GATA Factor-Dependent Positive-Feedback Circuit in Acute Myeloid Leukemia Cells

Abstract

The master regulatory transcription factor GATA-2 triggers hematopoietic stem and progenitor cell generation. GATA2 haploinsufficiency is implicated in myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML), and GATA2 overexpression portends a poor prognosis for AML. However, the constituents of the GATA-2-dependent genetic network mediating pathogenesis are unknown. We described a p38-dependent mechanism that phosphorylates GATA-2 and increases GATA-2 target gene activation. We demonstrate that this mechanism establishes a growth-promoting chemokine/cytokine circuit in AML cells. p38/ERK-dependent GATA-2 phosphorylation facilitated positive autoregulation of GATA2 transcription and expression of target genes, including IL1B and CXCL2. IL-1β and CXCL2 enhanced GATA-2 phosphorylation, which increased GATA-2-mediated transcriptional activation. p38/ERK-GATA-2 stimulated AML cell proliferation via CXCL2 induction. As GATA2 mRNA correlated with IL1B and CXCL2 mRNAs in AML-M5 and high expression of these genes predicted poor prognosis of cytogenetically normal AML, we propose that the circuit is functionally important in specific AML contexts.

Figure 1. Ras-p38/ERK- and GATA-2 DEF Motif-Mediated GATA-2 Phosphorylation in AML Cells

(A) Total Kasumi-1 cell protein was incubated with or without λ-phosphatase and analyzed by western blotting with anti-GATA-2 antibody. (B) Left: western blot analysis of endogenous GATA-2 in Kasumi-1 cells with or without expression of H-Ras(G12V). Right: densitometric analysis of relative protein levels is shown. The ratio of intensities of the upper to lower bands from control Kasumi-1 cells was designated as 1 (n = 3, mean ± SE, *p < 0.05). (C) Left: western blot analysis of MAP kinases and wild-type and mutant proteins transiently expressed in Kasumi-1 cells with or without H-Ras(G12V). Expressed GATA-2 was detected with anti-HA antibody. Right: densitometric analysis of relative protein levels is shown. The ratio of intensities of the upper to lower bands from HA-GATA-2-expressing Kasumi-1 cells was designated as 1 (n = 3, mean ± SE, **p < 0.01). (D) Top: western blot analysis of substrates of MAP kinases, GATA-2, and Ras proteins transiently expressed in Kasumi-1 cells with or without H-Ras(G12V) and MAP kinase inhibitors. Expressed GATA-2 was detected with anti-HA antibody. ERK, HSP27, and c-Jun phosphorylation was assessed to test inhibitor specificities. Bottom: densitometric analysis of relative protein levels is shown. The ratio of intensities of the upper to lower bands from Ras(G12V)-expressing Kasumi-1 cells was designated as 1 (n = 3, mean ± SE, *p < 0.05 and **p < 0.01). (E) Left: western blot analysis of MAP kinases, substrates of MAP kinases, and GATA-2 proteins transiently expressed in Kasumi-1 cells with or without constitutively active MEK1 or p38. Expressed GATA-2 protein was detected with anti-HA antibody. Right: densitometric analysis of relative protein levels is shown. The ratio of intensities of the upper to lower bands from GATA-2-expressing Kasumi-1 cells was designated as 1 (n = 3, mean ± SE, **p < 0.01). (F) DEF motif in GATA-2 is shown. (G) Left: western blot analysis of MAP kinases and wild-type and mutant proteins transiently expressed in Kasumi-1 cells with or without H-Ras(G12V). Expressed GATA-2 was detected with anti-HA antibody. Right: densitometric analysis of relative protein levels is shown. The ratio of intensities of upper to lower bands from HA-GATA-2-expressing Kasumi-1 cells was designated as 1 (n = 3, mean ± SE, **p < 0.01). (H) Total protein from Kasumi-1 cells expressing GATA-2 proteins was incubated with or without λ-phosphatase. Proteins were analyzed by western blotting with anti-HA antibody. (I) Left: western blot analysis of HA-GATA-2 and HA-GATA-2ΔDEF transiently expressed in MAE cells with or without Ras(G12V). Right: qRT-PCR analysis of Hdc mRNA levels in MAE cells transiently expressing HA-GATA-2 and HA-GATA-2ΔDEF with or without Ras(G12V) is shown (n = 3, mean ± SE, *p < 0.05). See also Figure S1.

Figure 2. Ras-p38/ERK Enhance GATA-2-Mediated Target Gene Transcription in AML Cells

(A) Top: western blot analysis of endogenous GATA-2 in Kasumi-1 cells stably infected with sh-luc virus or sh-GATA2 virus. Bottom: densitometric analysis of relative protein levels is shown. The protein expression in Kasumi-1 cells infected with sh-luc virus was designated as 1. (B) Real-time RT-PCR analysis of GATA2 mRNA and transcripts of GATA-2 target genes in Kasumi-1 cells infected with sh-luc virus or sh-GATA2 virus is shown (n = 5; mean ± SE; *p < 0.05, **p < 0.01, and ***p < 0.001). (C) ChIP-seq analysis of endogenous GATA-2 occupancy at GATA2, IL1B, and CXCL2 loci in Kasumi-3 cells, TF-1 cells (Mazumdar et al., 2015), and human CD34-positive hematopoietic cells (Beck et al., 2013) is shown. (D) Real-time RT-PCR analysis of transcripts of GATA-2 target genes in Kasumi-1 cells treated with 40 μM SB203580 (n = 4, mean ± SE). Samples were harvested at the designated times. (E) Western blot analysis of endogenous GATA-2 in Kasumi-1 cells treated with 40 μM SB203580 is shown (n = 5, mean ± SE, *p < 0.05). (F) Real-time RT-PCR analysis of GATA-2 target genes in Kasumi-1 cells treated with 20 μM U0126 (n = 4, mean ± SE). Samples were harvested at the designated times. (G) Western blot analysis of endogenous GATA-2 in Kasumi-1 cells treated with 20 μM U0126 is shown (n = 5, mean ± SE, *p < 0.05). (H) Left: western blot analysis of endogenous GATA-2 and MAP kinase proteins in Kasumi-1 cells expressing Ras(G12V). Right: real-time RT-PCR analysis of IL1B and CXCL2 expression in Kasumi-1 cells expressing Ras(G12V) is shown (n = 3, mean ± SE, *p < 0.05 and **p < 0.01). (I) Real-time RT-PCR analysis of IL1B and CXCL2 expression in Kasumi-1/sh-luc cells or Kasumi-1/sh-G2 cells expressing Ras(G12V) (n = 6, mean ± SE, *p < 0.05 and **p < 0.01). See also Figure S2.

Figure 3. p38/ERK Signaling Promotes GATA-2 Chromatin Occupancy and Chromatin Remodeling

(A) Gata2 locus map. Numbers represent distance to mouse 1S transcription start site. 1G is another GATA2 transcription start site. Quantitative ChIP analysis of GATA-2 occupancy in Kasumi-1 cells treated with 40 μM SB203580 or 20 μM U0126 is shown (n = 4, mean ± SE, *p < 0.05 and **p < 0.01). The western blot (anti-GATA-2 antibody) (inset) illustrates GATA-2 expression in representative samples used for ChIP. (B) Quantitative ChIP analysis of GATA-2 occupancy at GATA-2 target genes in Kasumi-1 cells treated with 40 μM SB203580 or 20 μM U0126 is shown (n = 4, mean ± SE, *p < 0.05). (C) Quantitative FAIRE analysis of chromatin accessibility in Kasumi-1 cells treated with 40 μM SB203580 or 20 μM U0126 is shown (n = 4, mean ± SE, *p < 0.05 and **p < 0.01). (D) Kinetics of GATA-2 expression and phosphorylation are shown (n = 3, mean ± SE, *p < 0.05). (E) GATA-2 occupancy at the −77 kb site during 40 μM SB203580 treatment is shown (n = 3, mean ± SE, *p < 0.05). (F) Kinetics of GATA-2 protein expression and phosphorylation are shown (n = 3, mean ± SE, *p < 0.05). (G) GATA-2 occupancy at the −77 kb site during 20 μM U0126 treatment is shown (n = 3, mean ± SE, *p < 0.05). (H) Left: rescue assay. Right: qRT-PCR analysis of IL1B and CXCL2 in Kasumi-1 cells expressing GATA-2 or control vector with or without 10 μM SB203580 is shown (n = 4; mean ± SE; *p < 0.05; PI, preimmune). See also Figure S3.

Figure 4. p38/ERK-GATA-2 Axis Establishes a Chemokine/Cytokine-Dependent Positive-Feedback Circuit

(A) Left: western blot analysis of endogenous GATA-2 in Kasumi-1 cells treated with 10 ng/ml recombinant human IL-1β. The cells were serum-starved overnight before IL-1β treatment. Right: densitometric analysis is shown. The ratio of intensities of the upper to lower bands from at 0 min was designated as 1 (n = 3, mean ± SE, *p < 0.05). (B) Western blot analysis of endogenous GATA-2 in Kasumi-1 cells treated with 100 ng/ml recombinant human CXCL2. Cells were serum-starved overnight before CXCL2 treatment. Right: densitometric analysis is shown. The ratio of intensities of the upper to lower bands at 0 min was designated as 1 (n = 3, mean ± SE, *p < 0.05). (C) Quantitative ChIP analysis of GATA-2 occupancy at GATA-2 target genes in Kasumi-1 cells treated with 10 ng/ml IL-1β or 100 ng/ml CXCL2 (n = 4, mean ± SE). The cells were serum-starved overnight before treatment with IL-1β or CXCL2 (*p < 0.05). (D) Left: western blot analysis of endogenous GATA-2 in primary AML cells treated with 10 ng/ml recombinant human IL-1β or 100 ng/ml recombinant human CXCL2 for 5 min. The cells were serum-starved for 90 min before treatment. Right: densitometric analysis is shown. The ratio of intensities of the upper to lower bands from the control sample was designated as 1 (n = 4, mean ± SE, *p < 0.05). (E) Real-time RT-PCR analysis of GATA2 mRNA in primary AML cells treated with 10 ng/ml IL-1β or 100 ng/ml CXCL2 (n = 4, mean ± SE). Primary AML cells were serum-starved for 90 min and treated with IL-1β or CXCL2 for 30 min (*p < 0.05). See also Figure S4.

Figure 5. p38/ERK-GATA2 Axis Stimulates Kasumi-1 Cell Proliferation

(A) Comparison of growth rates of Kasumi-1 cells stably infected with sh-luc virus or sh-GATA2 virus. The Kasumi-1/sh-luc cells and Kasumi-1/sh-G2 cells (1 × 10⁵) were plated, and cells were counted every second day. Cells were passaged at a density of 1 × 10⁵ cells at day 4 (n = 9, mean ± SE, *p < 0.05 and **p < 0.01). (B) Top: proliferation analysis. Bottom: representative plots from proliferation analysis of Kasumi-1/sh-luc cells and Kasumi-1/sh-G2 cells using Cell-Trace Violet dye are shown. A greater number of generations indicates the cells underwent more rounds of cell division. (C) Left: the average percentage of total cells in each daughter generation. Each bar represents three independent experiments from one clonal line. The western blot inset illustrates the efficacy of the GATA-2 knockdown (*p < 0.05, **p < 0.01, and ***p < 0.001). Right: proliferation index of Ka-sumi-1/sh-luc cells and Kasumi-1/sh-G2 cells is shown. Each bar represents three independent experiments from one clonal line (***p < 0.001). (D) Comparison of the growth rates of Kasumi-1 cells stably infected with sh-luc virus or sh-G2 virus with 100 ng/ml recombinant human CXCL2. The Kasumi-1/sh-luc cells and Kasumi-1/sh-G2 cells (1 × 10⁵) were plated, and cells were counted every second day. Cells were passaged at a density of 1 × 10⁵ cells at day 4 (n = 9, mean ± SE, *p < 0.05). (E) Representative plots from proliferation analysis of Kasumi-1/sh-luc cells and Kasumi-1/sh-G2 cells treated with or without 100 ng/ml recombinant human CXCL2 using CellTrace Violet dye. A greater number of generations indicates the cells underwent more rounds of cell division. (F) Proliferation index of Kasumi-1/sh-luc cells and Kasumi-1/sh-G2 cells. Cells were cultured with or without 100 ng/ml CXCL2 for 6 days (n = 3, mean ± SE, **p < 0.01 and ***p < 0.001). See also Figure S5.

Figure 6. GATA2, IL1B, and CXCL2 Expression in AML Patients

(A) Heatmap depicts 19,798 genes, based on ranking of the correlation with GATA2 expression. (B) Heatmap depicts GATA2, CXCL2, and IL1B mRNA correlations. (C) Scatterplot depicts correlations among GATA2, CXCL2, and IL1B mRNA expression in M5-AML patients. (D) Real-time RT-PCR analysis of CXCL2 and IL1B mRNA in Kasumi-1 cells treated with 10 ng/ml IL-1β or 100 ng/ml CXCL2 (n = 3, mean ± SE). Kasumi-1 cells were serum-starved overnight and treated with IL-1β or CXCL2 for 30 min (*p < 0.05). (E) Kaplan-Meier plots compare overall survival of patients with high versus low GATA2, CXCL2, or IL1B mRNA. (F) Kaplan-Meier plots depict overall survival of patients with high versus low GATA2/IL1B mRNA, GATA2/CXCL2 mRNA, and IL1B/CXCL2 mRNA. This analysis was based on the cohort of 163 patients with CN-AML from the GEO: GSE12417 dataset (Metzeler et al., 2008). See also Figures S6 and S7.

Figure 7. p38/ERK-GATA-2 Axis Function in AML Cells

Ras-p38/ERK signaling increases GATA-2 phosphorylation. GATA-2 phos-phorylation facilitates GATA-2 chromatin occupancy at GATA-2 target genes. GATA-2 stimulates GATA2 transcription through positive autoregulation. Loci are differentially sensitive to the signal-dependent GATA-2 mechanism. GATA-2 upregulates IL1B and CXCL2 expression. These factors activate p38 and ERK signaling, enhancing GATA-2 activity through a positive-feedback circuit. GATA-2 increases CXCL2 transcription directly or indirectly (through IL-1β), which constitutes a type I coherent feedforward loop (FFL). This GATA-2-chemokine/cytokine circuit is predicted to be an important determinant of AML cell proliferation.