A novel FOXM1-BCL2A1 axis determines unfavorable response to venetoclax in AML

Abstract

Forkhead box M1 (FOXM1), a Forkhead family transcription factor, is often overexpressed in a variety of human cancers, including acute myeloid leukemia (AML), and is strongly associated with therapy resistance and unfavorable outcomes. In AML with NPM1 mutations, NPM1-FOXM1 complex sequesters FOXM1 in the cytoplasm and confers favorable treatment outcomes for AML patients because of FOXM1 inactivation. Inhibition of FOXM1 in AML cell lines and animal models of AML sensitizes AML cells to the BCL2 inhibitor, venetoclax. In a recent study, the upregulation of the BCL2-family protein, BCL2A1, conferred resistance to venetoclax and multiple venetoclax combinations. In this study, we investigated the FOXM1-BCL2A1 axis and determined that FOXM1 specifically inhibits venetoclax-induced apoptosis in AML via upregulation of BCL2A1. The knockdown of BCL2A1 in AML in the presence of high levels of FOXM1 led to sensitization of AML cells to venetoclax, suggesting that BCL2A1 is a major target of FOXM1 responsible for resistance to venetoclax. Venetoclax in combination with FOXM1 inhibitor STL001 inhibited BCL2A1 and circumvented venetoclax resistance. Pharmacological inhibition of the FOXM1-BCL2A1 axis represents a therapeutic strategy to sensitize AML cells to venetoclax-induced apoptosis.

Keywords: AML; BCL2A1; FOXM1; apoptosis; venetoclax.

Figure 1

FOXM1 confers venetoclax resistance by stabilizing BCL2A1 expression levels.A, expression changes in genes contributing to venetoclax resistance in AML. The heatmap displays log2-transformed differences in gene expression levels between FOXM1-KD and untreated control KG-1 cells. Data were exported from our previously published RNA-Seq analysis in AML cell line KG-1 (https://doi.org/10.3389/fonc.2021.696532). B, relative fold change of BCL2A1 mRNA in KG-1-FOXM1-KD compared with control scramble vector–transduced KG-1 cells. GAPDH was used as an internal loading control. Data presented as mean ± SEM (n = 3 biologically independent replicates; ∗∗p < 0.002), for comparison two-tailed Student’s t-test was used. C, KG-1 cells with stable shRNA–mediated FOXM1-KD grown in the absence or the presence of venetoclax (50 nM) at indicated time points and compared with control KG-1 cells. In all cases, total protein samples were obtained from cells immediately after treatment and analyzed for FOXM1, BCL2A1, and cleaved caspase-3 levels via immunoblotting, and β-actin was used as an internal loading control. The blot shown is representative of two independent experiments with consistent results. D, KG-1 cells with stable shRNA–mediated FOXM1-KD, doxycycline (Dox)-inducible FOXM1b overexpression, or control vector were treated with indicated concentrations of venetoclax for 24 h. In all cases, total protein samples were obtained from cells immediately after treatment and analyzed for FOXM1, Bcl-XL, BCL2, BCL2A1, and cleaved caspase-3 levels via immunoblotting, and β-actin was used as internal loading control. The blot shown is representative of two independent experiments with consistent results. AML, acute myeloid leukemia; FOXM1, Forkhead box M1; KD, knockdown.

Figure 2

FOXM1-mediated stabilization of BCL2A1 expression is sufficient to protect cells from BCL2 inhibitor venetoclax (Ven)–induced apoptosis.A, KG-1 cells with stable shRNA–mediated FOXM1-KD, BCL2A1-KD, or control vector were treated with increasing concentrations of Ven (0, 50, and 100 nM). The percentage of dead cells was determined via a trypan blue exclusion test, n = 3 biologically independent replicates. B, KG-1 cells with stable shRNA–mediated BCL2A1-KD, FOXM1-KD, or control vector were treated with increasing concentrations of Ven (0, 50, and 100 nM). In all cases, total protein samples were obtained from cells immediately after treatment and analyzed for FOXM1, BCL2A1, and cleaved caspase-3 levels via immunoblotting, and β-actin was used as internal loading control. C–E, densitometry quantitation of bands in three independent experiments related to B. F, FOXM1-KD-KG-1 cells with doxycycline (Dox)-inducible BCL2A1 expression lentivirus, FOXM1-KD, or control vector were treated with increasing concentrations of Ven (0, 50, and 100 nM). In all cases, total protein samples were obtained from cells immediately after treatment and analyzed for FOXM1, BCL2A1, and cleaved caspase-3 levels via immunoblotting, and β-actin was used as internal loading control. G and H, densitometry quantitation of bands in three independent experiments related to F. I, KG-1 cells were treated with indicated concentrations of Ven alone or in combination with STL001 for 24 h. In all cases, total protein samples were obtained from cells immediately after treatment and analyzed for FOXM1, BCL2A1, and cleaved caspase-3 levels via immunoblotting, and β-actin was used as internal loading control. J–L, densitometry quantitation of bands in three independent experiments related to I. M, HL-60 cells with stable shRNA–mediated FOXM1-KD, BCL2A1-KD, or control vector were treated with increasing concentrations of Ven (0, 50, and 100 nM). The percentage of dead cells was determined via a trypan blue exclusion test, n = 3 biologically independent replicates. N, HL-60 cells with stable shRNA–mediated BCL2A1-KD, FOXM1-KD, or control vector were treated with increasing concentrations of Ven (0, 50, and 100 nM). In all cases, total protein samples were obtained from cells immediately after treatment and analyzed for FOXM1, BCL2A1, and cleaved caspase-3 levels via immunoblotting, and β-actin was used as an internal loading control. The blot shown is representative of two independent experiments with consistent results. O, HL-60 cells were treated with indicated concentrations of Ven alone or in combination with STL001 for 24 h. In all cases, total protein samples were obtained from cells immediately after treatment and analyzed for FOXM1, BCL2A1, and cleaved caspase-3 levels via immunoblotting, and β-actin was used as an internal loading control. The blot shown is representative of two independent experiments with consistent results. B, F, I, band intensities were quantified using ImageJ software; data are normalized to “Control/0 nM” sample; all data are presented as mean ± SEM; n = 3 biologically independent replicates; comparison of mean values between multiple groups was evaluated by one-way ANOVA; two-way ANOVA was used to compare groups with two independent variables (∗∗p < 0.002, ∗∗∗p < 0.001, ns, nonsignificant difference). FOXM1, Forkhead box M1; KD, knockdown.