AKT/FOXO signaling enforces reversible differentiation blockade in myeloid leukemias

Abstract

AKT activation is associated with many malignancies, where AKT acts, in part, by inhibiting FOXO tumor suppressors. We show a converse role for AKT/FOXOs in acute myeloid leukemia (AML). Rather than decreased FOXO activity, we observed that FOXOs are active in ∼40% of AML patient samples regardless of genetic subtype. We also observe this activity in human MLL-AF9 leukemia allele-induced AML in mice, where either activation of Akt or compound deletion of FoxO1/3/4 reduced leukemic cell growth, with the latter markedly diminishing leukemia-initiating cell (LIC) function in vivo and improving animal survival. FOXO inhibition resulted in myeloid maturation and subsequent AML cell death. FOXO activation inversely correlated with JNK/c-JUN signaling, and leukemic cells resistant to FOXO inhibition responded to JNK inhibition. These data reveal a molecular role for AKT/FOXO and JNK/c-JUN in maintaining a differentiation blockade that can be targeted to inhibit leukemias with a range of genetic lesions.

Figure 1. Constitutive Akt Activation Promotes Myeloid Maturation and Apoptosis of Leukemic Cells

(A) Lineage^low, Sca-1⁻, cKit^high, CD34⁺ cells purified from healthy and leukemic mice were stimulated with or without mSCF and then subjected to flow cytometry with phospho-Akt^Ser473 (CD34⁺ myeloid progenitors [MP] versus CD34⁺ leukemic progenitors [LP], p = 0.0478). Right panel is a histogram from a single experiment with the left panel representing mean ± standard error of the mean (SEM) from three experiments. (B) Mononuclear bone marrow cells (MNBCs) recovered from MLL-AF9 leukemic mice were infected with MSCV-IRES-GFP control (Ctrl) or myr-Akt-expressing retroviruses. Cells from each condition were then treated with vehicle or 10 nM or 100 nM rapamycin and evaluated for numbers of GFP⁺ cells every 2 days using flow cytometry (day 6 *Ctrl versus myr-Akt, vehicle, p = 0.0005; **Ctrl versus myr-Akt, 10 nM rapamycin, p = 0.0008; ***Ctrl versus myr-Akt, 100 nM rapamycin, p = 0.0046; n = 3). Data are represented as the mean ± standard deviation (SD). (C) GFP⁺ cells treated as described above were analyzed for CD11b expression (*Ctrl versus myr-Akt, vehicle, p < 0.0001; **Ctrl versus myr-Akt, 10 nM rapamycin, p < 0.0001; *** Ctrl versus myr-Akt, 100 nM rapamycin, p < 0.0001; n = 3) or (D) stained with May-Grünwald Giemsa. Data are represented as the mean ± SD. (E and F) Flow cytometric analysis of control and myr-Akt-infected cells incubated with pHrodo fluorescent-labeled E. coli particles. (E) Flow cytometric histogram plot of pHrodo-stained Ctrl versus myr-Akt GFP⁺ cells and (F) graphical representation of three replicates (*Ctrl versus myr-Akt, vehicle, p < 0.0001; **Ctrl versus myr-Akt, 10 nM rapamycin, p < 0.0001; ***Ctrl versus myr-Akt, 100 nM rapamycin, p = 0.0009; n = 3). Data are represented as the mean ± SD. (G) Flow cytometric analysis of Annexin V expression on Ctrl and myr-Akt-expressing cells (*Ctrl versus myr-Akt, p = 0.0006; n = 3). Data are represented as the mean ± SD. See also Figure S1.

Figure 2. FoxOs Are Active and Suppress Myeloid Maturation in Murine AML Cells

(A) Table of activated (left panel) and repressed (right panel) FoxO target genes differentially expressed between GMP and L-GMP microarray datasets (D-Chip analysis, p = 0.95). (B) Immunofluorescence of purified lineage^low, Sca-1⁻, cKit^high, CD34⁺, FcgRII/III⁺ cells from healthy and MLL-AF9-induced leukemic mice with FoxO3-specific antibodies (75D8) and DAPI contrast. (C) Mononuclear bone marrow leukemia cells expressing MLL-AF9 and bearing floxed alleles for FoxO1, FoxO3, and FoxO4 (FoxO1/3/4^floxed;MLL-AF9 cells) were infected with Ctrl or CreER-expressing recombinant retroviruses and then treated with vehicle or 400 nM 4-hydroxytamoxifen (4-OHT) for 4–6 hr. 48–72 hr following treatment, cells from all conditions were subjected to western blotting with FoxO3, Tubulin, and Cre antibodies. (D) Five days following treatment, Ctrl and CreER cells from each condition were assessed by flow cytometry for CD11b and Gr-1 expression (*CreER + 4-OHT versus CreER + vehicle, Ctrl + vehicle, or Ctrl + 4-OHT, p < 0.0001; n = 3, data represented as the mean ± SEM) and (E) stained with May-Grünwald Giemsa. See also Figure S2.

Figure 3. FOXO3 Is Active and Required to Preserve the Immature State of Human AML Cell Lines

(A)THP-1 and Mono-mac-6 (MM6) (both MLL-AF9⁺)and SKM-1 and NB4(both MLLAF9⁻) leukemia cell lines were fractionated into nuclear (N) and cytoplasmic (C) extracts and subjected to western blotting with FOXO3, ORC2 (nuclear), and Tubulin (cytoplasmic) antibodies. (B) MOLM-14, MM6, and SKM-1 cells were stably transduced with recombinant lentiviruses expressing either nontargeting (NT) or FOXO3 (F3-1 or F3-2) shRNAs. Cells were then subjected to western blotting with FOXO3 and Tubulin antibodies or (C) counted either everyday (SKM-1 and NB4) or every 2 days (MOLM-14 and MM6) following stable expression of designated shRNAs (day 6 MOLM-14, NT versus F3-1, *p = 0.0003; NT versus F3-2, **p < 0.0001; day 6 MM6, NT versus F3-1, *p < 0.0001; NT versus F3-2, **p = 0.0002; day 4 SKM-1, NT versus F3-1, *p = 0.0062; NT versus F3-2, **p = 0.0083; day 4 NB4, NT versus F3-1, *p = 0.0003; NT versus F3-2, **p = 0.0058). Data are represented as the mean ± SD. (D) Transduced MOLM-14 and SKM-1 cells were analyzed by flow cytometry for human CD11b expression (MOLM-14, NT versus F3-1, *p< 0.0001; NT versus F3-2, **p < 0.0001; SKM-1, NT versus F3-1, *p < 0.0001; NT versus F3-2, **p = 0.0115). Data are represented as the mean ± SD. (E) Flow cytometric analysis of transduced MM6 cells incubated with pHrodo particles. (F) May-Grünwald Giemsa staining of transduced MOLM-14, SKM-1, and NB4 cells. (G) Flow cytometric analysis of transduced SKM-1 cells stained with Annexin V and CD11b (*NT shRNA versus FOXO3 shRNA-1, CD11b⁺, p < 0.0001; **NT shRNA versus FOXO3 shRNA-2, CD11b⁺, p = 0.0007; n = 3). Data are represented as the mean ± SD. NT = NT shRNA, F3-1 = FOXO3 shRNA-1, and F3-2 = FOXO3 shRNA-2; n = 3. See also Figure S3.

Figure 4. Primary AMLs Derived from Patients Separate into Distinct Clusters of FOXO Activity

(A) BM cells derived from patients with AML were stained with human lineage cocktail and human CD34 (both BD Biosciences), and then lineage^low, CD34⁺ cells were isolated by flow cytometry. Nuclear (N) and cytoplasmic (C) extracts of total MNBCs (TBM) and lineage^low, CD34⁺ cells were subjected to western blotting with FOXO3, ORC2, and Tubulin antibodies. (B) Lineage^low, CD34⁺ cells from three AML patients analyzed as described above. (C) Patient samples #1 and #6 were transduced with recombinant lentiviruses expressing either NT shRNA or FOXO3 shRNA-1, then grown in liquid culture for 8 days, and then assessed for CD11b expression. Bar graphs are represented as the mean ± SD. (D) Transduced patient samples #1 and #6 cells were placed in methylcellulose supplemented with human cytokines. Graph represents the enumeration of colonies formed after 8 days of culture. Data are represented as the mean ± SD. (E) Transduced cells from patient sample #1 stained with Wright-Giemsa after 8 days of liquid culture. (F) Gene list comprising the FOXO-specific gene signature generated from comparing the gene expression array data of murine lineage^low, Sca-1⁺, cKit⁺ (LSK) cells in animals without (+/+) and with (Δ/Δ) FoxO1/3/4 deletion (comprehensive gene set located in Table S1). (G) Hierarchical cluster analysis based on the overlap of the murine FOXO gene signature stratified over the gene expression array data of 436 individual primary AML samples. See also Figure S4 and Table S1.

Figure 5. Deletion of FoxO Transcription Factors Suppresses MLL-AF9-Induced Leukemia In Vivo

(A) Experimental scheme used in Figures 5B–5H where MLL-AF9⁺FoxO1/3/4^floxed murine bone marrow (BM) cells recovered from a leukemic primary recipient mouse carrying the Mx1-Cre transgene (Cre⁺) or not (Cre⁻) were transplanted into recipient mice. Cre⁺ and Cre⁻ mice were administered saline or pI-pC and then all mice from each Cre condition were assessed for leukemic burden (day 29 for Cre⁻ and day 39 for Cre⁺). (B) Mean spleen weight of Cre⁺ (pI-pC) versus Cre⁺ (saline) (p < 0.0001; n = 4). Data represented as the mean ± SEM. (C) Gross anatomical view of spleens recovered from Cre⁺ mice administered pI-pC (right) or saline (left). (D) WBC analysis of peripheral blood collected every 4–14 days post-transplant from Cre⁺ saline- and pI-pC-treated mice (n = 4). (E and F) Kaplan-Meier survival curve analysis of mice transplanted and treated as described above (E: Cre⁻, p = 0.6899; n = 6 and F: Cre⁺, p = 0.0009; n = 10). (G) Leukemic BM cells isolated from Cre⁺ mice administered saline or pI-pC 7 days earlier were analyzed for the mean proportion ±SEM of L-GMPs (lineage^low, Sca-1 ^–, cKit^high, CD34⁺, FcgRII/III⁺, p = 0.039; n = 3). (H and I) Leukemic BM cells isolated from Cre⁺ mice administered saline (H) or pI-pC (I) 7 days earlier were transplanted into tertiary recipients at various cell numbers: 300 (n = 6), 3,000 (n = 6), 30,000 (n = 4), and 300,000 (n = 4). Kaplan-Meier analysis of animals that developed leukemia is shown. LIC frequencies were calculated using poisson statistics (H: LICfreq^+/+ = 1:5,314 and I: LICfreq^Δ/Δ = 1: 86,044). See also Figure S5.

Figure 6. The JNK/c-JUN Signaling Pathway Antagonizes Maturation and Apoptosis Mediated by FOXO Inhibition in AML

(A) MOLM-14 cells stably expressing either nontargeting (NT) or FOXO3 (FOXO3-1) shRNA were subjected to western blotting with antibodies that specifically recognize FOXO3, c-JUN, JNK, Tubulin, or the phosphorylated forms of c-JUN (pc-JUN^S63) and JNK (pJNK^T183/Y185). (B and C) MOLM-14 cells stably expressing nontargeting (NT) or FOXO3 (F3-1) shRNA were treated with 10 µM SP600125 (JNK inhibitor) or vehicle. Forty-eight hours later, cells from each condition were assessed for cell number (B, NT [vehicle] versus NT [SP600125], *p = 0.0095, NT [SP600125] versus F3-1 [vehicle], **p = 0.0008, F3-1 [vehicle] versus F3-1 [SP600125], ***p < 0.0001) and Annexin V staining (C, NT [vehicle] versus NT [SP600125], *p = 0.0282, NT [SP600125] versus F3-1 [vehicle], **p = 0.0133, F3-1 [vehicle] versus F3-1 [SP600125], ***p = 0.0002). Data are represented as the mean ± SD. (D) Control and myr-Akt-expressing MLL-AF9⁺ leukemia BM cells were subjected to western blotting with FoxO3, c-Jun, Jnk, Tubulin, or the phosphorylated forms of FoxO3 (pFoxO3^S256), c-Jun (pc-JUN^S63), and Jnk (pJnk^T183/Y185) antibodies. (E and F) Control and myr-Akt-expressing MLL-AF9⁺ leukemia BM cells were treated with 10 µM SP600125 (JNK inhibitor) or vehicle. Forty-eight hours after treatment, cells from each condition were assessed for cell number (E, Ctrl [vehicle] versus Ctrl [SP600125], *p = 0.0373, Ctrl [vehicle] versus myr-Akt [vehicle], **p=0.0381, myr-Akt [vehicle] versus myr-Akt [SP600125], ***p< 0.0001) and Annexin Vstaining (F, Ctrl [SP600125] versus myr-Akt [vehicle], *p =0.0011, myr-Akt [vehicle] versus myr-Akt [SP600125], **p < 0.0001). Data are represented as the mean ± SD. See also Figure S6.

Figure 7. c-JUN Activity Is Upregulated in AMLs Displaying Constitutive AKT Activation or FOXO Inhibition

(A) BM cells recovered from mice that succumbed to MLL-AF9-induced AML (refer to Figure 5F) that were wild-type or null for FoxO1/3/4 were subjected to western blotting with pc-Jun^S63 and Tubulin antibodies. (B) BM cells recovered from mice that succumbed to control or myr-Akt-expressing MLL-AF9-in-duced AML (refer to Figure S5G) were subjected to western bloting with antibodies that recognize Akt, Tubulin, or the phosphorylated forms of FoxO3 (pFoxO3^S256) or c-Jun (pc-Jun^S63) antibodies. (C) Mean expression levels of FOXO1, FOXO3, and c-JUN in the FOXO signature-based, hierarchical cluster-defined primary AML sample groups (p < 0.0001; see Figure 4G). Data are represented in a box plot with whiskers.