Purpose: The Forkhead transcription factors (FOXO) are tumor suppressor genes regulating differentiation, metabolism, and apoptosis that functionally interact with signal transduction pathways shown to be deregulated and prognostic in acute myelogenous leukemia (AML). This study evaluated the level of expression and the prognostic relevance of total and phosphorylated FOXO3A protein in AML.

Experimental Design: We used reverse-phase protein array methods to measure the level of total and phosphoprotein expression of FOXO3A, in leukemia-enriched protein samples from 511 newly diagnosed AML patients.

Results: The expression range was similar to normal CD34+ cells and similar in blood and marrow. Levels of total FOXO3A were higher at relapse compared with diagnosis. Levels of pFOXO3A or the ratio of phospho to total (PT) were not associated with karyotpe but were higher in patients with FLT3 mutations. Higher levels of pFOXO3A or PT-FOXO3A were associated with increased proliferation evidenced by strong correlation with higher WBC, percent marrow, and blood blasts and by correlation with higher levels of Cyclins B1, D1 and D3, pGSK3, pMTOR, and pStat5. Patients with High levels of pFOXO3A or PT-FOXO3A had higher rates of primary resistance and shorter remission durations, which combine to cause an inferior survival experience (P = 0.0002). This effect was independent of cytogenetics. PT-FOXO3A was a statistically significant independent predictor in multivariate analysis.

Conclusions: High levels of phosphorylation of FOXO3A is a therapeutically targetable, independent adverse prognostic factor in AML. Clin Cancer Res; 16(6); 1865–74

Translational Relevance

In this article, we show that the phosphorylation of FOXO3A, a protein that interacts with numerous signal transduction pathways, is an adverse prognostic factor in AML that is associated with increased proliferation, resistance to therapy, and shorter survival. There is active investigation into developing therapies aimed at restoring FOXO3A function by altering localization or inhibiting phosphorylation. Strategies designed to inhibit AKT activity might counteract this effect in patients with high phospho to total ratios. In contrast, strategies designed to drive resting cells into proliferation might counteract the inhibitory effect of high levels of unphosphorylated FOXO3A but would be unlikely to have additional effects on those that are already highly proliferative due to high PT-FOXO3A ratios. FOXO3A therefore makes an attractive target for therapy in AML but correct application of FOXO3A-directed therapies requires defining the mechanism by which FOXO3A is inactivated to optimize efficacy. This research therefore defines FOXO3A as a target and guides the application of FOXO3a directed therapies.

Forkhead transcription factors are a superfamily of evolutionary conserved proteins that function in diverse physiologic processes including cellular differentiation, tumor suppression, metabolism, cell cycle arrest, resistance, and apoptosis (1, 2). The forkhead genes are grouped into 19 subclasses of FOX genes (FOXA to FOXS) based on a conserved 100-residue DNA binding “forkhead box” domain (1, 35). The Forkhead O transcription factor subfamily (FOXO), which consists of four members FOXO1, FOXO3, FOXO4, and FOXO6, is deregulated in several tumor types (3). The role of FOXO proteins in tumorigenesis were initially identified by their involvement at sites of chromosomal rearrangements in humans tumors (68). For example, FOXO1, FOXO3A, and FOXO4 genes were found at chromosomal breakpoints in human soft-tissue tumors and leukemias. The FOXO1 gene was identified in the studies of t(2, 13)(q35;q14) and t(1, 13)(p36;q14) chromosomal translocations found in rhabdomyosarcomas, whereas the FOXO3 gene was identified at t(6;11)(q21;q23) chromosomal translocation from an acute myeloid leukemia patient and FOXO4 at t(x;11)(q13;q23) in acute lymphoblastic leukemia (9). FOXO3A overexpression inhibits tumor growth in vitro and tumor size in vivo in breast cancer cells and cytoplasmic location of FOXO3A correlated with poorer survival in patients with breast cancer (10, 11). Genetic deletion of five FOXO alleles showed modest neoplastic phenotypes, whereas deletion of all of the FOXO alleles generated thymic lymphomas and hemangiomas (12). These data define FOXO proteins as bona fide tumor suppressor genes (12).

Studies by Fei et al. (13) have suggested that low expression of FOXO3A expression is associated with the development of ovarian tumors. More recently, studies have shown the importance of FOXOs in preserving the self-renewal capacity of hematopoietic stem cells through undefined mechanisms (14). Studies have shown that FOXO3A transcriptional activity may be required to prevent B-chronic lymphocytic leukemia and chronic myelogenous leukemia (14, 15).

The FOXO transcription factors are functionally integrated (see Fig. 1 for a summary) with several signal transduction pathways including many kinase pathways, e.g., c-Jun-NH2-kinase (16), MST1 (17) AMPK (18), IkB (11), SGK (19), CDK2 (20), extracellular signal-regulated kinase (ERK)1/2 (10), DYRK1A (21), Caspase (22), phosphoinositide 3-kinase -AKT (23), and Wnt (24), which can regulate a broad array of cellular process that include stem cell proliferation (14, 25), aging (26), and malignancy (2). FOXO proteins can act not only as transcriptional activators but also as transcriptional repressors (12, 15, 27, 28). Their functions are tightly regulated at multiple levels by phosphorylation, ubiquitylation, acetylation, and protein-protein interactions (2931). Phosphorylation on threonine 24, serine 256, or serine 318 by AKT inhibits FOXO3A activity by increasing nuclear export and this in turn increases proliferation (32). Similar to the retinoblastoma tumor suppressor gene, FOXO function could therefore be lost by either diminished expression or the inactivation by phosphorylation. FOXO transcription factors are therefore emerging as master signaling regulators, which control various physiologic and pathologic processes, including cancer protection (15, 28, 33). This combination of functional activity combined with the frequency of abnormal expression in malignancy makes the FOXO proteins interesting targets for therapeutic intervention.

Fig. 1.

FOXO proteins are involved in multiple signaling pathways that control cell proliferation, apoptosis, survival, and immune functions in response to external and internal stimuli. Stimulated growth factors (insulin or insulin-like growth factor I) and their receptors (such as epidermal growth factor receptor) lead to activation of ERK (A), Akt (B) and CDK2 (C) through RAS and phosphoinositide 3-kinase–dependent phosphorylation of FOXOs. ERK and Akt-mediated phosphorylation of FOXO proteins inhibit transcription of the gene that promote apoptosis (Bim, FasL, and TRAIL) and cell cycle arrest (p27, p21, and c-myc) by exporting FOXO proteins to cytoplasm and inducing ubiquitination and proteosomal degradation. Akt-mediated phosphorylation allows FOXOs cytoplasmic retention through 14-3-3 protein binding, whereas ERK pathway activation leads to FOXO ubiquitination through MDM2 binding. Similarly, tumor necrosis factor α (TNFα) stimulates IKK, which phosphorylates FOXO and thus promote cytoplasmic retention. FOXOs phosphorylation by CDKs also leads to cytoplasmic localization. However, this activity is abolished on any DNA damage, leading to the activation of CHK1 and CHK2. In case of any oxidative stress, FOXO proteins translocate back to the nucleus with the help of c-Jun-NH2-kinase (JNK)–dependent and MST1 phosphorylation (D) and on interaction with β-catenins.

Fig. 1.

FOXO proteins are involved in multiple signaling pathways that control cell proliferation, apoptosis, survival, and immune functions in response to external and internal stimuli. Stimulated growth factors (insulin or insulin-like growth factor I) and their receptors (such as epidermal growth factor receptor) lead to activation of ERK (A), Akt (B) and CDK2 (C) through RAS and phosphoinositide 3-kinase–dependent phosphorylation of FOXOs. ERK and Akt-mediated phosphorylation of FOXO proteins inhibit transcription of the gene that promote apoptosis (Bim, FasL, and TRAIL) and cell cycle arrest (p27, p21, and c-myc) by exporting FOXO proteins to cytoplasm and inducing ubiquitination and proteosomal degradation. Akt-mediated phosphorylation allows FOXOs cytoplasmic retention through 14-3-3 protein binding, whereas ERK pathway activation leads to FOXO ubiquitination through MDM2 binding. Similarly, tumor necrosis factor α (TNFα) stimulates IKK, which phosphorylates FOXO and thus promote cytoplasmic retention. FOXOs phosphorylation by CDKs also leads to cytoplasmic localization. However, this activity is abolished on any DNA damage, leading to the activation of CHK1 and CHK2. In case of any oxidative stress, FOXO proteins translocate back to the nucleus with the help of c-Jun-NH2-kinase (JNK)–dependent and MST1 phosphorylation (D) and on interaction with β-catenins.

Close modal

There is limited data available on the link between FOXO3A and the progression of leukemia. A previous study showed that increased expression of nuclear FOXO3A was found to be associated with increased drug resistance in leukemic cells through enhanced phosphoinositide 3-kinase PI3K/AKT activity (34). A forthcoming publication found that high FOXO3a mRNA levels correlated with inferior survival in patients with normal karyotype acute myelogenous leukemia (AML; ref. 35) but protein expression and activation status has not been reported. We have previously shown that numerous signal transduction pathways that regulate FOXO and which are also regulated by FOXO are abnormally activated in AML with adverse prognostic consequences (36). This led us to study the role of FOXO3A expression and inactivation through phosphorylation, in a cohort of 511 cases of AML and 21 cases of APL using reverse-phase protein array (RPPA). In this study, we report a correlation between phopsho-FOXO3A expression and clinical characteristics, and French-American-British classification (FAB) and overall survival in leukemia patients.

Patient population

Peripheral blood and bone marrow specimens were collected from 511 patients with newly diagnosed AML and 21 with newly diagnosed acute promyelocytic leukemia (APL) evaluated at The University of Texas M.D. Anderson Cancer Center (MDACC) between September 1999 and March 2007. Samples were acquired during routine diagnostic assessments in accordance with the regulations and protocols (Lab 01-473) approved by the Investigational Review Board of MDACC. Informed consent was obtained in accordance with the Declaration of Helinski. Samples were analyzed under and Institutional Review Board–approved laboratory protocol (Lab 05-0654). Samples were enriched for leukemic cells by performing ficoll separation to yield a mononuclear fraction followed by CD3/CD19 depletion to remove contaminating T and B cells, if they were calculated to be >5% based on the differential. The samples were normalized to a concentration of 1 × 104 cells/μL and a whole-cell lysate was prepared as previously described (36). A total of 387 marrow and 283 blood samples were studied from the newly diagnosed cases with 140 having both available. Among the 142 AML and 4 APL cases that relapsed, a paired relapse sample was available for 49 of the AML and 1 of the APL cases. All but one relapse samples were from the marrow. Outcomes analysis for this report is restricted to the newly diagnosed patients. The associated demographics are described in Table 1. Although this population is typical for the MDACC referral pattern, the median age of these patients (65.7 y) is older than the national average of 58 y, and this population has a high percentage of cases with unfavorable cytogenetics (49%) and a very high percentage with an antecedent hematologic disorder (AHD; 40%).

Table 1.

Demographic and clinical characteristics of 511 newly diagnosed AML patients in the study set

All casesTreatedP
AllLowest 3rdMiddle 3rdHighest 3rd
511 414 129 142 144 
Gender 
    Male/female 291:220 218:197 73:56 71:70 75:69 0.66 
Age 
    Min 15.8 17.4 27 18.8 17.39 0.32 
    Max 87.23 86.8 86.6 85.6 86.8  
    Median 65.7 64.16 64.24 65.3 63.58  
FAB 
    M0 5.7% 5.1% 3.9% 6.3% 4.9% 0.67 
    M1 10.8% 11.6% 8.5% 11.3% 14.6% 0.29 
    M2 33.1% 37.0% 44.2% 44.4% 22.9% 0.00002 
    M4 22.7% 23.4% 19.4% 22.5% 27.8% 0.26 
    M5 10.0% 9.9% 5.4% 4.2% 19.4% 0.000008 
    M6 5.3% 5.1% 6.2% 4.9% 4.2% 0.72 
    M7 2.0% 1.9% 2.3% 0.7% 2.8% 0.41 
    Unknown 3.3% 0.7% 0.8% 1.4% 0.0%  
    RAEBT 7.2% 5.6% 9.3% 4.2% 3.5% 0.068 
WHO* 
    Abnormal cytogenetics 9.2% 10.4% 13.2% 11.3% 8.3% 0.57 
    Multilineage dysplasia 21.5% 18.8% 24.8% 18.3% 13.9% 0.16 
    Therapy related 14.3% 12.8% 16.3% 12.0% 10.4% 0.36 
    Not in Others 55.0% 58.0% 45.0% 58.5% 68.8% 0.0008 
Cyto 
    Favorable 6.7% 8.0% 10.9% 10.6% 2.8% 0.019 
    Intermediate 44.0% 45.2% 37.2% 45.1% 52.1%  
    Unfavorable 49.3% 47.1% 51.9% 44.4% 45.1%  
FLT3 
    ITD 14.9% 17.9% 7.0% 15.5% 29.9% 0.000004 
    D835 3.1% 3.9% 1.6% 4.2% 5.6% 0.045 
    Both 1.6% 1.9% 0.0% 2.8% 2.8% 0.16 
Zubrod PS 
    3 or 4 3.3% 3.1% 1.6% 2.8% 4.9% 0.42 
AHD 
    ≥2 Mo 39.9% 37.4% 48.8% 34.5% 29.9% 0.004 
Infection 
    Yes 19.8% 22.2% 14.0% 21.1% 30.6% 0.004 
WBC 
    Median 8.8 9.9 3.55 9.2 30.1 <0.00001 
Platelet 
    Median 56 55.5 53 53.5 59 0.11 
Hemoglobin 
    Median 9.6 9.6 9.7 9.5 9.5 0.56 
% marrow blast 
    Median 46 50 38 46 70 <0.00001 
% blood blast 
    Median 18 20.5 20.5 48 <0.00001 
Response 
    CR NA 55.8% 59.7% 59.2% 48.6% 0.11 
    Resistant NA 34.3% 31.8% 31.7% 38.9% 0.34 
    Fail NA 10.1% 8.5% 9.2% 12.5% 0.49 
    Relapse NA 61.5% 62.3% 57.1% 64.3% 0.64 
    Alive NA 25.1% 24.8% 33.8% 16.7%  
Overall survival, median, wk NA 48.7 64.85 60.14 33.714 0.0002 
Remission duration, median, wk NA 45.57 41 52.9 34.7 0.12 
All casesTreatedP
AllLowest 3rdMiddle 3rdHighest 3rd
511 414 129 142 144 
Gender 
    Male/female 291:220 218:197 73:56 71:70 75:69 0.66 
Age 
    Min 15.8 17.4 27 18.8 17.39 0.32 
    Max 87.23 86.8 86.6 85.6 86.8  
    Median 65.7 64.16 64.24 65.3 63.58  
FAB 
    M0 5.7% 5.1% 3.9% 6.3% 4.9% 0.67 
    M1 10.8% 11.6% 8.5% 11.3% 14.6% 0.29 
    M2 33.1% 37.0% 44.2% 44.4% 22.9% 0.00002 
    M4 22.7% 23.4% 19.4% 22.5% 27.8% 0.26 
    M5 10.0% 9.9% 5.4% 4.2% 19.4% 0.000008 
    M6 5.3% 5.1% 6.2% 4.9% 4.2% 0.72 
    M7 2.0% 1.9% 2.3% 0.7% 2.8% 0.41 
    Unknown 3.3% 0.7% 0.8% 1.4% 0.0%  
    RAEBT 7.2% 5.6% 9.3% 4.2% 3.5% 0.068 
WHO* 
    Abnormal cytogenetics 9.2% 10.4% 13.2% 11.3% 8.3% 0.57 
    Multilineage dysplasia 21.5% 18.8% 24.8% 18.3% 13.9% 0.16 
    Therapy related 14.3% 12.8% 16.3% 12.0% 10.4% 0.36 
    Not in Others 55.0% 58.0% 45.0% 58.5% 68.8% 0.0008 
Cyto 
    Favorable 6.7% 8.0% 10.9% 10.6% 2.8% 0.019 
    Intermediate 44.0% 45.2% 37.2% 45.1% 52.1%  
    Unfavorable 49.3% 47.1% 51.9% 44.4% 45.1%  
FLT3 
    ITD 14.9% 17.9% 7.0% 15.5% 29.9% 0.000004 
    D835 3.1% 3.9% 1.6% 4.2% 5.6% 0.045 
    Both 1.6% 1.9% 0.0% 2.8% 2.8% 0.16 
Zubrod PS 
    3 or 4 3.3% 3.1% 1.6% 2.8% 4.9% 0.42 
AHD 
    ≥2 Mo 39.9% 37.4% 48.8% 34.5% 29.9% 0.004 
Infection 
    Yes 19.8% 22.2% 14.0% 21.1% 30.6% 0.004 
WBC 
    Median 8.8 9.9 3.55 9.2 30.1 <0.00001 
Platelet 
    Median 56 55.5 53 53.5 59 0.11 
Hemoglobin 
    Median 9.6 9.6 9.7 9.5 9.5 0.56 
% marrow blast 
    Median 46 50 38 46 70 <0.00001 
% blood blast 
    Median 18 20.5 20.5 48 <0.00001 
Response 
    CR NA 55.8% 59.7% 59.2% 48.6% 0.11 
    Resistant NA 34.3% 31.8% 31.7% 38.9% 0.34 
    Fail NA 10.1% 8.5% 9.2% 12.5% 0.49 
    Relapse NA 61.5% 62.3% 57.1% 64.3% 0.64 
    Alive NA 25.1% 24.8% 33.8% 16.7%  
Overall survival, median, wk NA 48.7 64.85 60.14 33.714 0.0002 
Remission duration, median, wk NA 45.57 41 52.9 34.7 0.12 

NOTE: The demographics of only those that were treated is shown in the second column and the distribution based on the ratio of phospho to total FOXO3A levels are shown in the next three columns. The P values refer to comparisons of the 3 PT-FOXO3A groups.

Abbreviation: NA, not applicable.

*Based on the 2008 classification schema.

Favorable, t(8:21) or inversion(16); intermediate, Diploid or −y; unfavorable, all others including −5, −7, +8, and t(6:9); miscellaneous changes, +21, 11q23.

Because not all cases were treated, outcome measurements were not applicable.

Of the 511 AML cases, 415 patients were treated at MDACC and are evaluable for outcome. Among these, 277 received high-dose ara-C (HDAC), 191 with an anthracycline, 49 with fludarabine (FLAG, FA), 28 with clofarabine, 8 with other agents, and 1 with an HDAC alone. Another 35 received standard-dose ara-C with clofarabine (n = 33) or with daunorubicin and etoposide (n = 2), whereas 8 received low-dose ara-C–based regimens. Idarubicin and troxacitabine was used in 2 cases and VNP40101M was used for 25 cases. Demethylating or histone deacetylating agents were used alone or in combination for 45 patients. Targeted agents were used in 13, 6 received gemtuzumab ozogamicin (GO) in combination with interleukin-11, and 4 received phase 1 agents. For the 21 APL cases, 15 received arsenic trioxide plus ATRA, and 6 of these also receive gemtuzumab ozogamicin. Four received idarubicin and all trans retinoicacid (ATRA), one in conjunction with gemtuzumab ozogamicin, and one each receive liposomal ATRA and gemtuzumab ozogamicin plus ATRA.

RPPA method

Proteomic profiling was done on samples from patients with AML using RPPA. The method and validation of the technique are fully described in previous publications (3739). Briefly, patient samples were printed in five serial dilutions onto slides along with normalization and expression controls. Slides were probed with a strictly validated primary antibody against total FOXO3A or phospho serine 318/321 (Cell Signaling) and a secondary antibody to amplify the signal, and finally a stable dye (40) is precipitated. The stained slides were analyzed using the Microvigene software (Vigene Tech) to produce quantified data.

Statistical analysis

Supercurve algorithms were used to generate a single value from the five serial dilutions (41). Loading control (42) and topographical normalization procedures accounted for protein concentration and background staining variations. Analysis using unbiased clustering, perturbation bootstrap clustering, and principle component analysis was then done as fully described in ref. (38). A composite variable based on the ratio of phospho to total FOXO3A was generated (PT-FOXO3A). Use of the ratio allows for recognition that a high ratio in the setting of low total levels would produce the same biological consequence of inactivation as a high ratio in the face of high levels of total FOXO3A. That information is not present when the results for the individual total or phosphor protein are analyzed. These variables were divided into thirds (low, medium, and high cohorts) based on the range of expression of all 511 samples.

Comparison of the protein levels between paired samples was done by performing paired t test. Association between protein expression levels and categorical clinical variables were assessed in R using standard t tests, linear regression, or mixed effects linear models. Association between continuous variable and protein levels were assessed by using the Pearson and Spearman correlation and linear regression. Bonferroni corrections were done to account for multiple statistical parameters for calculating statistical significance. The Kaplan-Meier method was used to generate the survival curves. Univariate and multivariate Cox proportional hazard modeling was done to investigate association with survival with protein levels as categorized variables using the Statistica version 6 software (StatSoft). This data set contains patients treated before some more recent prognostic markers were discovered (e.g., NPM1); therefore, the multivariate analysis did not contain all known AML prognostic markers.

FOXO3A, phospho FOXO3A (pFOXO3A), and the ratio of phospho to total FOXO3A (PT-FOXO3A) protein expression was analyzed in 511 newly diagnosed AML patients by RPPA. To determine the effect of source on protein levels, the expression in the 140 same day paired blood and marrow samples was compared. (Fig. 2, top row). No statistically significant differences were observed between blood- and marrow-derived samples, thus both source samples were combined for the subsequent data analysis. Compared with expression in normal CD34+ cells, levels of total and pFOX3A were ≤2 SDs below the mean in 30.3% and 17.8% of cases, respectively, and were infrequently ≥2 SDs above the mean (7.3% and 1.6%; Fig. 2. bottom row). Among the 49 paired diagnosis/relapse paired samples, the expression levels of pFOXO3A did not change with disease progression, but levels of total FOXO3A were significantly higher in relapsed samples (median fold change, 1.49; P = 0.0007).

Fig. 2.

Distribution of expression of FOXO3A and phosphoFOXO3A in AML based on source, disease status, and relative to normal CD34+ cells. Top, the ratio of expression between 140 paired blood and marrow specimens is shown. There was no significant difference between the two sources of leukemic blasts. Middle, the ratio of expression between 49 paired diagnosis and relapse specimens is shown. FOXO3A levels were significantly higher in relapse compared with diagnosis, but pFOXO3A levels did not differ. Bottom, comparison of expression distribution histograms of 511 AML samples (gray) relative to 10 normal CD34+ samples (black) is shown. Both FOXO3A and pFOXO3A showed expression levels that were significantly below that of the normal CD34+ cells.

Fig. 2.

Distribution of expression of FOXO3A and phosphoFOXO3A in AML based on source, disease status, and relative to normal CD34+ cells. Top, the ratio of expression between 140 paired blood and marrow specimens is shown. There was no significant difference between the two sources of leukemic blasts. Middle, the ratio of expression between 49 paired diagnosis and relapse specimens is shown. FOXO3A levels were significantly higher in relapse compared with diagnosis, but pFOXO3A levels did not differ. Bottom, comparison of expression distribution histograms of 511 AML samples (gray) relative to 10 normal CD34+ samples (black) is shown. Both FOXO3A and pFOXO3A showed expression levels that were significantly below that of the normal CD34+ cells.

Close modal

FOXO3, pFOXO3A, and PT ratio expression, and clinical characteristics

Variation in the level of total, phospho, and PT-FOXO3A by FAB classification or by cytogenetics is shown in Supplementary Fig. S1A and B. Cases of monocytic leukemia (M5) had significantly higher levels of pFOXO3A expression and higher PT ratios (P = 0.0000008). Patients with FAB M2 were more likely to have low PT ratios (P = 0.00002). Levels of total FOXO3A trended higher (P = 0.09) and levels of pFOXO3A were significantly higher (P = 0.002) in patients with intermediate and unfavorable cytogenetics. Patients with favorable cytogenetics were significantly less likely to have high PT-FOXO3a levels (0.017). Cases with either a FLT3-ITD or D835 mutation had significantly higher levels of total, (P = 0.001), phospho (P = 0.008), or PT-FOXO3A (P = 0.000004; Fig. 3A). Neither showed different expression on the basis of gender or performance status. Total (P = 0.008) and PT-FOXO3A (P = 0.004) but not phospho levels were higher in patients with AHD. Levels of pFOXO3A were significantly lower in cases with a history of prior malignancy, chemotherapy, or radiation therapy (P = 0.03, 0.02, and 0.01). Levels of total FOXO3A were significantly lower (P = 0.0004), and pFOXO3A (P = 0.02) and PT-FOXO3A (P = 0.004) were significantly higher in patients with active infections. The protein levels for total, pFOXO3A, and PT ratio were modestly but statistically significantly (R < 0.30; P < 0.000001) correlated with WBC, percent marrow blasts, percent blood blasts, and the absolute peripheral blood blast count (Fig. 3B–D), and the PT ratio was correlated with CD34. Categorized levels were strongly associated with these characteristics (see Table 1).

Fig. 3.

Correlation of phospho to total FOXO3A levels with (A) FLT3 ITD mutation status, (B) percentage bone marrow blasts, (C) WBC count, and (D) absolute peripheral blood blast count. Black box, 24 to 75 percentile levels; white box, the median level. Bars, 2 SDs. Outliers (O) and extremes (*) are shown. The results of the Kruskal-Wallis and F tests are shown in the other box.

Fig. 3.

Correlation of phospho to total FOXO3A levels with (A) FLT3 ITD mutation status, (B) percentage bone marrow blasts, (C) WBC count, and (D) absolute peripheral blood blast count. Black box, 24 to 75 percentile levels; white box, the median level. Bars, 2 SDs. Outliers (O) and extremes (*) are shown. The results of the Kruskal-Wallis and F tests are shown in the other box.

Close modal

Correlation with other proteins

This same RPPA array has been probed with 174 other antibodies and the full analysis will be published separately. Notably pFOXO3A or PT-FOXO3a levels were significantly positively correlated with proliferation markers Cyclins B1, D1, D3, GSK3, pGSK3, LYN, MTOR, pPKCδ, pPKCγ, and pStat3, (R = 0.42, 0.23, 0.36, 0.39, 0.30, 0.31, 0.35, 0.33, 0.21, and 0.23, all with P < 0.0000004) and negatively correlated with pAKT-thr308, pRB, Yap, and pYAP (R = −0.21, −0.25, −0.35, and −0.39, all P < 0.0000001). PT-FOXO levels were also strongly correlated with apoptosis proteins Bax, BAK, and inactivated Bad-pSerine136 (R = 0.25, 0.24, and 0.22; all P < 0.000001). A list of all proteins with R ≥ 0.2 is shown in Supplementary Fig. S2.

FOXO3A expression, overall survival, and remission duration

For outcome analysis, the values were categorized into thirds (low, medium, and high) based on the range of expression in all 511 cases. A higher proportion of those with low levels were not treated at MDACC, making the size of the “low” cohort smaller than the other two. Total FOXO3A levels were not prognostic of remission attainment. For pFOXO3A or PT-FOXO3A remission rates were similar for those with low or medium levels (∼60%) and were significantly higher compared with patients in the highest third (49%, P = 0.001). Relapse rates were not associated with total, p-FOXO3a, or PT-FOXO3a levels (P = 0.64). Higher levels of pFOXO3A and PT-FOXO3A were both highly prognostic adverse markers for overall survival (P = 0.0002 for both) with a median survival of 60 and 64 weeks for low and middle compared with 33 weeks for high PT-FOXO3A (Fig. 4A), but total FOXO3A was not (P = 0.34). This was also true among patients treated with more traditional HDAC-based therapies (n = 285; P = 0.0007) or those treated with nontraditional therapies (e.g., demethylating or histone deacetylating agents, gemtuzumab ozogamicin or VNP40101M; n = 72; P = 0.06). Neither total (P = 0.83), pFOXO3A (P = 0.21), or PT-FOXO3A (P = 0.11) were prognostic of remission duration (P = 0.83; Fig. 4B), although for all three, those with medium levels had the longest median remission duration. Identical results are observed if the variables are split at the median instead of into thirds.

Fig. 4.

Kaplan Meier curves for overall survival (top) and remission duration (bottom) for the ratio of phospho to total FOXO3A are shown.

Fig. 4.

Kaplan Meier curves for overall survival (top) and remission duration (bottom) for the ratio of phospho to total FOXO3A are shown.

Close modal

The effect of total, pFOXO3A, and PT-FOXO3A within different cytogenetic groups was evaluated as well. A high PT-FOXO3A level was adverse for patients with intermediate (P = 0.034) and unfavorable (P = 0.024) cytogenetics with a similar but not significant trend noted for those with favorable cytogenetics. This was independent of the presence or absence of a FLT3-ITD or D835 mutation. In agreement with the findings of Santamaria et al. (35), we observed that high levels of total FOXO3A was adverse for cases with intermediate cytogenetics (P = 0.02) and this effect was predominant among those with a FLT3 abnormality. High levels of pFOXO3A were adverse for those with favorable (P = 0.015) or unfavorable cytogenetics (P = 0.008)

For APL, there were eight, six, and seven patients with low, medium, or high PT-FOXO3A. All but one achieved complete remission (CR) (the lone failure presented with a WBC of 180K and had medium PT-FOXO3A). Four have relapsed, one with middle, and three with high PT-FOXO3A. Although the numbers are small, this suggests that high PT-FOXO3A may predict for relapse in APL (P = 0.09).

PT-FOXO3A level is an independent predictor of outcome in multivariate analysis

A Cox proportional hazard model was done evaluating for factors that were independent predictors of overall survival. Starting with 16 variables that were univariate predictors in this data set, or that have traditionally been prognostic (e.g., performance status, antecedent hematologic disorder), a stepwise analysis was conducted until only significant (P < 0.05) variables remained. This was followed by sequential add back of all previously removed variables, one by one, until a final model with only significant variables remained. The final model contained six variables, age, cytogenetics (favorable, intermediate, unfavorable) FLT3 mutation, albumin, PT-FOXO3A, and gender, as shown in Table 2. Performance status likely was not prognostic as only 3% of cases had a Zubrod PS of 3 or 4. This shows that the PT ratio of FOXO3A is an independent predictor of outcome in AML.

Table 2.

Multivariate analysis, final model

VariableβExponent βWald statisticP
Age (y) 0.03 1.04 65.66 <0.000001 
Cytogenetics −0.657 0.518 37.67 <0.000001 
Albumin −0.375 0.687 21.54 0.000003 
FLT3-abnormal 0.184 1.201 7.846 0.005 
PT-FOXO3A 0.191 1.21 6.98 0.008 
Gender 0.261 1.298 4.705 0.03 
VariableβExponent βWald statisticP
Age (y) 0.03 1.04 65.66 <0.000001 
Cytogenetics −0.657 0.518 37.67 <0.000001 
Albumin −0.375 0.687 21.54 0.000003 
FLT3-abnormal 0.184 1.201 7.846 0.005 
PT-FOXO3A 0.191 1.21 6.98 0.008 
Gender 0.261 1.298 4.705 0.03 

In the present study, total and phosphoprotein expression of FOXO3A was assessed by RPPA in AML patients. Expression was generally similar to, or at the lower end of the range of expression seen in normal CD34+ cells, and levels were similar in blood and marrow blasts. Levels of total FOX3A but not pFOXO3A were higher at relapse compared with diagnosis. Higher levels of pFOXO3A or the PT ratio were observed in FAB M5 and lower levels in FAB M2. Levels of p-FOXO3A or PT-FOXO3A were higher among patients with FLT3 mutations, substantiating observations in FLT3-ITD+ cell lines (43), but otherwise were not associated with specific cytogenetic abnormalities. Higher levels of p-FOXO3A or PT-FOXO3A were clearly associated with a higher proliferative potential as seen by the strong correlation between inactivated (phosphorylated) FOXO3A with higher WBC, percent marrow, and blood blasts. This was reinforced by the association of higher p-FOXO3A or PT-FOXO3A levels with higher levels of other proteins associated with proliferation including Cyclins B1, D1 and D3, pGSK3, pMTOR, and pStat5 as well as by negative correlations with pRB. Paradoxically levels were inversely correlated with PT-AKT-thr308 levels. Based on current understanding that phosphorylation leads to cytoplasmic localization and inactivation of FOXO3A, these findings are consistent with the idea that increased proliferation results when FOXO3A is inactivated in AML.

Although there is compelling evidence for a role for the FOX proteins in malignancy, the prognostic implications of their expression and activation has not been extensively studied (2). For FOXO3A, divergent observations are reported; high mRNA was adverse in diploid AML (35); high total FOXO3A protein levels were favorable in lymphoma (44); and low protein expression was adverse in ovarian and hepatocellular cancer (13, 45). The prognostic implications of phosphorylation has not been previously reported, and the activation state imparted by phosphorylation may account for these opposite findings. We found that high levels of phosphorylation of FOXO3A was an adverse prognostic factor for overall survival. These patients had higher rates of primary resistance and shorter remission durations that combine to cause an inferior survival experience. This effect was independent of cytogenetics as high p-FOXO3A or PT-FOXO3A was prognostic within individual cytogenetic groups, and remained a statistically significant independent variable in multivariate analysis along with cytogenetics. Validation of this finding in a data set with other prognostic factors such as NPM1 requires future research. Although not statistically significant, the remission duration of patients with low levels of p-FOXO3A or PT-FOXO3A was shorter than patients with middle ratios and their survival experience diverged after 104 weeks. These patients would be expected to have more FOX3A in the nucleus and this could lead to protection of leukemic stem cells from chemotherapy and ultimately to relapse. This would be consistent with the finding of Hui et al. (34) that increased expression of nuclear FOXO3A was associated with increased drug resistance in leukemic cells. This would imply that FOXO levels can influence responsiveness in two manners. When inactivated by phosphorylation, proliferation is increased and this creates a repopulation advantage for leukemic blasts that survive therapy, leading to higher rates of resistance and more rapid leukemic recurrence. When there are high levels of unphosphorylated FOXO3A present, it contributes to resistance of stem cells, possibly by maintaining them in a resting state, thereby enabling some to survive and eventually repopulate the marrow with leukemia. The first is a more rapid effect, the second more delayed. There is active investigation into developing therapies aimed at restoring FOXO3A function by altering localization or inhibiting phosphorylation (4648). Strategies designed to inhibit AKT activity might counteract this effect in patients with high PT ratios but would not be expected to affect the second mechanism. In contrast, strategies designed to drive resting cells into proliferation (G or granulocyte macrophage colony-stimulating factor, CXCR4 antagonists) might counteract the inhibitory effect of high levels of unphosphorylated FOXO3A but would be unlikely to have additional effects on those that are already highly proliferative due to high PT-FOXO3A ratios. FOXO3A therefore makes an attractive target for therapy in AML but correct application of FOXO3A-directed therapies may require defining the mechanism by which FOXO3A is inactivated to optimize efficacy.

We evaluated phosphorylation on a single site, but different kinases achieve inactivation by phosphorylating different sites on FOXO3A (e.g., AKT on T32, S253 and S315 and ERK on S294, S344 and S425, and IκK on S644; ref. 48). A more comprehensive analysis of FOXO3A phosphorylation might detect additional phosphorylated cases and improve the prognostic discrimination. Recent reports suggest that IκK phosphorylation is important in AML (49). However, there is likely to be significant overlap as we have previously shown that simultaneous activation of multiple signal transduction pathways is a common event in AML (36). In this data set, increased pAKT and pERK2 was strongly associated with increased FOXO3A (P < 0.00001; data not shown). In summary, phosphorylation of FOXO3A is an adverse prognostic factor in AML that is associated with increased proliferation, resistance to therapy, and shorter survival.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Greer
EL
,
Brunet
A
. 
FOXO transcription factors at the interface between longevity and tumor suppression
.
Oncogene
2005
;
24
:
7410
25
.
2
Myatt
SS
,
Lam
EW-F
. 
The emerging roles of forkhead box (Fox) proteins in cancer
.
Nat Rev Cancer
2007
;
7
:
847
59
.
3
Arden
KC
. 
Multiple roles of FOXO transcription factors in mammalian cells point to multiple roles in cancer
.
Exp Gerontol
2006
;
41
:
709
17
.
4
Tsai
KL
,
Sun
YJ
,
Huang
CY
,
Yang
JY
,
Hung
MC
,
Hsiao
CD
. 
Crystal structure of the human FOXO3a-DBD/DNA complex suggests the effects of post-translational modification
.
Nucleic Acids Res
2007
;
35
:
6984
94
.
5
Kaestner
KH
,
Knochel
W
,
Martinez
DE
. 
Unified nomenclature for the winged helix/forkhead transcription factors
.
Genes Dev
2000
;
14
:
142
6
.
6
Davis
RJ
,
D'Cruz
CM
,
Lovell
MA
,
Biegel
JA
,
Barr
FG
. 
Fusion of PAX7 to FKHR by the variant t(1;13)(p36;q14) translocation in alveolar rhabdomyosarcoma
.
Cancer Res
1994
;
54
:
2869
72
.
7
Borkhardt
A
,
Repp
R
,
Haas
OA
, et al
. 
Cloning and characterization of AFX, the gene that fuses to MLL in acute leukemias with a t(X;11)(q13;q23)
.
Oncogene
1997
;
14
:
195
202
.
8
Anderson
MJ
,
Viars
CS
,
Czekay
S
,
Cavenee
WK
,
Arden
KC
. 
Cloning and characterization of three human forkhead genes that comprise an FKHR-like gene subfamily
.
Genomics
1998
;
47
:
187
99
.
9
Barr
FG
. 
Gene fusions involving PAX and FOX family members in alveolar rhabdomyosarcoma
.
Oncogene
2001
;
20
:
5736
46
.
10
Yang
JY
,
Zong
CS
,
Xia
W
, et al
. 
ERK promotes tumorigenesis by inhibiting FOXO3a via MDM2-mediated degradation
.
Nat Cell Biol
2008
;
10
:
138
48
.
11
Hu
MC
,
Lee
DF
,
Xia
W
, et al
. 
IκB kinase promotes tumorigenesis through inhibition of forkhead FOXO3a
.
Cell
2004
;
117
:
225
37
.
12
Paik
JH
,
Kollipara
R
,
Chu
G
, et al
. 
FoxOs are lineage-restricted redundant tumor suppressors and regulate endothelial cell homeostasis
.
Cell
2007
;
128
:
309
23
.
13
Fei
M
,
Zhao
Y
,
Wang
Y
, et al
. 
Low expression of Foxo3a is associated with poor prognosis in ovarian cancer patients
.
Cancer Invest
2009
;
27
:
52
9
.
14
Tothova
Z
,
Kollipara
R
,
Huntly
BJ
, et al
. 
FoxOs are critical mediators of hematopoietic stem cell resistance to physiologic oxidative stress
.
Cell
2007
;
128
:
325
39
.
15
Kikuchi
S
,
Nagai
T
,
Kunitama
M
,
Kirito
K
,
Ozawa
K
,
Komatsu
N
. 
Active FKHRL1 overcomes imatinib resistance in chronic myelogenous leukemia-derived cell lines via the production of tumor necrosis factor-related apoptosis-inducing ligand
.
Cancer Sci
2007
;
98
:
1949
58
.
16
Barthelemy
C
,
Henderson
CE
,
Pettmann
B
. 
Foxo3a induces motoneuron death through the Fas pathway in cooperation with JNK
.
BMC Neurosci
2004
;
5
:
48
.
17
Lehtinen
MK
,
Yuan
Z
,
Boag
PR
, et al
. 
A conserved MST-FOXO signaling pathway mediates oxidative-stress responses and extends life span
.
Cell
2006
;
125
:
987
1001
.
18
Calnan
DR
,
Brunet
A
. 
The FoxO code
.
Oncogene
2008
;
27
:
2276
88
.
19
Leong
ML
,
Maiyar
AC
,
Kim
B
,
O'Keeffe
BA
,
Firestone
GL
. 
Expression of the serum- and glucocorticoid-inducible protein kinase, Sgk, is a cell survival response to multiple types of environmental stress stimuli in mammary epithelial cells
.
J Biol Chem
2003
;
278
:
5871
82
.
20
Huang
H
,
Regan
KM
,
Lou
Z
,
Chen
J
,
Tindall
DJ
. 
CDK2-dependent phosphorylation of FOXO1 as an apoptotic response to DNA damage
.
Science
2006
;
314
:
294
7
.
21
Woods
YL
,
Rena
G
,
Morrice
N
, et al
. 
The kinase DYRK1A phosphorylates the transcription factor FKHR at Ser329 in vitro, a novel in vivo phosphorylation site
.
Biochem J
2001
;
355
:
597
607
.
22
Obexer
P
,
Geiger
K
,
Ambros
PF
,
Meister
B
,
Ausserlechner
MJ
. 
FKHRL1-mediated expression of Noxa and Bim induces apoptosis via the mitochondria in neuroblastoma cells
.
Cell Death Differ
2007
;
14
:
534
47
.
23
Burgering
BM
,
Kops
GJ
. 
Cell cycle and death control: long live Forkheads
.
Trends Biochem Sci
2002
;
27
:
352
60
.
24
Maiese
K
,
Chong
ZZ
,
Shang
YC
,
Hou
J
. 
Rogue proliferation versus restorative protection: Where do we draw the line for Wnt and forkhead signaling?
Expert Opin Ther Targets
2008
;
12
:
905
16
.
25
Miyamoto
K
,
Araki
KY
,
Naka
K
, et al
. 
Foxo3a is essential for maintenance of the hematopoietic stem cell pool
.
Cell Stem Cell
2007
;
1
:
101
12
.
26
Partridge
L
,
Bruning
JC
. 
Forkhead transcription factors and ageing
.
Oncogene
2008
;
27
:
2351
63
.
27
Ramaswamy
S
,
Nakamura
N
,
Sansal
I
,
Bergeron
L
,
Sellers
WR
. 
A novel mechanism of gene regulation and tumor suppression by the transcription factor FKHR
.
Cancer Cell
2002
;
2
:
81
91
.
28
Ticchioni
M
,
Essafi
M
,
Jeandel
PY
, et al
. 
Homeostatic chemokines increase survival of B-chronic lymphocytic leukemia cells through inactivation of transcription factor FOXO3a
.
Oncogene
2007
;
26
:
7081
91
.
29
Vogt
PK
,
Jiang
H
,
Aoki
M
. 
Triple layer control: phosphorylation, acetylation and ubiquitination of FOXO proteins
.
Cell Cycle
2005
;
4
:
908
13
.
30
Brunet
A
,
Park
J
,
Tran
H
,
Hu
LS
,
Hemmings
BA
,
Greenberg
ME
. 
Protein kinase SGK mediates survival signals by phosphorylating the forkhead transcription factor FKHRL1 (FOXO3a)
.
Mol Cell Biol
2001
;
21
:
952
65
.
31
Brunet
A
,
Bonni
A
,
Zigmond
MJ
, et al
. 
Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor
.
Cell
1999
;
96
:
857
68
.
32
Arden
KC
. 
FoxO: linking new signaling pathways
.
Mol Cell
2004
;
14
:
416
8
.
33
Zou
Y
,
Tsai
WB
,
Cheng
CJ
, et al
. 
Forkhead box transcription factor FOXO3a suppresses estrogen-dependent breast cancer cell proliferation and tumorigenesis
.
Breast Cancer Res
2008
;
10
:
R21
.
34
Hui
RC
,
Gomes
AR
,
Constantinidou
D
, et al
. 
The forkhead transcription factor FOXO3a increases phosphoinositide-3 kinase/Akt activity in drug-resistant leukemic cells through induction of PIK3CA expression
.
Mol Cell Biol
2008
;
28
:
5886
98
.
35
Santamaria
CM
,
Chillon
MC
,
Garcia-Sanz
R
, et al
. 
High FOXO3a expression is associated with a poorer prognosis in AML with normal cytogenetics
.
Leuk Res
2009
;
33
:
1706
9
.
36
Kornblau
SM
,
Womble
M
,
Jackson
CE
,
Chen
W
,
Estey
EH
,
Andreeff
M
. 
Simultaneous activation of multiple signal transduction pathways confers poor prognosis in acute myelogenous leukemia
.
Blood
2006
;
108
:
2358
65
.
37
Kornblau
SM
,
Qiu
YH
,
Chen
W
, et al
. 
Proteomic profiling of 150 proteins in 511 acute myelogenous leukemia (AML) patient samples using reverse phase proteins arrays (RPPA) reveals recurrent proteins expression signatures with prognostic implications
.
Blood
2008
;
112
:
281
2
.
38
Kornblau
SM
,
Tibes
R
,
Qiu
Y
, et al
. 
Functional proteomic profiling of AML predicts response and survival
.
Blood
2009
;
113
:
154
64
.
39
Tibes
R
,
Qiu
YH
,
Lu
Y
, et al
. 
Reverse phase protein array (RPPA): validation of a novel proteomic technology and utility for analysis of primary leukemia specimens and hematopoetic stem cells (HSC)
.
Mol Cancer Ther
2006
;
5
:
2512
21
.
40
Hunyady
B
,
Krempels
K
,
Harta
G
,
Mezey
E
. 
Immunohistochemical signal amplification by catalyzed reporter deposition and its application in double immunostaining
.
J Histochem Cytochem
1996
;
44
:
1353
62
.
41
Hu
J
,
He
X
,
Baggerly
KA
,
Coombes
KR
,
Hennessy
BT
,
Mills
GB
. 
Non-parametric quantification of protein lysate arrays
.
Bioinformatics
2007
;
23
:
1986
94
.
42
Neeley
ES
,
Kornblau
SM
,
Coombes
KR
,
Baggerly
KA
. 
Variable slope normalization of reverse phase protein arrays
.
Bioinformatics
2009
;
25
:
1384
9
.
43
Scheijen
B
,
Ngo
HT
,
Kang
H
,
Griffin
JD
. 
FLT3 receptors with internal tandem duplications promote cell viability and proliferation by signaling through Foxo proteins
.
Oncogene
2004
;
23
:
3338
49
.
44
Zhao
Y
,
Fei
M
,
Wang
Y
,
Lu
M
,
Cheng
C
,
Shen
A
. 
Expression of Foxo3a in non-Hodgkin's lymphoma is correlated with cell cycle inhibitor p27kip
.
Eur J Hematol
2008
;
81
:
83
93
.
45
Lu
M
,
Ma
J
,
Xue
W
, et al
. 
The expression adn prognosis of FOXO3a and SKP2 in uman hepatocellular carcinoma
.
Pathol Oncol Res
2009
;
epub
:
1
9
.
46
Maiese
K
,
Chong
ZZ
,
Shang
YC
,
Hou
J
. 
Clever cancer strategies with FoxO transcription factors
.
Cell Cycle
2008
;
7
:
3829
39
.
47
Maiese
K
,
Chong
ZZ
,
Shang
YC
. 
OutFOXOing disease and disability: the therapeutic potential of targeting FoxO proteins
.
Trends Mol Med
2008
;
14
:
219
27
.
48
Yang
J-Y
,
Hung
M-C
. 
A new fork for clinical application: targeting forkhead transcription factors in cancer
.
Clin Cancer Res
2009
;
15
:
752
7
.
49
Chapuis
N
,
Park
S
,
Leotoing
L
, et al
. 
IκB kinase overcomes P13K to control Fox03a subcellular localization in acute myeloid leukemia
.
Blood
2009
;
114
:
197
.

Supplementary data