Ornithine Decarboxylase Inhibition by α-Difluoromethylornithine Activates Opposing Signaling Pathways via Phosphorylation of Both Akt/Protein Kinase B and p27^Kip1 in Neuroblastoma Free

Neuroblastoma (NB) is the most common extracranial childhood tumor and is derived from the neural crest cells of the sympathetic nervous system. NB patients diagnosed under the age of 1 year often experience complete regression of tumors, whereas older patients often struggle with more advanced stages of the disease, malignant progression, and poor prognosis partly due to the emergence of multidrug resistance (1–3). The mechanisms that control the progression or regression of NB tumors have yet to be elucidated and thus present a difficult challenge for the treatment of this disease. Therefore, there is a need for alternative therapeutic strategies for the treatment of NB.

We and others previously showed that the treatment of MYCN-amplified NB cells with α-difluoromethylornithine (DFMO), an irreversible suicide inhibitor of ornithine decarboxylase (ODC), depleted polyamine pools and caused growth inhibition associated with p27^Kip1/retinoblastoma (Rb)–coupled G₁ cell cycle arrest in the absence of apoptosis (4, 5). DFMO-induced polyamine depletion, alone or in combination with S-adenosylmethionine decarboxylase (AdoMetDC) inhibitor SAM486A, effectively increased the expression of cyclin-dependent kinase (CDK) inhibitor p27^Kip1, inhibited the hyperphosphorylation of Rb protein, and down-regulated MYCN. Other studies have shown that DFMO treatment of chondrocytes and intestinal epithelial cells (IEC-6) induced cell cycle arrest in the absence of apoptosis and activated the antiapoptotic protein Akt/protein kinase B (PKB; refs. 6, 7). Akt/PKB is also critical in the regulation of cell cycle progression by modulating the phosphorylation state and stability of p27^Kip1 (8–13).

In the present study, we continued to investigate the role of p27^Kip1 in NB by analyzing the phosphorylation patterns of p27^Kip1 in response to polyamine inhibition by DFMO. In addition, we focused on the protein Akt/PKB and determined the effect of DFMO on the phosphatidylinositol 3-kinase (PI3K)/Akt signaling pathway in the presence/absence of PI3K inhibitor LY294002 and Akt/PKB inhibitor IV. The paradoxical effect of DFMO by activation of two separate pathways, one activating Akt/PKB via PI3K and one inducing cell cycle arrest by regulating p27^Kip1, may explain the moderate efficacy of DFMO-based monotherapies in clinical trials. We provide further evidence that combination therapies may prove to be essential for the development of more effective therapeutic strategies for the treatment of NB.

Chemicals, reagents, and antibodies. The ODC inhibitor DFMO (14) and the AdoMetDC inhibitor SAM486A (CGP48644; refs. 15, 16) were provided by Dr. Patrick Woster (Wayne State University, Detroit, MI) and Novartis, respectively. LY294002, Akt/PKB inhibitor IV, putrescine, spermidine, spermine, aminoguanidine, trichloroacetic acid (TCA), acetic acid, sulforhodamine B (SRB), and mouse monoclonal β-actin antibody (A5316) were obtained from Sigma Chemical Co. Rabbit polyclonal phospho-Akt/PKB (Ser⁴⁷³), rabbit polyclonal phospho-glycogen synthase kinase-3β (GSK-3β; Ser⁹), rabbit polyclonal phospho-FKHR (Ser²⁵⁶), rabbit polyclonal phospho-PTEN (Ser³⁸⁰), and rabbit polyclonal phospho-PDK1 (Ser²⁴¹) were from Cell Signaling Technology, Inc. Rabbit polyclonal p27^Kip1, rabbit polyclonal phospho-p27^Kip1 (Ser¹⁰), rabbit polyclonal phospho-p27^Kip1 (Thr¹⁸⁷), and rabbit polyclonal agarose-conjugated p27^Kip1 were from Santa Cruz Biotechnology, Inc. Rabbit polyclonal phospho-p27^Kip1 (Thr¹⁹⁸) was purchased from R&D Systems. Secondary anti-mouse and anti-rabbit antibodies coupled to horseradish peroxidase (HRP) were from Amersham Biosciences. Protein assay dye reagent was from Bio-Rad Laboratories.

Cell lines and treatment of cultured cells. The human NB cell line LAN-1 (17) was maintained in RPMI 1640 (Biosource) containing 10% heat-inactivated fetal bovine serum (Invitrogen), penicillin (100 IU/mL), and streptomycin (100 μg/mL). If cells were treated with polyamines, 1 mmol/L aminoguanidine was included as an inhibitor of serum polyamine oxidation. Cells were seeded 3 to 4 h before treatment with 5 mmol/L DFMO and analyzed after 4 d. For polyamine studies, 10 μmol/L of putrescine, spermidine, or spermine were added together with DFMO. For PI3K and Akt/PKB inhibitor studies, 20 μmol/L of LY294002 or 500 nmol/L of Akt/PKB inhibitor IV were added after 3 d of DFMO treatment.

SRB assay. The SRB colorimetric assay was used to determine cell proliferation following the protocol outlined in ref. 18. Briefly, cells were seeded at a density of 750 per well on a transparent, flat-bottom, 96-well plate and allowed to settle overnight. At the initiation of each experiment (t = 0), and after drug treatments, 100 μL of 10% (w/v) TCA were added to each well, incubated for 1 h at 4°C, washed with deionized water, and dried at room temperature. One hundred microliters of 0.057% (w/v) SRB solution were added to each well, incubated for 30 min at room temperature, rinsed four times with 1% (v/v) acetic acid, and allowed to dry at room temperature. Finally, 200 μL of 10 mmol/L Tris base solution (pH 10.5) were added to each well, and after shaking for 5 min at room temperature, the absorbance was measured at 510 nm in a microplate reader. The absorbance at t = 0 was compared with the absorbance at the end of the experiment to determine cell growth in treated cells compared with control cells.

3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt assay. The CellTiter 96 AQueous One Solution Cell Proliferation Assay is a colorimetric method for determining the number of viable cells in proliferation or cytotoxicity assays (Promega). The CellTiter 96 AQueous One Solution Reagent contains a tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS)] and was used to determine the viability of cells treated with DFMO alone and in combination with PI3K inhibitor LY294002 or Akt/PKB inhibitor IV and compared with untreated control cells. Briefly, cells were seeded at a density of 750 per well on a transparent, flat-bottom, 96-well plate in a total volume of 100 μL. After cell treatments, 20 μL of CellTiter 96 AQueous One Solution Reagent were added to wells and incubated for 1 to 4 h at 37°C. The absorbance was measured at 492 nm using a microplate reader.

Western blot analysis. Cell lysates were prepared in radioimmunoprecipitation assay (RIPA) buffer [20 mmol/L Tris-HCl (pH 7.5), 0.1% (w/v) sodium lauryl sulfate, 0.5% (w/v) sodium deoxycholate, 135 mmol/L NaCl, 1% (v/v) Triton X-100, 10% (v/v) glycerol, 2 mmol/L EDTA, supplemented with Complete protease inhibitor cocktail (Roche Molecular Biochemicals), and phosphatase inhibitors, 20 mmol/L sodium fluoride (NaF), and 0.27 mmol/L sodium vanadate (Na₃VO₄)]. Western blot analysis was performed as previously described (5). The total protein concentration was determined using the Bradford dye reagent protein assay (Bio-Rad Laboratories). Cell lysates in SDS sample buffer were boiled for 5 min and equal amounts of total protein were resolved by 10% SDS-PAGE and electrotransferred onto polyvinylidene difluoride Immobilon-P membrane (Millipore). Blots were incubated with phospho-Akt/PKB (Ser⁴⁷³; 1:1,000), phospho-GSK-3β (Ser⁹; 1:1,000), phospho-FKHR (Ser²⁵⁶; 1:1,000), phospho-PTEN (Ser³⁸⁰; 1:1,000), phospho-PDK1 (Ser²⁴¹; 1:1,000), β-actin (1:5,000), p27^Kip1 (1:250), phospho-p27^Kip1 (Ser¹⁰; 1:250), phospho-p27^Kip1 (Thr¹⁸⁷; 1:200), or phospho-p27^Kip1 (Thr¹⁹⁸; 1:200) with gentle agitation, washed with deionized water, and then incubated with secondary anti-mouse (1:5,000) or anti-rabbit (1:5,000) antibodies coupled to HRP. After washing the blot with deionized water, proteins were detected using the enhanced chemiluminescence (ECL) Plus reagents (Amersham Biosciences) and Kodak BioMax XAR film (Fisher Scientific). Membranes were stripped at 50°C for 30 min with ECL stripping buffer [62.5 mmol/L Tris-HCl (pH 6.7), 2% SDS, 100 mmol/L 2-mercaptoethanol] and sequentially probed. Bands were quantified using a Bio-Rad Multi Imager and Quantity One Quantitation Software (Bio-Rad Laboratories).

Immunoprecipitation. Cell lysates were prepared in RIPA buffer plus inhibitors, and total protein concentration was determined as above. Equal amounts of total protein (1,000 and 250 μg in experiments of Fig. 5A and B, respectively) for each treatment sample were mixed with 1 μg p27^Kip1 agarose-conjugated antibody (Santa Cruz Biotechnology) and incubated in the cold room overnight with gentle agitation. Agarose-protein immune complexes were washed with RIPA buffer, and proteins were eluted by adding SDS sample buffer and boiling for 10 min. The samples were then briefly centrifuged to separate agarose beads from immunoprecipitated proteins. Samples were analyzed by Western blot and quantification was performed as indicated above.

Polyamine depletion–induced phosphorylation of Akt/PKB and GSK-3β. To investigate the effects of polyamine depletion on the cell signaling response in NB cells, we treated LAN-1 cells with DFMO and/or SAM486A and determined the response of protein Akt/PKB as well as Akt/PKB-associated proteins that either regulate Akt/PKB activity (PTEN and PDK1) or are themselves regulated by Akt/PKB (GSK-3β and FKHR; refs. 19–21). DFMO treatment, either alone or in combination with SAM486A, induced a rapid increase in Akt/PKB phosphorylation at Ser⁴⁷³, whereas the total protein levels of Akt/PKB remained unchanged (Fig. 1A). Identical treatment conditions did not induce the phosphorylation of Akt/PKB at Thr³⁰⁸. Furthermore, the phosphorylation status of PDK1, FKHR, or PTEN remained unchanged (Fig. 1A). Interestingly, identical cell treatments also increased GSK-3β phosphorylation at Ser⁹ (Fig. 1A). SAM486A alone had no detectable effect on Akt/PKB protein expression or Akt/PKB phosphorylation at Ser⁴⁷³ and Thr³⁰⁸. To verify that the observed effects are specific and due to DFMO-mediated polyamine depletion, polyamines were added exogenously to DFMO-treated cells. We found that the phosphorylation of Akt/PKB at Ser⁴⁷³ and GSK-3β at Ser⁹ was comparably attenuated by the addition of putrescine, spermidine, or spermine (Fig. 1B).

To further investigate the DFMO-dependent regulation of Akt/PKB, we next examined the effect of PI3K inhibitor LY294002 and Akt/PKB inhibitor IV on the phosphorylation state of Akt/PKB and GSK-3β. Whereas LY294002 inhibits PI3K and affects the signaling to numerous downstream proteins including Akt/PKB, Akt/PKB inhibitor IV inhibits a kinase directly upstream of Akt/PKB. We found that cell treatments with either LY294002 or Akt/PKB inhibitor IV reduced DFMO-induced phosphorylation of Akt/PKB and, to a lesser degree, of GSK-3β (Fig. 2). These results show that DFMO induces the phosphorylation of Akt/PKB at Ser⁴⁷³ via the PI3K pathway, which in turn phosphorylates and thus inactivates GSK-3β.

The effects of DFMO and Akt/PKB inhibitors on proliferation and viability of NB cells. To determine the role of Akt/PKB in DFMO-induced cell cycle arrest in LAN-1 NB cells, the dose-dependent effects of inhibitors LY294002 and Akt/PKB inhibitor IV, alone or in combination with DFMO, were studied using the SRB assay. The results clearly show that DFMO treatment alone inhibited proliferation in LAN-1 cells (Fig. 3A). Addition of LY294002 or Akt/PKB inhibitor IV further decreased the proliferation in both DFMO-treated and control cells in a dose-dependent fashion (Fig. 3B). Overall, the inhibitory effects of LY294002 were less pronounced than those of Akt/PKB inhibitor IV.

To confirm these findings, we examined the effects of DFMO, LY294002, and Akt/PKB inhibitor IV using the MTS cell viability assay. The results supported the findings observed with the SRB assays. DFMO treatment alone decreased cell viability (Fig. 3C). LY294002 and Akt/PKB inhibitor IV further decreased the viability of cells in both DFMO-treated and control cells (Fig. 3D). However, whereas the decrease in cell viability was similar between Akt/PKB inhibitor IV and LY294002 in DFMO-treated cells, the effects of Akt/PKB inhibitor IV alone were much more striking than those of LY294002 in control cells.

To further investigate the cytotoxic effects of LY294002 and Akt/PKB inhibitor IV, both inhibitors were examined in DFMO-treated and control cells by Western blot analysis and probing for poly(ADP-ribose) polymerase (PARP) cleavage (a late apoptotic event). The results showed that LY294002 and Akt/PKB inhibitor IV induced PARP cleavage in DFMO-treated and control cells (Fig. 4). PARP cleavage was not detected with DFMO treatment alone. The combination of DFMO with LY294002 or Akt/PKB inhibitor IV also led to a decrease in the total protein level of noncleaved (116 kDa) PARP protein.

The effects of DFMO on p27^Kip1 phosphorylation. Although we previously showed that the treatment of LAN-1 cells with DFMO induced p27^Kip1 accumulation and p27^Kip1/Rb-coupled G₁ cell cycle arrest (5), the mechanism by which DFMO induces these effects has not been established. Therefore, we examined the phosphorylation state of p27^Kip1 in response to DFMO treatment. Total p27^Kip1 protein was immunoprecipitated from DFMO-treated and control cells and immunoblotted to assess the phosphorylation state of p27^Kip1 using phospho-specific antibodies targeted against individual p27^Kip1 phosphorylation sites (Fig. 5A). Although there was a marked increase in the phosphorylation of p27^Kip1 at Ser¹⁰ and Thr¹⁹⁸ with DFMO, the phosphorylation of p27^Kip1 at the Thr¹⁸⁷ site was not affected.

Because p27^Kip1 is a downstream target of Akt/PKB, we next tested whether DFMO-induced activation of Akt/PKB directly modulates the phosphorylation state of p27^Kip1. To accomplish this, we examined the effects of inhibitors LY294002 and Akt/PKB inhibitor IV on the phosphorylation state of p27^Kip1 in DFMO-treated and control cells. We immunoprecipitated total p27^Kip1 protein and immunoblotted for phospho-p27^Kip1 (Ser¹⁰, Thr¹⁸⁷, and Thr¹⁹⁸) using phospho-specific p27^Kip1 antibodies. We found that LY294002 did not attenuate DFMO-induced phosphorylation of p27^Kip1, which would be expected if Akt/PKB was the major phosphorylating enzyme (Fig. 5B). Surprisingly, LY294002 alone and in combination with DFMO increased p27^Kip1 protein expression and phosphorylation of p27^Kip1 at Ser¹⁰ and Thr¹⁹⁸ (Fig. 5B). This effect was significantly augmented in LY294002-treated cells compared with DFMO-treated and control cells without LY294002. Akt/PKB inhibitor IV had little or no effect on p27^Kip1 protein and phosphorylation of p27 at Ser¹⁰ and Thr¹⁹⁸ (Fig. 5B). Both LY294002 and Akt/PKB inhibitor IV did not have any detectable effects on phosphorylation of p27^Kip1 at Thr¹⁸⁷.

Previous studies showed that DFMO treatment of LAN-1 NB cells depleted putrescine levels and decreased spermidine and spermine levels by 84% and 47%, respectively (5). DFMO also induced G₁ cell cycle arrest but did not cause apoptosis. In the present study, we investigated the role of Akt/PKB in DFMO-treated LAN-1 NB cells. We found that the survival of DFMO-treated cells is mediated, at least in part, by the phosphorylation of Akt/PKB via the PI3K/Akt signaling pathway. DFMO also induced GSK-3β phosphorylation at residue Ser⁹. This phosphorylation event leads to the inactivation of GSK-3β, which is also associated with inhibition of apoptosis. The differences in GSK-3β phosphorylation with DFMO compared with SAM486A treatment may reflect the changes in intracellular polyamine content resulting from inhibition of different polyamine biosynthesis enzymes, ODC and AdoMetDC, respectively. With DFMO, putrescine and spermidine levels are depleted, whereas spermine levels remain relatively stable. SAM486A induces strong accumulation of putrescine and depletion of spermidine and spermine (5). The different effects of these inhibitors on polyamine contents could account for the different effects of DFMO and SAM486A on GSK-3β phosphorylation.

The DFMO-induced Akt/PKB phosphorylation was mediated by polyamines, as addition of individual polyamines alleviated this effect. However, it is not clear whether the attenuation of the DFMO effects was due to one particular polyamine because all three polyamines (putrescine, spermidine, and spermine) exerted the same response. The effect of DFMO on Akt/PKB was found to occur via PI3K and not PDK2. This is an interesting finding, as previous studies showed that Ser⁴⁷³ phosphorylation of Akt/PKB was primarily mediated by PDK2 (22). Although Thr³⁰⁸ phosphorylation of Akt/PKB was not detected in DFMO-treated or control cells, we were able to detect Akt/PKB phosphorylation at Thr³⁰⁸ in Akt/PKB-activated positive control extracts, thus verifying that the Akt/PKB Thr³⁰⁸ phospho-antibody is able to detect phosphorylation of Akt/PKB at this site (data not shown). Therefore, the activation of Akt/PKB and the Akt/PKB-mediated effects observed in this study (such as GSK-3β phosphorylation) did not require Thr³⁰⁸ Akt/PKB phosphorylation.

Next, we investigated the role of Akt/PKB in cell survival during DFMO-induced polyamine depletion by examining the effects of LY294002 and Akt/PKB inhibitor IV on cell proliferation, cell viability, and PARP cleavage, a marker for late apoptosis. LY294002 and Akt/PKB inhibitor IV decreased proliferation and cell viability and induced apoptosis in both untreated and DFMO-treated cells. The effects of LY294002 and Akt/PKB inhibitor IV seemed to be stronger when combined with DFMO. However, Akt/PKB inhibitor IV was more effective than LY294002 in control cells. Interestingly, the total levels of noncleaved (116 kDa) PARP protein decreased in DFMO/LY294002–treated and DFMO/Akt/PKB inhibitor IV–treated cells, suggesting that this combination treatment also affects the synthesis or degradation pathways of PARP. The different effects of LY294002 compared with Akt/PKB inhibitor IV on these processes may reflect the different molecular targets of these two inhibitors. LY294002 is a known inhibitor of PI3K but it has also been shown to directly inhibit mammalian target of rapamycin (mTOR; ref. 23). Furthermore, LY294002 has the potential to alter many more effectors and proteins that are downstream from PI3K compared with Akt/PKB inhibitor IV. Akt/PKB inhibitor IV inhibits the phosphorylation and activation of a kinase directly upstream of Akt/PKB but downstream of PI3K. One potential candidate protein that may be differently modulated by LY294002 and Akt/PKB inhibitor IV is mTOR, a protein that is known to function as a sensor of cellular energy and as a regulator of cell survival and cell proliferation (24–29). Therefore, it may influence the LY294002 effects on cell proliferation, viability, and survival observed in this study. The results from this study clearly show that cell survival during DFMO treatment is mediated by the PI3K/Akt pathway. In addition, our results reveal a unique mechanism by which DFMO-induced phosphorylation of Ser⁴⁷³, in the absence of Thr³⁰⁸ phosphorylation, promotes survival in NB cells through a process that is mediated by the PI3K/Akt signaling pathway.

To further explore DFMO-induced effects in NB cells, we looked at the phosphorylation state of p27^Kip1. We found that DFMO increased p27^Kip1 phosphorylation (Ser¹⁰ and Thr¹⁹⁸) and led to p27^Kip1 accumulation. Next, we investigated the role of Akt/PKB in this process, as Akt/PKB has been shown to directly phosphorylate p27^Kip1 (10–13). We found that LY294002 increased the DFMO-induced phosphorylation of p27^Kip1 at Ser¹⁰ and Thr¹⁹⁸ and accumulation of p27^Kip1, and Akt/PKB inhibitor IV had little or no effect on p27^Kip1. Therefore, Akt/PKB is either not involved or plays a minimal or indirect role in this process.

Skp2 is part of the SCFskp2 ubiquitin ligase complex that regulates the ubiquitination and degradation of p27^Kip1 (30–35). DFMO-induced p27^Kip1 phosphorylation and accumulation may reflect a disruption of the SCFskp2-mediated degradation of p27^Kip1. We observed that Thr¹⁹⁸ phosphorylation of p27^Kip1 increased with DFMO, potentially stabilizing the protein and decreasing the degradation of p27^Kip1, suggesting that disruption of p27^Kip1 degradation may be involved in the accumulation of p27^Kip1. This is supported by our previous work in which DFMO treatment led to hypophosphorylation of Rb protein (5). Rb is involved in the regulation of SCFskp2 ubiquitin ligase-mediated p27^Kip1 degradation via the Skp2 autoinduction loop (36, 37). DFMO-induced hypophosphorylation of Rb and accumulation of p27^Kip1 may disrupt the Skp2 autoinduction loop, consequently promoting p27^Kip1/Rb-coupled G₁ cell cycle arrest. Furthermore, the LY294002-induced stabilization and accumulation of p27^Kip1 may be regulated by mTOR/SCFskp2 signaling and/or GSK-3β–mediated regulation of CDK2 assembly with p27^Kip1 and Skp2 protein expression (38, 39).

Based on our new findings and those from previous studies, DFMO induces two opposing pathways: one leading to p27^Kip1-mediated G₁ cell cycle arrest and one leading to Akt/PKB-mediated cell survival. The two pathways and the potential role of Akt/PKB, mTOR, and GSK-3β in regulating the effects of DFMO are illustrated in Fig. 6. These effects occur as a consequence of polyamine depletion induced by DFMO inhibition of ODC, which is a sentinel metabolic enzyme of polyamine biosynthesis, further supporting the potential value of targeting metabolic enzymes as a treatment for cancer. Whereas cell cycle arrest induced by DFMO is a positive aspect of this drug, DFMO-induced Akt/PKB phosphorylation is not ideal for the treatment of cancer. However, our study provides important information as to why several clinical cancer trials with DFMO-based monotherapies may have failed in the past (40–42). DFMO exerts cytostatic effects, slowing the progression of some cancers possibly by inducing prolonged growth inhibition, but does not exhibit cytotoxic effects in most cancers. The lack of cytotoxicity in cancer cells may be due to DFMO-induced activation of Akt/PKB, conferring resistance to apoptosis. The results from the present study support the argument for combination therapies, which may overcome the obstacle of chemoresistance. Further investigation into targeting metabolic enzymes and combination therapies that affect PI3K/Akt signaling and cell cycle regulation in cancer may prove essential for the development of more effective treatment strategies.

No potential conflicts of interest were disclosed.

Grant support: NIH R01 grants CA111419 and R01-supplement CA111419-S1 (A.S. Bachmann).

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.

We thank Dr. Patrick Woster for providing the ODC inhibitor DFMO, Novartis for providing the AdoMetDC inhibitor SAM486A, Dr. D.J. Feith (Pennsylvania State University) and Dr. D. Geerts (University of Amsterdam) for the critical review of the manuscript, David Albert for excellent technical support, and Kelsie Takasaki for her initial contribution to this project.