2 deletion mutants versus IRF2BP1. (B) Western blot of a co-immunoprecipitation experiment of IRF2BP1 deletion mutants versus IRF2BP2. (C) Western blot of a co-immunoprecipitation experiment of IRF2BP2 deletion mutants versus ATF7. (D) Western blot of a co-immunoprecipitation experiment of IRF2BP2 deletion mutants versus JDP2. (E) Western blot of a co-immunoprecipitation experiment of IRF2BP1 deletion mutants versus ATF7. (F) Western blot of a co-immunoprecipitation experiment of IRF2BP1 deletion mutants versus JDP2- (G) AlphaFold predicted dimerization of ATF7 and JDP2 (34,54) modeled onto DNA using PDB: 1FOS (55). All co-immunoprecipitation experiments were performed upon ectopic expression in HEK293 cells. Representative results of at least two biological replicates are shown. Red proteins indicate those that do not interact in the respective experiment.

Figure 4.

ATF7 chromatin binding requires JDP2. (A) Western blot of THP1 cells with CRISPR/Cas9-mediated KOs of ATF7 and JDP2. (B) ChIP-qPCR of ATF7 in control, ATF7 KO, and JDP2 KO THP1 cells. Data represent the mean ± s.d. of at least three biological replicates. The significance was evaluated via a two-tailed unpaired Student's t-test. (C) Overlap of ATF7 peaks from THP1 cells (this study) with ATF7 peaks from K562 cells (44), and JDP2 peaks from Loucy cells (45). (D) Genomic distribution of ATF7 and JDP2 compared to FOS (43). (E) Motifs at ATF7 binding sites, identified via HOMER (26). (F) Heatmap of ATF7 ChIP-Seq results in control, ATF7 KO, and JDP2 KO THP1 cells. The two ChIP-Seq replicates were merged. (G) Violin plot showing the distribution of ATF7 ChIP-Seq signals at ATF7 peaks. Significance was evaluated via a two-sided Kolmogorov–Smirnov test. n.s = no significant; * P < 0.05; ** P < 0.01; *** P < 0.001

Figure 5.

IRF2BP2 chromatin binding is reduced in the absence of ATF7 and JDP2. (A) Venn diagram showing overlap of IRF2BP2 ChIP-Seq peaks with IRF2BP2 ChIP-Seq data from Ellegast et al. (7). See also Supplementary Figure S5A and B. (B) Genomic distribution of IRF2BP2. (C) Venn diagram showing the overlap of IRF2BP2 peaks with ATF7 ChIP-Seq peaks. (D) Venn diagram showing the overlap of IRF2BP2-bound and ATF7-bound promoters. Significance was evaluated via a hypergeometric probability test. (E) Motifs enriched at IRF2BP2 binding sites, identified via HOMER. (F) Enriched gene ontologies at IRF2BP2/ATF7 overlapping peaks, and at ATF7 and IRF2BP2 peaks. Analysis performed via GREAT (27). (G) ChIP-qPCR experiments of IRF2BP1 and IRF2BP2 upon ATF7 and JDP2 KO. Data represent the mean ± s.d. of at least two biological replicates. The significance was evaluated via a two-tailed unpaired Student's t-test. (H) Heatmaps showing IRF2BP2 ChIP-Seq signals at significant IRF2BP2 peaks upon ATF7 and JDP2 KO. (I) Violin plot showing IRF2BP2 levels at IRF2BP2 peaks upon ATF7 and JDP2 KO. Significance was evaluated via a two-sided Kolmogorov–Smirnov test. (J) Genome browser image showing the reduced levels of IRF2BP2 at the JUN and IRF2 genes upon ATF7 and JDP2 KO. n.s = no significant; * P < 0.05; ** P < 0.01; *** P < 0.001

Figure 6.

IRF2BP2 depletion activates inflammatory pathways in THP1 cells. (A) Volcano plots showing significantly (fold change > 0.75, p-value < 0.01) differentially expressed genes upon KD of IRF2BP2 or KO of ATF7 and JDP2. (B) GSEA of the top 250 IRF2BP2 target genes and upon IRF2BP2 depletion. (C) GSEA of the top 250 ATF7 target genes and upon ATF7 deletion. (D) Principal component analysis of the RNA-Seq data. The data for IRF2BP2 were obtained from two independent shRNAs (3 replicates each). Two independent KO clones for ATF7 and JDP2 were analyzed (2 replicates each). (E) Correlation of gene expression changes upon ATF7 and JDP2 KO. Significance was evaluated via ANOVA. (F) GSEA plot of the top 2 hallmarks affected upon IRF2BP2 depletion in THP1 cells. Compare to Figure 1  I. (G) Bubble plot comparing the results from GSEA obtained from our RNA-Seq data (IRF2BP2 KD versus Control) and analyzing TCGA samples (IRF2BP2^low versus IRF2BP2^high). (H) Bubble plot comparing the results from GSEA obtained from our RNA-Seq data (IRF2BP2 KD versus control in THP1 cells) and RNA-Seq data from Ellegast et al. (IRF2BP2 dTAG versus DMSO in MV4-11 cells for 24 h)(7). (I) Overlap of significantly upregulated genes in our KD experiments and experiments from Ellegast et al. Overlapping genes are linked to innate immune response and inflammation. Significance of the overlap was evaluated via a hypergeometric probability test. (J) Bubble plot comparing the GSEA results from ATF7 KO and JDP2 KO. (K) Bubble plot comparing the GSEA results from IRF2BP2 KD and JDP2 KO. (L) Bubble plot comparing the GSEA results from IRF2BP2 KD and ATF7 KO. NES = Normalized Enrichment Score, KO = Knockout, KD = Knockdown. GSEA results from (G), (H), (J), (K) and (L) are presented in Supplementary Figure S6.

Figure 7.

IRF2BP2 depletion influences differentiation from mouse bone marrow stem and progenitor cells. (A) Western blot of components of the IRF2BP2 complex in cells isolated from mouse bone marrow. NUC = nucleoplasm, CHR = chromatin. (B) Representative bright field photography of colonies (here CFU-GM) obtained upon colony formation assays of myeloid progenitor cells with and without IRF2BP2 KD in methylcellulose medium. Additional examples are shown in Supplementary Figure S9B. Scale = 500 μm. (C) Quantification of the area of colonies upon differentiation in methylcellulose. For each condition, 15 colonies were measured from two biological replicates. Significance was evaluated via a two-tailed unpaired Student's t-test. (D) Representative bright field microscopy photography of dendritic cell-shaped cells (arrows), seen in the IRF2BP2 KD cells compared to the control cells. Scale = 50 μm. (E) RT-qPCR analysis of collected cells after colony formation assays. Data represents the mean of three biological replicates. Significance was evaluated via a two-tailed paired Student's t-test. (F) FACS quantification of surface makers after 10 days of differentiation of Lin-negative bone marrow cells in liquid culture. Data represent the mean ± s.d. of at least three biological replicates. Significance was evaluated via a two-tailed unpaired Student's t-test. Example histograms are shown in Supplementary Figure S9B. (G) GSEA of TCGA and our own RNA-Seq data regarding pathways involved in myeloid leukocyte migration and stem cell maintenance. n.s. = no significant; * P < 0.05; ** P < 0.01; *** P < 0.001

Figure 8.

Model of the role of the ATF7/JDP2-IRF2BP2 axis in AML. (A) ATF7/JDP2 activates inflammatory pathways, while IRF2BP2 is involved in counteracting ATF7/JDP2. This balanced regulation establishes an optimal level of inflammation required for high proliferation and rapid cancer progression. (B) In the absence of IRF2BP2, inflammatory genes become strongly upregulated, leading to inflammatory overactivation and cell death. (C) In the absence of ATF7/JDP2, their activating function is abolished. Due to alternative recruitment mechanisms, the repressive role of IRF2BP2 is partially retained, leading to a reduction in inflammatory pathways.