Programmed Cell Death-4 Tumor Suppressor Protein Contributes to Retinoic Acid–Induced Terminal Granulocytic Differentiation of Human Myeloid Leukemia Cells

Introduction

Acute myeloid leukemia (AML), the most common type of acute leukemia, is a heterogeneous group of hematologic malignancies characterized by a differentiation block in hematopoietic progenitor cells at the early stages of myelopoiesis, proliferation of immature blasts, and invasion of bone marrow. Acute promyelocytic leukemia, a subtype of AML, is characterized by a t(15;17) translocation involving the genes encoding promyelocytic leukemia and retinoic acid receptor α. This translocation results in differentiation arrest at the promyelocytic stage of myeloid cell differentiation (1). Despite recent improvements in our understanding of terminal cell differentiation, the molecular mechanisms regulating myeloid cell differentiation are not fully understood.

Differentiation therapy is based on the concept that differentiation-inducing agents can force cancer cells arrested at an immature or poorly differentiated state to resume the process of maturation (2). This type of treatment has the advantage of being potentially less toxic than standard chemotherapy. Induction of differentiation restores a natural cell death program and inhibits proliferation. Treatment of acute promyelocytic leukemia with all-trans retinoic acid (ATRA), the first model of differentiation therapy, has proved extremely successful in inducing clinical remission in most patients (3). ATRA, a naturally occurring derivative of vitamin A (retinol), is a potent inducer of cellular differentiation, growth arrest, and apoptosis in various tumor cell lines. ATRA induces terminal differentiation of immature leukemic promyelocytes into normal mature granulocytes in vitro and in vivo (4, 5). Thus, this system provides an excellent in vitro model for studying the molecular events taking place during the terminal differentiation of myeloid cells.

The ATRA-induced granulocytic differentiation program is a complex process that requires transcriptional and translational regulation of many specific targets (6-9). We previously found that ATRA inhibits the translational machinery through multiple posttranscriptional suppression mechanisms, including down-regulation of translation factors and up-regulation of a repressor of translation initiation, DAP5/p97, in myeloid progenitor cells during granulocytic differentiation (10). Currently, the posttranscriptional molecular mechanisms controlling translation initiation and the role of translational control in terminal myeloid cell differentiation remain largely unknown.

Programmed cell death 4 (PDCD4) is a recently discovered tumor suppressor gene that has attracted great interest as an inhibitor of tumor promoter–induced neoplastic transformation and as a specific inhibitor of cap-dependent mRNA translation in vitro and in vivo (11–15). PDCD4 was originally isolated from a human glioma library and is homologous to the mouse Pdcd4 (MA-3/TIS/A7-1) gene (16, 17). Human PDCD4 gene is localized to chromosome 10q24. PDCD4 has been shown to inhibit the activation of AP-1-dependent transcription, skin tumorigenesis, and tumor progression in transgenic mice (12, 18). PDCD4 suppresses translation initiation by specifically inhibiting the helicase activity of eukaryotic translation initiation factor 4A (eIF4A), a component of the translation initiation complex (13, 15). Binding of PDCD4 to eIF4A and consequent inhibition of translation is required for transrepression by PDCD4 of target activities such as AP-1 (15). PDCD4 is ubiquitously expressed in normal tissues, but its expression is lost or suppressed in several tumors, including lung, breast, colon, brain, and prostate cancers (19). Loss of PDCD4 expression in human lung cancer cells correlates with tumor progression and poor prognosis (20). The mechanism by which PDCD4 expression is suppressed is not understood. Recently, the chicken Pdcd4 gene has been identified as a direct target of the transcription factor c-Myb, which is essential for the development of the hematopoietic system, and plays an important role as a switch that directs hematopoietic progenitor cells to alternative fates, such as proliferation, differentiation, and apoptosis (21–23). We therefore hypothesized that PDCD4 plays a role in terminal differentiation and lineage commitment of human myeloid cells.

In the present study, we show that PDCD4 expression was markedly induced in AML cell lines and primary promyelocytic leukemia cells undergoing granulocytic differentiation. Cells that are undergoing monocytic/macrophagic differentiation and are resistant to granulocytic differentiation failed to up-regulate PDCD4 and translocate it into the nucleus after ATRA treatment. Targeted inhibition of PDCD4 expression by siRNA resulted in a significant inhibition of ATRA-induced granulocytic differentiation, suggesting that PDCD4 induction contributes to granulocytic differentiation. We also showed that the phosphatidylinositol 3-kinase (PI3K)/Akt pathway negatively regulates induced PDCD4 expression in AML cells and ATRA induces PDCD4 through inhibition of this pathway. Knockdown of PDCD4 antagonized the expression of c-myc, p27^Kip1, DAP5, and Wilms tumor 1 (WT1), suggesting that PDCD4 is involved in the regulation of these downstream proteins. Overall, PDCD4 may exert its effects on differentiation by altering the expression of these proteins.

Results

Evaluation of ATRA-Induced Differentiation

We first examined the surface expression of CD11b, a marker of myeloid differentiation, in NB4 cells and their differentiation-resistant derivatives, NB4.R1 cells. Cells were treated with 1 μmol/L ATRA, collected at 12, 24, 48, and 72 h, and analyzed by fluorescence-activated cell sorting with anti-CD11b antibody (Fig. 1A ). NB4 cells expressed CD11b after ATRA treatment, whereas NB4.R1 cells lacked surface CD11b expression up to 96 h after ATRA treatment (Fig. 1B), indicating that NB4.R1 cells did not undergo ATRA-induced differentiation. Morphologic changes in the cells were assessed by May-Grünwald-Giemsa staining, which revealed that the untreated NB4 cells were predominantly promyelocytes with characteristic cytoplasmic granules, large nuclei, and prominent nucleoli. The ATRA-treated NB4 cells acquired a granulocytic morphology that included a decreased nuclear/cytoplasmic ratio, appearance of cytoplasmic granules, chromatin condensation, and loss of nucleoli (Fig. 1C). To further confirm the induction of differentiation in NB4 cells, we examined the reorganization of promyelocytic leukemia nuclear bodies in the cells after 72 h of ATRA treatment. Immunostaining with anti–promyelocytic leukemia antibodies revealed a diffusely microspeckled pattern of promyelocytic leukemias in the nuclei of untreated control NB4 cells. However, in cells treated with ATRA, the microspeckled pattern disappeared and the size and brightness of the promyelocytic leukemia bodies returned to normal appearance. In contrast, the normal nuclear organization of promyelocytic leukemia protein was not seen in ATRA-treated NB4.R1 cells, indicating that these cells did not differentiate into granulocytes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

ATRA induces granulocytic differentiation in NB4 but not differentiation-resistant NB4.R1 cells. A. Time-dependent expression of myeloid differentiation marker CD11b. NB4 cells were treated with 1 μmol/L ATRA for up to 72 h, stained with monoclonal anti-CD11b antibody, and analyzed by flow cytometry to detect induction of granulocytic differentiation. B. ATRA-induced differentiation within 96 h in NB4 and NB4.R1 cells as detected by fluorescence-activated cell sorting (FACS) analysis of CD11b expression. C. ATRA induces morphologic changes in promyelocytic leukemia cells undergoing granulocytic differentiation. After treatment with ATRA, NB4 cells were stained with May-Grünwald-Giemsa dye to reveal formation of myeloperoxidase-containing granules in differentiated cells. D. ATRA induces reorganization of promyelocytic leukemia nuclear bodies in NB4 cells but not NB4.R1 cells. Cells were treated with 1 μmol/L ATRA for 72 h, stained with monoclonal anti–promyelocytic leukemia primary and FITC-labeled secondary antibodies, and analyzed by immunofluorescence microscopy. Nuclei were stained with 4′,6-diamidino-2-phenylindole (blue).

ATRA Induces PDCD4 Expression during NB4 and HL60 Cell Differentiation but not in Differentiation-Resistant Cells

Because the translational machinery and protein synthesis are significantly inhibited in myeloid cells undergoing terminal differentiation (10, 24–26), we investigated whether PDCD4 expression is induced during the granulocytic differentiation of myeloid cells. NB4 cells were treated with ATRA at differentiation-inducing concentrations (0.1 or 1 μmol/L) and PDCD4 protein levels were examined by Western blot analysis. ATRA markedly up-regulated the expression of PDCD4 protein in a time-dependent manner, peaking at 72 h of ATRA treatment (Fig. 2A and B ). To determine whether induction of PDCD4 expression is regulated by transcriptional or posttranscriptional mechanisms, we analyzed PDCD4 mRNA levels in ATRA-treated NB4 cells by reverse transcription-PCR (RT-PCR) using specific primers. Induction of PDCD4 mRNA expression was detectable at 24 h and markedly increased at 48 h of ATRA treatment (Fig. 2C), suggesting that PDCD4 protein expression is regulated at the transcriptional level during differentiation of promyelocytic cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

ATRA induces marked PDCD4 expression in NB4 and HL60 cells but not in their differentiation-resistant counterparts. A. Cells were treated with differentiation-inducing doses (0.1 or 1 μmol/L) of ATRA and collected at the indicated time points. Equal amounts of total cell lysates were immunoblotted with anti-PDCD4 antibody. β-actin was used as a loading control. B. Bands representing PDCD4 protein expression in (A) were analyzed by densitometry. Results were expressed as the relative ratio of PDCD4 to β-actin. C. ATRA induced PDCD4 mRNA expression in NB4 cells. Following treatment with 1 μmol/L ATRA, RNA was extracted from NB4 cells at the indicated time points. PDCD4 mRNA expression was detected by RT-PCR using PDCD4-specific primers. D. Bands representing PDCD4 mRNA expression in (C) were analyzed using densitometry. Results were expressed as the relative ratio of PDCD4 to β-actin. E. PDCD4 protein expression is not induced by ATRA in NB4.R1 cells, which are unable to undergo granulocytic differentiation. Cells were treated with 1 μmol/L ATRA and harvested at the indicated time points. NB4 cells were used as a control. F. ATRA induces PDCD4 expression in HL60 but not HL60R cells. HL60 and HL60R cells were treated with 1 μmol/L ATRA and collected at the indicated time points for Western blot analysis of PDCD4 expression.

Because differentiation-defective myeloid cells provide a useful experimental model to study the molecular mechanisms involved in terminal cell differentiation, we compared the expression of PDCD4 in differentiation-sensitive (NB4 and HL60) and differentiation-resistant (NB4.R1 and HL60R) cells, which are unable to undergo ATRA-induced differentiation. In contrast to the differentiation-sensitive cells, treatment of NB4.R1 and HL60R cells with ATRA did not induce PDCD4 expression (Fig. 2D and E). ATRA also induced PDCD4 mRNA expression in HL60 cells but not in HL60R cells treated with ATRA for up to 96 h (data not shown). We also examined PDCD4 expression in the primary promyelocytic leukemia cells isolated from newly diagnosed acute promyelocytic leukemia patients. A significant up-regulation of PDCD4 protein by ATRA was observed in two different acute promyelocytic leukemia patient cells during ATRA-induced granulocytic differentiation, which was assessed by morphology and surface expression of myeloid (CD11b) and granulocytic (CD11c) differentiation markers (Fig. 3A-D ), supporting our hypothesis that PDCD4 expression is induced during granulocytic differentiation of myeloid cells. To determine whether ATRA induces PDCD4 in normal bone marrow progenitors, we treated CD34⁺ hematopoietic progenitor cells with ATRA for 72 h. We observed that these early progenitor cells could also up-regulate PDCD4 by ATRA treatment (Fig. 3E), suggesting that PDCD4 expression can be regulated in bone marrow microenvironment by retinoic acid.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

ATRA induces PDCD4 expression in primary human promyelocytic leukemia (AML-M3) and CD34⁺ normal bone marrow hematopoietic progenitor cells. The primary promyelocytic leukemia samples obtained from newly diagnosed acute promyelocytic leukemia (APL) patients and normal bone marrow progenitor cells were treated with ATRA at indicated time points. Primary acute promyelocytic leukemia cells were divided in two groups; the first group was lysed for Western blotting for the detection of PDCD4 expression and the rest of the cells were analyzed for induction of differentiation by examining CD11b and CD11c expression by fluorescence-activated cell sorting or stained for morphologic analysis. A. ATRA induced a significant PDCD4 expression during granulocytic differentiation of acute promyelocytic leukemia cell as indicated by appearance of granulocytic morphology including multilobular nucleus and increases cytoplasmic to nuclear ratio (B) and up-regulation of differentiation markers (CD11b and CD11c; C and D). E. ATRA also induced PDCD4 expression in normal CD34⁺ bone marrow progenitor cells examined at 72 h.

PDCD4 Expression Increases during Granulocytic but not Monocytic/Macrophagic Differentiation

We next investigated whether elevation of PDCD4 expression is specific to ATRA-induced granulocytic differentiation or also takes place during monocytic/macrophagic differentiation. NB4 cells were first treated with ATRA and other granulocytic differentiation-inducing agents, such as arsenic trioxide (27) and 1% DMSO (28). Granulocytic differentiation induced by ATRA, arsenic trioxide, or DMSO was associated with marked up-regulation of PDCD4 (Fig. 4A ). As expected, ATRA-induced PDCD4 expression was detectable at 48 h and peaked at 72 h of treatment. In contrast, arsenic trioxide at a differentiation-inducing dose (0.4 μmol/L) did not induce PDCD4 at early time points, but significant induction of PDCD4 was observed after 72 h of treatment (Fig. 4A). DMSO, on the other hand, induced strong PDCD4 expression at 48 h. These results showed that induction of PDCD4 expression generalizes to granulocytic differentiation induced by multiple inducers.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

PDCD4 expression is associated with granulocytic but not monocytic/macrophagic differentiation in AML cells. A. Granulocytic differentiation induced by 1 μmol/L ATRA, 0.4 μmol/L arsenic trioxide, and 1% DMSO was accompanied by increased PDCD4 expression in NB4 cells. Cells were collected at the indicated time points for Western blot analysis of PDCD4 expression. NT, nontreated control cells. B and C. Monocytic/macrophagic differentiation induced by 0.1 μmol/L PMA and 0.1 μmol/L 1,25-dihydroxyvitamin D3 did not up-regulate PDCD4 expression in NB4 cells. Equal amounts of total cell lysates were analyzed by Western blotting for PDCD4 protein levels. β-actin was used as a loading control. D. PMA induced morphologic changes associated with monocytic/macrophagic differentiation in NB4 leukemia cells. E. ATRA did not induce PDCD4 expression in THP1 myelomonocytic AML cells during monocytic differentiation. The cells were treated with 1 μmol/L ATRA for 24, 48, and 72 h and PDCD4 expression was detected by Western blotting.

We next treated NB4 cells with phorbol 12-myristate 13-acetate (PMA; refs. 29, 30) and 1,25-dihydroxyvitamin D3 (31, 32) agents that induce monocytic/macrophagic differentiation. Differentiation-inducing doses of PMA (0.1 μmol/L; Fig. 4B) and 1,25-dihydroxyvitamin D3 (0.1 μmol/L; Fig. 4C) did not induce PDCD4 expression. Higher doses (up to 1 μmol/L) of these compounds also failed to up-regulate PDCD4 expression in the cells (data not shown). Induction of monocytic/macrophagic differentiation by the two agents was confirmed by assessment of morphologic changes and adhesion to tissue culture flasks (Fig. 4D).

To confirm the association between PDCD4 induction and granulocytic differentiation, we investigated PDCD4 expression in THP1 myelomonocytic AML cells (33, 34), which undergo monocytic/macrophagic differentiation by ATRA. The cells were treated with ATRA (1 μmol/L) for 24, 48, and 72 h, and differentiation was assessed by morphologic characterization and adherence to plastic tissue culture flasks (data not shown; ref. 34). Although ATRA effectively induced monocytic differentiation in THP1 cells, it did not up-regulate PDCD4 expression (Fig. 4E). These findings provided further evidence of an association between induction of PDCD4 expression and granulocytic differentiation of AML cells.

ATRA Induces Nuclear Translocation of PDCD4 during Granulocytic Differentiation

The PDCD4 protein contains two basic NH₂- and COOH-terminal domains that may be nuclear localization signals. Intense nuclear staining of PDCD4 has been found in normal cells, such as fibroblasts, endothelial cells, and other cells of normal prostate, colon, breast, and lung tissues, compared with corresponding tumor cells (19). To elucidate the role of PDCD4 during granulocytic differentiation, we examined its subcellular localization in ATRA-responsive and differentiation-resistant leukemia cells. To this end, NB4 and NB4.R1 cells were treated with ATRA for 72 h and stained with anti-PDCD4 antibody. PDCD4 was located mainly in the cytoplasm in untreated NB4 and NB4.R1 cells, whereas marked nuclear translocation of PDCD4 was seen in ATRA-treated NB4 but not NB4.R1 cells (Fig. 5A and B ), suggesting that nuclear translocation of PDCD4 is strongly associated with granulocytic differentiation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

ATRA induces nuclear translocation of PDCD4 in differentiation-sensitive but not in differentiation-resistant cells. NB4 (A) and NB4.R1 cells (B) were treated with 1 μmol/L ATRA for 72 h, stained with rabbit anti-PDCD4 primary and FITC-labeled donkey anti-rabbit secondary antibodies, and analyzed by immunofluorescence microscopy. Nuclei were stained with 4′,6-diamidino-2-phenylindole (blue).

ATRA-Induced Reduction in PI3K/Akt Activity Is Associated with PDCD4 Induction during Granulocytic Differentiation

The PI3K/Akt (protein kinase B) pathway and its downstream component mammalian target of rapamycin (mTOR) constitutes a key signaling cascade that links diverse extracellular stimuli to cell proliferation, differentiation, and survival (35). Because the activity of the PI3K/Akt/mTOR pathway has been associated with increased proliferation and translation in cancer cells (36, 37), including AML (38, 39), and PDCD4 functions as a translational suppressor, we hypothesized that the PI3K/Akt/mTOR pathway negatively regulates PDCD4 expression and, thus, ATRA induces PDCD4 via inhibition of this pathway during granulocytic differentiation. We therefore sought to determine whether the PI3K/Akt/mTOR pathway is down-regulated during ATRA-induced granulocytic differentiation of NB4 cells. We first examined the phosphorylation status of Akt during ATRA treatment in NB4 cells. PI3K activity was reduced by ATRA, as indicated by a reduction in phosphorylated (p) Akt (Ser⁴⁷³) levels and the p-Akt/Akt ratio, reaching maximal inhibition after 48 to 72 h of treatment (Fig. 6A ). These findings suggest that the PI3K/Akt/mTOR pathway is inhibited during ATRA-induced granulocytic differentiation. The nadir p-Akt expression corresponded with the peak PDCD4 expression at 48 to 72 h of ATRA treatment (Fig. 2A), indicating an inverse association between activation of PI3K/Akt activity and PDCD4 expression.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

PI3K/Akt/mTOR signaling pathway represses PDCD4 expression. A. ATRA inhibits the PI3K/Akt/mTOR pathway during granulocytic differentiation in NB4 cells. NB4 cells were incubated with 1 μmol/L ATRA for up to 72 h or with 0.1 μmol/L ATRA for up to 48 h. Equal amounts of total cell lysates were analyzed by Western blotting for phosphorylated Akt (Ser⁴⁷³) and Akt. β-actin was used as a loading control. B. Inhibition of the PI3K/Akt/mTOR pathway enhances ATRA-induced PDCD4 expression in NB4 cells. The cells were treated with PI3K inhibitor (20 μmol/L LY294002) or mTOR inhibitor (20 nmol/L rapamycin) for 72 h, with or without ATRA. Equal amounts of total cell lysates were analyzed by Western blotting for PDCD4, p-Akt, and β-actin. C. Inhibition of PI3K pathway by LY294002 inhibits p-Akt. NB4 cells were treated with LY294002 in the presence or absence of ATRA for 48 h. p-Akt was detected by Western blotting. D. Inhibitors of the PI3K/Akt/mTOR pathway induce PDCD4 mRNA expression in NB4 cells. Cells were treated with PI3K inhibitors (200 nmol/L wortmannin, 20 μmol/L LY294002) or 20 nmol/L rapamycin in the presence or absence of 1 μmol/L ATRA. The cells were collected and total cellular RNA was extracted to detect PDCD4 mRNA expression by RT-PCR using PDCD4-specific primers. ATRA markedly induced PDCD4 mRNA expression after 24 h of treatment. Bands representing PDCD4 mRNA expression in the gel were analyzed by densitometry. Results were expressed as relative ratios of PDCD4 to β-actin mRNA.

PI3K/Akt/mTOR Signaling Pathway Suppresses PDCD4 Expression in Leukemia Cells

Because ATRA down-regulates activity of PI3K/Akt survival pathway under conditions in which it up-regulates PDCD4 expression, we sought to determine whether the PI3K/Akt/mTOR pathway plays a role in the regulation of PDCD4 by ATRA. To that end, we blocked PI3K/Akt/mTOR activity using specific inhibitors of PI3K (LY294002 and wortmannin; refs. 38, 39) and mTOR (rapamycin; ref. 40) and analyzed PDCD4 levels in the presence and absence of ATRA in NB4 cells by Western blotting. As expected, ATRA enhanced PDCD4 expression compared with untreated control cells (Fig. 6B), and treatment of cells with LY294002 or rapamycin enhanced the of PDCD4 expression alone and produced significant up-regulation of PDCD4 expression (Fig. 6B). To confirm the inhibition of PI3K pathway, we examined p-Akt levels in the cells and found that treatment with LY294002 markedly reduced p-Akt levels (Fig. 6C). These observations suggest that the PI3K/Akt/mTOR pathway represses PDCD4 expression in leukemia progenitors. The inhibition of this pathway by ATRA and/or by the pathway-specific inhibitors seems to release suppression of PDCD4.

To determine whether the induction of PDCD4 expression is mediated at the transcriptional or posttranslational level, we assessed PDCD4 mRNA expression after treatment with ATRA and/or the inhibitors. LY294002, wortmannin, and rapamycin up-regulated PDCD4 mRNA compared with untreated NB4 cells (Fig. 6D). Thus, the ATRA and PI3K/Akt/mTOR pathway inhibitors seem to regulate PDCD4 at the level of mRNA expression either by increasing transcription or by inhibiting mRNA degradation or both in AML cells.

PDCD4 Induction Is Important in Granulocytic Differentiation of AML Cells

To elucidate the role of PDCD4 in ATRA-induced granulocytic differentiation of myeloid progenitor cells, we knocked down PDCD4 expression using PDCD4-specific siRNA. NB4 cells transfected with PDCD4 or nonsilencing (control) siRNA, or left untreated, were treated with 1 μmol/L ATRA for 72 h, followed by assessment of the differentiation markers CD11b and CD11c by flow cytometry. Under these conditions, we consistently reached a transfection efficiency of ∼60%, without a significant reduction in cell viability (data not shown). For CD11b and CD11c, analyses were done including all cells with or without transfection. As expected, ATRA treatment of cells transfected with transfection reagent only resulted in no inhibition of CD11b and CD11c expression in all cells (without excluding untransfected cells). In contrast, concomitant treatment with ATRA and PDCD4 siRNA resulted in a significant inhibition of granulocytic differentiation in the cells (P < 0.05), as indicated by reduced expression of CD11b and CD11c granulocytic differentiation markers (Fig. 7A-D ) and by morphology (Fig. 7E), compared with those transfected with control nonsilencing siRNA. These results suggest that PDCD4 expression contributes to granulocytic differentiation in AML cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

PDCD4 is involved in granulocytic differentiation of promyelocytic leukemia cells. A and B. NB4 cells were transfected with transfection reagent (TR) alone, PDCD4 siRNA, or nonsilencing control siRNA, followed by ATRA treatment for 72 h. Induction of granulocytic differentiation was determined by flow cytometric analysis of surface CD11b and CD11c expression using all cells (transfected and untransfected). Data shown as percent reduction in the number of cells undergoing differentiation in transfection reagent–, PDCD4 siRNA–, control siRNA–transfected, and ATRA-treated cells compared with untransfected control cells treated with ATRA. C. NB4 cells after transfection with PDCD4 and control siRNA were stained for morphologic analysis of differentiation. Granulocytic phenotype including multilobular nucleus was observed in the majority of control siRNA–treated cells. In contrast, fewer differentiated cells were observed in PDCD4 siRNA–treated cells.

PDCD4 Mediates Expression of ATRA-Regulated Important Cellular Proteins

Although indirect transcriptional targets of PDCD4 have been found, most of the direct targets of PDCD4 have not yet been identified (14). To identify potential downstream targets of PDCD4 that may be functionally significant in the mechanism by which ATRA induces granulocytic differentiation, we examined the expression of several important cellular proteins known to be regulated by ATRA in NB4 cells and important in ATRA-induced growth inhibition and granulocytic differentiation. These proteins include c-jun (7), c-myc, the cyclin-dependent kinase inhibitors p21^Waf1/Cip1 (41) and p27^Kip1(42), WT1 (43), tissue transglutaminase (TG2; refs. 33, 44), and the novel translational inhibitor DAP5 (10). DAP5 inhibits cap-dependent and cap-independent mRNA translation by competing with eIF4G and sequestering eIF4A and eIF3 and is essential for terminal differentiation (10, 45, 46). WT1 is aberrantly overexpressed in majority of AML blasts isolated from patients, a bad prognostic factor, and inhibited by ATRA during differentiation.

NB4 cells treated with PDCD4 or nonsilencing siRNA, or left untreated, were incubated with or without ATRA for 72 h, followed by Western blot and RT-PCR analyses. PDCD4 protein levels were markedly down-regulated within 48 h of transfection with PDCD4 siRNA (Fig. 8A ). In addition, down-regulation in the levels of p27^Kip1 and DAP5, up-regulation of c-myc and WT1 protein, and no change in p21^Waf1/Cip1 levels were observed in the same samples (Fig. 8A-C). ATRA could not induce expression of PDCD4 protein when PDCD4 siRNA was used, as observed after a 48-h treatment (Fig. 8C, lanes 3 and 4). However, at higher doses of PDCD4 siRNA, PDCD4 expression was markedly reduced even when the cells were stimulated with ATRA. The knockdown of PDCD4 expression was accompanied by a reduction in the ATRA-induced expression of DAP5 and a block in ATRA-induced down-regulation of WT1; in contrast, the ATRA modulation of c-jun, p21^Waf1/Cip1, and TG2 expression remained unchanged (Fig. 8D). RT-PCR analysis showed that c-myc, DAP5, p27^Kip1WT1, and mRNA levels were not altered by siRNA-induced down-regulation of PDCD4 (Fig. 8E), suggesting that PDCD4 posttrancriptionally regulates ATRA-modulated expression of these cellular proteins.

• Download figure
• Open in new tab
• Download powerpoint

Figure 8.

PDCD4 regulates expression of key cellular proteins. To identify the role of PDCD4 in regulation of proteins, we examined proteins that are regulated by ATRA and involved in growth arrest and differentiation in myeloid cells. A. Cell cycle and cyclin-dependent kinase inhibitor protein p27^Kip1 is regulated by PDCD4. PDCD4 expression was knocked down PDCD4 by siRNA in NB4 cells and analyzed at 48 h by Western blotting for p27^Kip1 expression. Inhibition of PDCD4 expression by PDCD4 siRNA resulted in down-regulation of PDCD4 and p27^Kip1 protein expression, but not housekeeping protein β-actin, suggesting that PDCD4 is required for the expression of p27^Kip1. Right, relative inhibition of PDCD4 by densitometric analysis of the Western blot after normalizing to actin expression. B. PDCD4 inhibition leads to induction of c-myc and reduction in DAP5 protein expression but no change in p21^Waf1/Cip1 cyclin-dependent kinase inhibitor levels. C. WT1 expression is up-regulated by inhibition of PDCD4 by siRNA. D. The expression of ATRA-modulated proteins, including DAP5, TG2, WT1, and p21^Waf1/Cip1, was determined in NB4 cells after siRNA-mediated knockdown of PDCD4 compared with control cells. Cells were treated with ATRA after 48 h of siRNA transfection. β-actin was used as a loading control. PDCD4 inhibition by siRNA prevented ATRA-mediated up-regulation of DAP5 and down-regulation of WT1 proteins. However, PDCD4 deficiency did not alter ATRA-induced levels of p21 and TG2. E. Knockdown of PDCD4 by siRNA does not result in alteration in mRNA levels of DAP5, c-myc, and WT1 detected by semiquantitative RT-PCR analysis.

Discussion

The results of the present study show that the PDCD4 is involved in granulocytic differentiation induced by ATRA. ATRA-induced PDCD4 expression is mediated by inhibition of PI3K/Akt/mTOR survival pathway that constitutively represses PDCD4 expression in AML cells. This study is the first to implicate PDCD4 in myeloid cell differentiation and reveals a novel mechanism of ATRA-induced granulocytic differentiation of myeloid cells (Fig. 9 ).

• Download figure
• Open in new tab
• Download powerpoint

Figure 9.

Model for the role of PDCD4 in mediating ATRA-induced granulocytic differentiation. The PI3K/Akt/mTOR signaling pathway negatively regulates PDCD4 expression. ATRA inhibits this pathway and enhances PDCD4 expression in myeloid leukemia cells, which in turn leads to granulocytic differentiation. PDCD4 regulates ATRA-modulated proteins, such as p27^Kip1, c-myc, WT1, and DAP5, which are important to induction of granulocytic differentiation. DAP5, a novel translational suppressor, is an important mediator of granulocytic differentiation, and lack of DAP5 expression prevents ATRA-induced differentiation, leading to resistance to ATRA (10, 47).

Recent studies suggested that terminal cell differentiation is associated with the inhibition of proliferation and repression of mRNA translation, leading to a decreased rate of protein synthesis (24, 25). We previously reported significant down-regulation of the eukaryotic initiation factors, including eIF4A, eIF4G, eIF2, and eIF5, and up-regulation of a translation initiation repressor, DAP5/p97, during the granulocytic differentiation induced by ATRA (10). Induction of PDCD4 translational repressor is in agreement with previous studies and supports the hypothesis that ATRA inhibits translation initiation, which is a tightly regulated step of translation, and contributes to the posttranscriptional regulation of genes during myeloid cell differentiation.

Two lines of evidence obtained in this study suggest that the expression of PDCD4 is important to granulocytic differentiation of myeloid cells and resistance to retinoic acid-induced differentiation. First, in contrast to the differentiation-sensitive NB4 and HL60 cells, the differentiation-resistant NB4.R1 and HL60R cells did not show up-regulation and nuclear translocation of PDCD4 in response to ATRA (Fig. 2D and E). Second, down-regulation of PDCD4 reduced the number of cells undergoing ATRA-induced granulocytic differentiation (Fig. 7A and B). Our findings are in agreement with a recent study that showed that PDCD4 is highly expressed in normal tissues with predominant nuclear localization, but its nuclear localization is reduced in solid tumors (19, 47), supporting the hypothesis that lack of nuclear localization of PDCD4 may play a role in leukemogenesis/carcinogenesis (19). It is also possible that PDCD4 may interact with promyelocytic leukemia, which is also localized to nuclear domains and shown to be involved in translational control (48–52). Many tumor cell types are undifferentiated or poorly differentiated; deficiency of PDCD4 expression seems to correlate with undifferentiated phenotype and may contribute to the differentiation block seen in AML cells.

The PI3K/Akt/mTOR pathway is overactivated in >80% of AML patients and plays an important role in regulating global and specific mRNA translation (35, 37). Activation of PI3K has also been linked with tumorigenesis, metastasis, and resistance to chemotherapy (53). Activation of PI3K/Akt by growth factors or mitogens leads to phosphorylation of mTOR, subsequent phosphorylation of p70 S6 kinase and eIF4E-binding protein 1, and activation of translation initiation factor eIF4E, resulting in an increase in mRNA translation (35, 36). The present study shows for the first time that the PI3K/Akt/mTOR pathway represses expression of PDCD4 tumor suppressor protein at the transcriptional level, revealing a novel mechanism of PDCD4 regulation and inactivation in AML. In addition, a recent report suggested that Akt phosphorylates and inactivates PDCD4 tumor suppressor function as an inhibitor of AP-1-mediated transcription.(54). Because this pathway is crucial to promoting cell growth, survival, and antiapoptotic responses in AML cells (38, 39), our findings also shed light on mechanism of antileukemic action of rapamycin, which induces marked PDCD4 expression in AML cells (39). Inhibitors of mTOR, such as rapamycin analogues (CCI-779 and RAD001) have shown promising results in AML, suggesting that targeting translational pathways is a viable treatment strategy in AML (39, 55, 56). Inhibitors of mTOR prevent cyclin-dependent kinase activation, inhibit Rb protein phosphorylation, and down-regulate cyclin D1, all of which may contribute to G₁ phase arrest (55–59). Therefore, induction of PDCD4 by inhibition of mTOR by rapamycin analogues provides a novel rationale for this treatment in AML patients.

The present study shows that ATRA-induced granulocytic differentiation is associated with the inhibition of PI3K/Akt activity (Fig. 6). This finding is in agreement with previous reports that ATRA down-regulates PI3K activity (60–62). The PI3K pathway has been linked not only with ATRA resistance but also with ATRA-induced differentiation in promyelocytic leukemia cells (63–65). We found that ATRA-resistant cells are unable to enhance PDCD4 expression, thus suggesting that this pathway may contribute to resistance to ATRA-induced differentiation through down-regulation of PDCD4. In fact, inactivation or reduced expression of PDCD4 has been implicated in drug resistance in breast cancer cells (66), and knocking down of PDCD4 prevented ATRA-induced differentiation (Fig. 7), supporting this hypothesis.

Although the expression of several proteins, among them ornithine decarboxylase, cyclin-dependent kinase 4 (18), and carbonic anhydrase type II (67), has been reported to be down-regulated by PDCD4 expression, the downstream targets of PDCD4 have not yet been identified. SiRNA-mediated knockdown of PDCD4 helped us to identify important cellular proteins as downstream targets of PDCD4, including c-myc, p27, DAP5, and WT1. ATRA down-regulates c-myc and WT1 and p27 up-regulates CDC-inhibitor during ATRA-induced differentiation of NB4 and HL60 cells. Attenuation by PDCD4 of ATRA up-regulation of DAP5 and down-regulation of WT1 and c-myc occurred at the level of protein but not RNA expression, suggesting that PDCD4 regulates expression of these proteins involved in granulocytic differentiation (Figs. 8 and 9). DAP5 is an important mediator of differentiation, and lack of DAP5 expression prevents ATRA-induced differentiation and causes resistance to ATRA (10, 45). Because siRNA to PDCD4 attenuates down-regulation of WT1, WT1 may be a direct translational target of PDCD4, a possibility that requires further testing.

Overall, the present results suggest that PDCD4-induced inhibition of translation initiation may play a role in controlling hyperactivated translation in cancer cells and may lead to growth inhibition and differentiation in response to the granulocytic differentiation inducers. A better understanding of this posttranscriptional mechanism may help identify targets for new differentiation therapies for cancer.

Materials and Methods

Cell Lines, Culture Conditions, and Reagents

The human acute promyelocytic cell line NB4 (M3-type AML based on French-American-British classification) was obtained from Dr. Michael Andreeff (The University of Texas M.D. Anderson Cancer Center, Houston, TX) with permission of Dr. Michael Lanotte. The NB4.R1 cell line, an ATRA-resistant derivative of NB4, was generously provided by Dr. Ethan Dmitrovsky (Norris Cotton Cancer Center, Dartmouth Medical School, Hanover, NH; ref. 68). HL60 (M2-type myeloblastic AML) and THP1 (M5-type myelomonocytic AML) myeloid leukemia cells were purchased from the American Type Culture Collection (Manassas, VA). The granulocytic differentiation-resistant HL60R cell line, an ATRA-resistant subline of HL60, was provided by Dr. Steven Collins (Fred Hutchinson Cancer Center, Seattle, WA; ref. 69). Primary human promyelocytic (AML-M3) cells isolated from newly diagnosed acute promyelocytic leukemia patients were provided by the leukemia tissue bank through an Institutional Review Board protocol. A highly purified population of CD34⁺ primary human hematopoietic progenitor cells was purchased from Cambrex (Cambrex Bio Science, Walkersville). The cells were grown in RPMI 1640 (Life Technologies, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum at 37°C under 5% CO₂ in a humidified incubator. ATRA, arsenic trioxide, 1,25-dihydroxyvitamin D3, PMA, and DMSO were purchased from Sigma (St. Louis, MO). For primary cells, 20% fetal bovine serum was used. The PI3K-specific inhibitors LY294002 and wortmannin and the mTOR inhibitor rapamycin were purchased from Calbiochem (La Jolla, CA).

Cell Treatments and Growth Assays

Cells were seeded at 1 × 10⁵/mL in RPMI medium in six-well tissue culture plates (Costar, Cambridge, MA). After dilution with saline from a 10 mmol/L stock, ATRA (Sigma) at a final concentration of 0.1 or 1 μmol/L was incubated with the cells for the indicated time points. 1,25-Vitamin D3, PMA, or arsenic trioxide (dissolved in 5 N NaOH) was added at the indicated concentrations. Cell viability was determined by trypan blue (Sigma) exclusion test.

Evaluation of Cellular Differentiation

Granulocytic differentiation of the cells was determined by examining CD11b and CD11c expression, morphologic changes, and reorganization of promyelocytic leukemia nuclear bodies. For the CD11b and CD11c analysis, cells were collected after 3 to 5 days of treatment with differentiation-inducing agents and washed with PBS. Cells (5 × 10⁵) in 100 μL of PBS were incubated for 30 min with FITC-conjugated anti-CD11b antibody (1:200), phycoerythrin-conjugated anti-CD11b, or FITC-conjugated immunoglobulin G1 isotype control (Becton Dickinson, Franklin Lakes, NJ) at room temperature in the dark, as previously described (34). The cells were then washed again to remove unbound antibodies and resuspended in 500 μL of PBS. The percentage of CD11b⁺ and CD11c⁺ cells was determined by fluorescence-activated cell sorting analysis (Flow Cytometry and Cellular Imaging Facility, M. D. Anderson Cancer Center). Morphology was evaluated by May-Grünwald-Giemsa staining. Briefly, cytospin preparations of 2 × 10⁵ cells were air-dried, incubated sequentially in pure May-Grünwald solution (Sigma) for 5 min and 50% May-Grünwald/water solution for 10 min, and washed with water. The slides were then incubated in a 20% Giemsa (Sigma)/water solution for 20 min, washed again with water, air-dried, and examined under a Nikon microscope.

Immunofluorescence Staining

Cells were collected from control and ATRA-treated cultures and washed twice with ice-cold PBS (pH 7.4). Cytospin preparations of cells were fixed with methanol for 10 min at room temperature, fixed in cold acetone for 2 min at −20°C, and air-dried. The slides were then washed with PBS, blocked with 1% bovine serum albumin solution in PBS for 60 min, and incubated with either anti–promyelocytic leukemia antibody (Santa Cruz Biotechnology, Santa Cruz, CA) diluted 1:50 in PBS containing 1% bovine serum albumin or anti-PDCD4 antibody (1:200; Santa Cruz Biotechnology) for 45 min at room temperature. After washing with PBS containing 1% bovine serum albumin, the slides were incubated with FITC-labeled goat anti-mouse immunoglobulin G (1:100; Sigma) for 45 min at room temperature. The cells were then incubated with blocking buffer containing 1 μg/mL 4′,6-diamidino-2-phenylindole for 5 min at room temperature and washed thrice in PBS. Coverslips were mounted on the slides using a ProLong antifade kit (Molecular Probes, Carlsbad, CA) to retard fading and analyzed under a Nikon fluorescence microscope.

Western Blot Analysis

Following treatments with differentiation inducers, cells were collected and centrifuged, and whole-cell lysates were prepared using a lysis buffer. Total protein concentration was determined using a detergent-compatible protein assay kit (Bio-Rad, Hercules, CA). For the inhibition experiments, cells were preincubated with LY294002, wortmannin, and rapamycin for 1 to 4 h before treatment with ATRA for the indicated time periods. Aliquots containing 30 μg of total protein from each sample were subjected to SDS-PAGE and electrotransferred to nitrocellulose membranes. The membranes were blocked with 5% dry milk in TBST [100 mmol/L Tris-HCl (pH 8.0), 150 mmol/L NaCl, and 0.05% Tween 20], probed with primary antibodies diluted in TBST containing 5% dry milk, and incubated at 4°C overnight. Primary antibodies against Akt, p-Akt (Ser⁴⁷³), and DAP5 were obtained from Cell Signaling Technology (Beverly, MA); antibodies against p27^Kip1, p21^Waf1/Cip1, WT1, c-myc, and c-jun were obtained from Santa Cruz Biotechnology. Transglutaminase 2 (TG2) antibody was purchased from Neomarkers (Fremont, CA).

Serum containing anti-PDCD4 peptide antibodies was diluted 1:10,000 in TBST (50). After washing, the membranes were incubated with horseradish peroxidase–conjugated antirabbit secondary antibody (Amersham Life Science, Cleveland, OH). Mouse anti–β-actin and donkey anti-mouse secondary antibodies were purchased from Sigma to analyze β-actin expression for equal loading. The bands were visualized by enhanced chemiluminescence (KPL, Gaithersburg, MD). Images were scanned and quantitated using a densitometer with the Alpha Imager application program (Alpha Innotech, San Leandro, CA). All experiments were independently repeated at least thrice.

RNA Isolation and RT-PCR Analysis. Cells were seeded in six-well plates (1 × 10⁶/mL) and treated with ATRA at a final concentration of 1 μmol/L or with the specific inhibitors of PI3K/Akt/mTOR at the indicated concentrations. The cells were collected at various time points and total cellular RNA was isolated with TRIzol reagent (Life Technologies). cDNA was obtained from 5 μg of total RNA using a Superscript II RT kit (Life Technologies) as previously described (70). Briefly, 5 μL of the total 20 μL of reverse-transcribed product were used for PCR in 1× PCR buffer containing 1.5 mmol/L MgCl₂, 250 μmol/L deoxynucleotide triphosphates, 0.5 units of Taq polymerase (Life Technologies), and 100 ng of primers for PDCD4 (primer I, 5′-ATGGATGTAGAAAATGAGCAG-3′; primer II, 5′-TTAAAGTCTTCTCAAATGCCC-3′), DAP5 (primer I, 5′-CAGCAGTGAGTCGGAGCTCTATGG-3′; primer II, 5′-GTGGAGAGTGCGATTGCAGAAG-3′), c-myc (primer I, 5′-TCAAGAGGTGCCACGTCTCC-3′, primer II, 5′-TCTTGGCAGCAGGATAGTCCTT-3′) and WT1 (71) or β-actin (Sigma-Genosys, Houston, TX). The following programs were used for PCR amplification of PDCD4: 1 cycle at 94°C for 2 min, 25 to 35 cycles, denaturation at 94°C for 1 min, annealing at 55°C to 65°C for 1 min, and extension at 72°C for 1 min. A cycle of 72°C for 7 min was added to complete the reaction. The reaction products were analyzed on 2% agarose gels containing ethidium bromide and cDNA synthesis was verified by detection of the β-actin transcript.

Targeted Down-Regulation of PDCD4 by siRNA

Exponentially growing, untreated NB4 cells were harvested and used for siRNA transfection. Separate aliquots of 2 × 10⁶ cells were transfected with double-stranded siRNA targeting PDCD4 mRNA or control nonsilencing siRNA (all from Qiagen, Valencia, CA) using the Amaxa Nucleofector electroporation technique (Amaxa, Gaithersburg, MD) according to the manufacturer's guidelines. The siRNA sequence (5′-AAGGUGGCUGGAACAUCUAUU-3′) targeting PDCD4 was designed using siRNA-designing software (Qiagen). Untransfected cells, control siRNA–transfected cells, and transfection reagent alone were used as negative controls. Forty-eight hours after transfection with siRNA, fresh medium with or without 1 μmol/L ATRA was added to the cell cultures. After treatment, the cells were harvested for Western blot analysis of PDCD4 protein expression or fluorescence-activated cell sorting analysis of CD11b expression.

Statistical Analysis

Data were expressed as mean ± SD of three or more independent experiments. Statistical analysis was done using two-tailed Student's t test for paired data. P < 0.05 was considered statistically significant.