The tumour suppressor FOXO3 is a key regulator of mantle cell lymphoma proliferation and survival

Authors

  • Antònia Obrador-Hevia,

    1. Cancer Cell Biology Group, Institut Universitari d’Investigació en Ciències de la Salut (IUNICS)
    Search for more papers by this author
  • Margalida Serra-Sitjar,

    1. Cancer Cell Biology Group, Institut Universitari d’Investigació en Ciències de la Salut (IUNICS)
    Search for more papers by this author
  • José Rodríguez,

    1. Cancer Cell Biology Group, Institut Universitari d’Investigació en Ciències de la Salut (IUNICS)
    2. Hospital Universitario Severo Ochoa, Department of Internal Medicine, Leganés, Madrid, Spain
    Search for more papers by this author
  • Priam Villalonga,

    1. Cancer Cell Biology Group, Institut Universitari d’Investigació en Ciències de la Salut (IUNICS)
    2. Departament de Biologia Fonamental, Universitat de les Illes Balears, Palma, Illes Balears
    Search for more papers by this author
  • Silvia Fernández de Mattos

    1. Cancer Cell Biology Group, Institut Universitari d’Investigació en Ciències de la Salut (IUNICS)
    2. Departament de Biologia Fonamental, Universitat de les Illes Balears, Palma, Illes Balears
    Search for more papers by this author

S. Fernández de Mattos, Cancer Cell Biology Group, Edifici Cientificotècnic, Institut Universitari d’Investigació en Ciències de la Salut (IUNICS), Universitat de les Illes Balears, Crta Valldemossa km 7·5. E-07122 Palma, Illes Balears, Spain.
E-mail: silvia.fernandez@uib.es

Summary

The FOXO3 (Forkhead/winged helix box class O 3) transcription factor is a crucial regulator of haematopoietic cell fate that controls proliferation and apoptosis, among other processes. Despite the central role of FOXO3 as a tumour suppressor and phosphatidylinositol 3-kinase (PI3K)/AKT effector, little is known about its involvement in mantle cell lymphoma (MCL) biology. This study investigated the expression and activity of FOXO3 in MCL cell lines and in primary cultures. We analysed the expression of key FOXO regulators and targets, and studied the effect of modulators of FOXO function on cell viability and apoptosis. FOXO3 was constitutively inactivated in MCL cell lines, and showed cytoplasmic localization in patient-derived cells. PI3K and AKT, but not mammalian target of rapamycin (mTOR), inhibitors induced FOXO3 nuclear translocation and activation in correlation with their impact on MCL proliferation and survival. Moreover, FOXO3-defective cells were resistant to PI3K/AKT inhibitors. Reactivation of FOXO function with a nuclear export inhibitor had a profound effect on cell viability, consistent with FOXO3 nuclear accumulation. Interestingly, inhibition of FOXO3 nuclear export enhanced the effect of doxorubicin. Taken together, our results confirm that FOXO3 is a relevant regulator of proliferation and apoptosis in MCL, and suggest that reactivation of FOXO3 function might be a useful therapeutic strategy in MCL patients.

Mantle cell lymphoma (MCL) is a distinct biological and clinical subtype of B-cell non-Hodgkin lymphoma characterized by the t(11;14) (q13;q32) chromosomal translocation, which results in aberrant expression of cyclin D1 (CCND1) (Bosch et al, 1994). Although this chromosomal translocation is present in more than 95% of MCL cases, overexpression of cyclin D1 is not sufficient to cause lymphoma (Bodrug et al, 1994), which means that elucidation of other genetic lesions is essential and will provide insights towards a specific therapy. Additional molecular alterations that occur in subsets of MCL combine the deregulation of cell proliferation (Gronbaek et al, 1998; Pinyol et al, 1998) and survival pathways (Rummel et al, 2004), with a high level of chromosome instability and disruption of the DNA damage response pathways (Camacho et al, 2002) [reviewed in Obrador-Hevia et al (2009)]. Despite this knowledge, limited information is available on the functional significance of components of key signal transduction pathways in transformation of mantle cells and development of MCL.

The Forkhead box, class O (FOXO) family of transcription factors are relevant tumour suppressors that control cell fate inducing cell cycle arrest, programmed cell death and stress detoxification (reviewed in Birkenkamp & Coffer (2003)). FOXO3 is the predominant member of the FOXO family in lymphoid peripheral tissues (Furuyama et al, 2000; Lin et al, 2004), and its inactivation is essential for proliferation of immune cells, as shown in B lymphocytes (Yusuf et al, 2004), T lymphocytes (Stahl et al, 2002) and mastocytes (Moller et al, 2005). FOXO factors are also relevant for malignant transformation of haematopoietic cells (Coffer & Burgering, 2004). The importance of FOXOs in this context is underscored by findings that leukaemia and lymphoma oncogenes, including those that encode the BCR/ABL and NPM/ALK fusion proteins, mediate proliferation and survival signalling in part by inhibition of FOXOs (Fernández de Mattos et al, 2004; Gu et al, 2004). In addition, FOXO3 and FOXO4 are fusion partners of MLL in acute myeloid leukaemias (AMLs) associated with t(6;11)(q21;q23) or t(X;11)(q13;q23), respectively (Borkhardt et al, 1997; Hillion et al, 1997).

FOXO proteins are important targets of the phosphatidylinositol 3-kinase (PI3K)/AKT (also called PKB) pathway (Brunet et al, 1999; Kops et al, 1999; Tang et al, 1999). This pathway is essential for proliferation of haematopoietic cells and for immune homeostasis, and is involved in the development of haematological neoplasia [reviewed in Okkenhaug & Vanhaesebroeck (2003)]. Activation of the PI3K/AKT pathway through disruption of function of the PTEN phosphatase results in AML in murine models (Zhang et al, 2006). The phosphorylation by AKT is one of the major regulatory mechanisms by which FOXO-mediated transcription is repressed (Brunet et al, 1999; Kops & Burgering, 1999; Tang et al, 1999). AKT-phosphorylated FOXO proteins bind to 14-3-3 chaperone proteins and become sequestered in the cytoplasm, where they are unable to regulate gene expression. Interestingly, recent studies have demonstrated the constitutive activation of the PI3K/AKT pathway in MCL. Gene expression profiling studies have revealed that genes coding for the catalytic subunit of PI3K (PI3KCA), the protein kinases PDK1 (PDPK1) or AKT1 (AKT1) show consistent overexpression in leukaemic MCL patient cells, when compared with their normal counterparts (Rizzatti et al, 2005). Also, AKT is phosphorylated in aggressive blastoid MCL primary cells and cell lines (Rudelius et al, 2006), and to a lesser extent in cases of typical MCL. The constitutive activation of AKT in MCL has been correlated with expression of the inactive form of the lipid phosphatase PTEN (Dal Col et al, 2008). Two separate studies show that treatment of MCL cell lines with chemical inhibitors of PI3K and AKT leads to cell-cycle arrest and apoptosis (Rudelius et al, 2006; Dal Col et al, 2008), and reduced telomerase activity (Dal Col et al, 2008).

However, despite their central role as tumour suppressors and PI3K/AKT effectors, little is known about the role of FOXO factors, in particular FOXO3, in the context of MCL. We have aimed to investigate FOXO3 expression and activity in a panel of MCL cell lines and primary MCL cultures. We have also analysed the effect of different modulators of FOXO function on MCL cell viability, proliferation and apoptosis. Our results show that FOXO3 is constitutively inactivated in MCL cells, and importantly that re-activation of FOXO3 function can effectively impair viability of MCL cell lines and primary cultures. Thus, FOXO3 represents a potential target for the development of novel and effective chemotherapeutic regimens for MCL patients.

Materials and methods

Cell lines and reagents

The human MCL cell lines Jeko-1 (Jeon et al, 1998), JVM-2 (Melo et al, 1986), Granta-519 (Rudolph et al, 2004) and Rec-1 (Rimokh et al, 1994) were maintained in RPMI 1640 medium supplemented with 10% heat-inactivated foetal bovine serum and 100 units/ml penicillin/streptomycin (all products from Biological Industries Ltd., Kibbutz Beit Haemek, Israel), in a humidified incubator of 5% CO2 at 37°C. LY294002 and Rapamycin were purchased from Calbiochem (Darmstadt, Germany). Triciribine phosphate (TCB-P; NSC 280594) and doxorubicin (NSC 123127) were obtained from the Developmental Therapeutics Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute (Bethesda, MD, USA). Psammaplysene A (PsA) was synthesized in the laboratory of Prof. J. Clardy, Harvard Medical School (Boston, MA, USA) (Schroeder et al, 2005). QVD (Q-VD-OPH non-O-methylated) was purchased from Calbiochem (Merck, Darmstadt, Germany).

Isolation of MCL primary cells

Tumour cells from peripheral blood from MCL patients diagnosed according to the World Health Organization classification (Harris et al, 1999) were studied. Cyclin D1 overexpression was demonstrated by immunohistochemistry. Informed consent was obtained from each patient in accordance with the guidelines of the Ethical Committee of Clinical Investigation (CEIC-IB, Spain) and the Declaration of Helsinki. Peripheral blood samples were obtained from patients immediately before treatment. Tumour cells were isolated from peripheral blood by Ficoll/Hypaque sedimentation (Seromed, Berlin, Germany). Cells were cryopreserved in liquid nitrogen in the presence of 10% dimethyl sulfoxide and 20% heat-inactivated foetal calf serum. Upon reconstitution, mononuclear cells were maintained for up to 5 days in normal growing media at a density of approximately 0·5–1 × 106 cells/ml, at 37°C in a humidified atmosphere containing 5% CO2.

Cell viability assay

The number of viable cells in culture was determined based on quantification of ATP, which signals the presence of metabolically active cells, using the Cell Titer-Glo Luminiscent assay kit (Promega, Madison, WI, USA) following the manufacturer’s instructions. Briefly, cells (4 × 105 cells/ml) were treated with drugs for the indicated times followed by addition of Cell Titer-Glo reagent. Luminiscence was detected using a multi-well scanning spectrophotometer (Plate chameleon, Hidex, Finland).

Apoptosis analysis

Cellular apoptosis was determined by measurement of caspases 3 and 7 activity by means of the luminometric Caspase-Glo 3/7 assay (Promega) according to the manufacturer’s protocol, using a Synergy HT multidetection microplate reader (Bio-Tek, Winooski, VT, USA).

Cell cycle analysis

Cell cycle analysis was performed using propidium iodide staining as described previously (Fernández de Mattos et al, 2004). Briefly, cells were washed in phosphate-buffered saline (PBS) and fixed in 90% ethanol. Fixed cells were then washed twice in PBS and stained in 50 μmol/l propidium iodide (Sigma-Aldrich, St Louis, MO, USA) containing 5 μg/ml DNase-free RNase for 1 h, then analysed by flow cytometry using a FACScan (Coulter Epics XL-MSL; Beckman Coulter, Fullerton, CA, USA) and winMDI software.

Western blot analysis and antibodies

Western blot whole-cell extracts were prepared by lysing cells with lysis buffer (1% Nonidet P-40, 20 mmol/l Tris–HCl pH 7·4, 100 mM NaCl, 10 mmol/l NaF, 1 mmol/l Na3VO4 and protease inhibitors ‘Complete’ (Roche, Basel, Switzerland)) on ice for 15 min. Protein yield was quantified by Bio-Rad Dc protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA). 50 μg of lysate was separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis, transferred to Immobilon-NC nitrocellulose membranes (Millipore, Billerica, MA, USA) and specific proteins were recognized by specific antibodies. The antibodies used were the following: FOXO3 (#07-702) and phospho-FOXO3-Thr32 (#07-695) from Millipore; AKT (SC-1618) and p27kip1 (SC-528) from Santa Cruz Biotechnology (Santa Cruz, CA, USA); phospho-AKT-Ser473 (#9271), Caspase-3 (#9661), PARP (#9542), phospho-Bcl2-Ser70 (#2827), Bcl-2 (#2870), MCL-1 (#4572) and Bcl-xL (#2764) from Cell Signaling (Beverly, MA, USA); cyclin D1 (DCS6) from NeoMarkers (Fremont, CA, USA); α-tubulin (#T9026) from Sigma-Aldrich. The antibodies were detected using horseradish peroxidase-linked secondary antibodies (Dako, Glostrup, Denmark), and visualized by the enhanced chemiluminiscent (ECL) detection system (Millipore). Protein levels were evaluated by densitometric scanning of the immunoblots using Quantity One software (Bio-Rad laboratories).

Immunofluorescence

Cells were spun on poly-lysine treated 13-mm-diameter coverslips, fixed in 4% formaldehyde before being permeabilized in 0·1% v/v Triton X-100. Coverslips were then blocked in PBS containing 0·3% bovine serum albumin, and antibody recognizing FOXO3 (#07-702; Millipore) was added (1:30 dilution). Specific staining was visualized with a secondary antibody conjugated to Alexa 488 (Dako). Coverslips were mounted using fluorescent mounting medium (Dako), and analysed on a Leica TCS SP2 confocal microscope (Leica Microsystems, Wetzlar, Germany). Nuclei were stained with either ethidium homodimer-1 (Invitrogen, Carlsbad, CA, USA) or 4′,6-diamidino-2-phenylindole (DAPI, Dako).

Statistical analysis

The statistical significance of differences was assessed by Student’s t test. Statistically significant differences are indicated by ***P < 0·001, **P < 0·01 and *P < 0·05.

Results

Analysis of FOXO in MCL cell lines and primary cultures

We first examined the expression and phosphorylation status of components of the FOXO3 pathway in a panel of MCL cell lines, namely Jeko-1, Rec-1, JVM-2 and Granta-519. FOXO3 expression and its AKT-mediated phosphorylation were detected in Jeko-1, Rec-1 and Granta-519 cells (Fig 1A). Interestingly, both total and phosphorylated FOXO3 levels were almost undetectable in JVM-2 cells. The levels of the FOXO3 target p27kip1 (CDKN1B) were high in Jeko-1 and Rec-1 cells, despite FOXO3 phosphorylation, in accordance with a heterogeneous pattern of expression reported for other tumour cell lines (Sunters et al, 2003). Expression levels and phosphorylation of the upstream regulatory kinase AKT showed some variation between the cell lines analysed. Total levels of AKT were higher for Jeko-1 and Rec-1 cells, while JVM-2 and Granta-519 cells showed the highest level of phosphorylated AKT. The levels of cyclin D1 were high in all cell lines except for JVM-2, which showed the lowest level of cyclin D1 expression, in accordance with published data (Tucker et al, 2006). These results indicate that, when expressed, FOXO3 is constitutively phosphorylated and hence inactivated in MCL cell lines. They also show that some MCL cell lines have a dramatic reduction in FOXO3 expression.

Figure 1.

 Expression of components of the FOXO pathway in MCL cells. (A) Western blot analyses of MCL cell lines. Whole cell lysates were prepared from the panel of MCL cell lines. The expression of both phosphorylated (P-) and total FOXO3, phosphorylated and total AKT, cyclin D1 and p27Kip1 was analysed by Western blotting. α-tubulin was used as a loading control. (B) FOXO3 Immunocytochemistry in MCL primary cultures. Primary MCL cells were fixed in 4% formaldehyde. FOXO3 was visualized with a rabbit polyclonal antibody (Millipore) followed by the addition of ALEX488 (green) labelled anti-rabbit antisera. DAPI was added to visualize the nuclei. Bottom panels show the indicated white-boxed areas at higher magnification.

We next conducted immunocytochemistry experiments to assess the expression and subcellular localization of FOXO3 in patient-derived primary cells. FOXO3 expression was detected, and its staining was predominantly cytoplasmic in MCL primary cells (Fig 1B). This observation indicates that FOXO3 is inactive and unable to transactivate its target genes in MCL primary cultures, in agreement with the detection of phosphorylated (inactive) FOXO3 in MCL cell lines.

Effect of PI3K pathway inhibitors on proliferation and viability of MCL cells

In order to modulate FOXO activity, we used small-molecule PI3K pathway inhibitors and characterized their effects in our panel of MCL cells. To this end LY294002 (inhibitor of PI3K), TCB-P (inhibitor of AKT) and Rapamycin (inhibitor of mTOR) were added at various concentrations and cell viability was analysed 48 h after drug addition. The results showed that MCL cell lines responded differently to PI3K and AKT inhibition (Fig 2). Treatment of Jeko-1, Rec-1 and Granta-519 cells with LY294002 resulted in a significant reduction in cell viability, particularly in Jeko-1 cells. Both Jeko-1 and Rec-1 cells also showed a strong sensitivity to TCB-P, which induced an 80% inhibition in viable cell number at 20 μmol/l. In contrast, FOXO3-deficient JVM-2 cells were refractory to LY294002 treatment and displayed a reduced sensitivity to TCB-P. Rapamycin had a lower effect on all cell lines, inducing an average 30% reduction in viable cell number at all concentrations tested. Similar results were obtained at 24 h and in growth curve assays of cells treated with the panel of inhibitors (data not shown). Taken together, these results indicate that inhibition of mTOR is a less effective strategy than PI3K or AKT inhibition in order to reduce MCL viability and growth. Interestingly, our data also suggest that reduced FOXO3 expression confers resistance to PI3K and AKT inhibition in JVM-2 cells.

Figure 2.

 Cell viability analysis of MCL cells after treatment with LY294002, TCB-P and Rapamycin. Cells were treated with LY294002, Triciribine-phosphate (TCB-P) or Rapamycin at the indicated concentrations and incubated for 48 h, followed by addition of Cell Titer-Glo reagent and detection of luminescence with a multi-well scanning spectrophotometer. Cell viability is represented as a percentage relative to untreated cells, and data is means ± SEM from three independent determinations performed in duplicate. Statistically significant differences are indicated by ***P < 0·001, **P < 0·01 and *P < 0·05 compared with untreated cells.

Analysis of apoptosis and cell cycle profile in MCL cells

In order to assess whether the observed effects on cellular viability upon PI3K pathway inhibition were due to an induction of apoptosis, the activity of caspases 3/7 was measured in MCL cells treated with the various inhibitors. Interestingly, whereas in Jeko-1 and Rec-1 cells LY294002 and TCB-P clearly increased caspase activity, this effect was negligible in JVM-2 cells (Fig 3A). In agreement with our previous observations, no differences in apoptosis induction were observed in Rapamycin-treated cells.

Figure 3.

 Cell cycle and apoptosis analysis of MCL cells after PI3K pathway inhibition. (A) Caspase activity determination. The indicated MCL cell lines were treated with LY294002 (25 μmol/l), Triciribine-phosphate (TCB-P, 10 μmol/l) or Rapamycin (20 nmol/l) for 24 h and caspase activity was determined using the Caspase-Glo 3/7 assay kit. Caspase activity is represented as fold induction relative to untreated cells, and data is means ± SEM from three independent determinations performed in duplicate. Statistically significant differences are indicated by *P < 0·05 for JVM-2 cells compared with Jeko-1 cells, or ++P < 0·01 for JVM-2 cells compared with Rec-1 cells. (B) Cell cycle analysis of MCL cells. Jeko-1 (top panel) and JVM-2 cells (bottom panel) were treated with LY294002 (10 μmol/l), TCB-P (10 μmol/l) or Rapamycin (20 nmol/l) for 24 h, fixed in ethanol, stained with propidium iodide, and DNA content determined by flow cytometry. A representative cell cycle profile is shown for each cell line, with the percentage of cells in each phase of the cell cycle: sub-G1 (<2N), G1, S and G2/M. (C) Determination of apoptosis induction by flow cytometry. The extent of apoptosis induction with the various drugs (sub-G1 population) is represented as fold increase (compared to untreated cells) for Jeko-1 and JVM-2 cells. Data is means ± SEM from three independent determinations. Statistically significant differences are indicated by **P < 0·01 for JVM-2 cells compared with Jeko-1 cells (D) Expression of anti-apoptotic proteins and apoptotic markers. Whole cell lysates were prepared from Jeko-1 and JVM-2 cells treated with LY294002 (10 μmol/l) and TCB-P (10 μmol/l) for the times indicated. The expression of Bcl-2, MCL-1, Bcl-XL, phosphorylated Bcl-2 (P-Bcl-2) and cleavage of caspase 3 and PARP was analyzed by Western blotting. α-tubulin was used as a loading control. The levels of phospho-Bcl2 and Mcl-1 were evaluated by densitometric scanning of the immunoblots and corrected for the amount of Bcl-2 and α-tubulin, respectively.

We next investigated the cell cycle phase distribution of inhibitor-treated MCL cells by flow cytometry. The data obtained indicated that there was a marked heterogeneity in the extent of apoptosis induction between the cell lines analysed, as revealed by the percentage of cells with a <2N DNA content (sub-G1 phase) (Fig 3B). Upon LY294002 treatment, Jeko-1 cells not only showed an increase in the apoptotic fraction, but were also clearly arrested in G1. TCB-P had a very pronounced effect on apoptosis induction in Jeko-1 cells, in agreement with the results obtained in cell viability assays. Similar results were obtained in Rec-1 cells treated with PI3K and AKT inhibitors (data not shown). In contrast, in JVM-2 cells neither the apoptotic nor the G1 fractions were increased upon LY294002 or TCB-P treatment, also correlating with previous analysis. Interestingly, Rapamycin had a minor effect on the accumulation of MCL cells in the G1 phase, while no apoptotic fraction was detected after Rapamycin treatment. The comparison between the percentages of sub-G1 cells (Fig 3C) clearly indicated that FOXO3-deficient JVM-2 cells showed a lower sensitivity to PI3K or AKT inhibitors when compared with FOXO3-expressing Jeko-1 cells. It also indicated that mTOR inhibition with Rapamycin had a negligible effect on cell survival in both cell lines.

In order to extend these observations, we monitored the expression and phosphorylation levels of a number of pro-survival and apoptotic markers following the treatment of MCL cells with LY294002 and TCB-P. The results confirmed a differential response between the cell types analysed (Fig 3D). In sensitive Jeko-1 cells, PI3K/AKT inhibitors led to the downregulation of the pro-survival Bcl-2-family member MCL-1, together with the inhibition of Bcl-2 phosphorylation, while no changes were observed for Bcl-XL. LY294002 and TCB-P caused a 6-fold decrease in Mcl-1 levels, while phospho-Bcl-2 levels decreased up to 8-fold in LY294002-treated Jeko-1 cells, and up to 3-fold after TCB-P treatment. Accordingly, cleavage of caspase 3 and PARP was clearly detected in Jeko-1 cells. In contrast, no significant changes in expression, phosphorylation or cleavage of the analysed proteins could be detected in JVM-2 cells upon LY294002 or TCB-P treatments. These results confirm that inhibition of PI3K or AKT induce apoptosis in Jeko-1 cells, while this response is blocked in JVM-2 cells lacking FOXO3 expression.

Molecular changes induced by PI3K pathway inhibitors in MCL cells

To further investigate the connection between FOXO3 activity and sensitivity to PI3K/AKT inhibition, we monitored alterations on relevant components of the pathway following treatment of MCL cells with the various inhibitors (Fig 4A). In the sensitive Jeko-1 cell line, LY294002 induced a sustained activation of FOXO3 up to 16 h after treatment, as indicated by its rapid dephosphorylation. This correlated with a sustained inhibition of AKT phosphorylation. In the resistant JVM-2 cell line LY294002 was also able to dephosphorylate and inactivate AKT but, as expected, expression of FOXO3 could not be detected. Interestingly, the cell cycle inhibitor p27Kip1, a direct target of FOXO3, was clearly up-regulated in Jeko-1 cells upon LY294002 treatment, whereas its levels remained unaltered in JVM-2 cells.

Figure 4.

 Analysis of the FOXO3 pathway in response to PI3K pathway inhibition. (A) Whole cell lysates were prepared from Jeko-1 and JVM-2 cells treated with LY294002 (10 μmol/l), Triciribine-phosphate (TCB-P, 10 μmol/l) or Rapamycin (20 nmol/l) for the times indicated. The expression and phosphorylation of the indicated proteins were analysed by Western blotting. Also shown is a blot for α-tubulin as a loading control. (B) p27kip1 expression levels are represented as fold induction relative to untreated cells for LY294002, TCB-P and Rapamycin treatments. The levels of p27kip1 for each time point were evaluated by densitometric scanning of the immunoblots and corrected for the amount of α-tubulin.

As expected, the AKT inhibitor TCB-P reduced AKT phosphorylation levels in both cell lines, although at later times in the case of JVM-2. Again, this correlated in Jeko-1 cells, but not in JVM-2 cells, with dephosphorylation and activation of FOXO3 and induction of p27kip1 expression. This suggests that reduced expression of FOXO3 in JVM-2 cells correlates with the inability to increase the expression of its target p27Kip1 after PI3K/AKT inhibition.

Given that the effects of Rapamycin on cell viability and apoptosis induction were only modest, we evaluated the molecular changes induced by this drug upon prolonged (24–48 h) treatment. Although Rapamycin induced the dephosphorylation of AKT, phosphorylated and inactive FOXO3 was observed in Jeko-1 cells at all times analysed (Fig 4A). As expected, FOXO3 expression was undetectable in JVM-2 cells. Accordingly with its inability to induce FOXO3 reactivation, Rapamycin treatment did not affect p27kip1 expression (Fig 4A).

The analysis of p27kip1 levels after treatment with the panel of inhibitors confirmed the clear differences in p27kip1 induction between Jeko-1 and JVM-2 cells for LY294002 and TCB-P, but not Rapamycin, treatments (Fig 4B).

FOXO nuclear export inhibitor Psammaplysene A affects viability of MCL cells

To further confirm the relevant role of FOXO3 in MCL, we used the small molecule compound Psammaplysene A (PsA), which has been demonstrated to induce FOXO nuclear localization in PTEN-deficient cancer cells without altering the activity of its upstream kinase AKT (Kau et al, 2003). Accordingly, molecular analysis of Jeko-1 cells confirmed that neither AKT nor FOXO3 phosphorylation were altered upon PsA treatment (Fig 5A). Expression of AKT also remained constant for the time analyzed, while accumulation of FOXO3 protein became evident after 2–4 h of drug treatment.

Figure 5.

 Effect of Psammaplysene A on cell viability and FOXO3 activity in MCL cells. (A) Analysis of FOXO3 and AKT after PsA treatment of Jeko-1 cells. Whole cell lysates were prepared from Jeko-1 cells treated with Psammaplysene A (PsA, 2·5 μmol/l) for the times indicated. The expression and phosphorylation of FOXO3 and AKT were analysed by Western blotting. Also shown is a blot for α-tubulin as a loading control. (B) Cell viability assay. Cells were treated with PsA at the indicated concentrations and incubated for the indicated times, followed by addition of Cell Titer-Glo Reagent and detection of luminescence with a multi-well scanning spectrophotometer. Cell viability is represented as a percentage relative to untreated cells, and data is means ± SEM from three independent determinations performed in duplicate. Statistically significant differences are indicated by **P < 0·01 and *P < 0·05 for JVM-2 cells compared with Jeko-1 cells. (C) Caspase activity determination. Jeko-1 and JVM-2 cells were treated with PsA (2·5 μmol/l) for the indicated times and caspase activity was determined using the Caspase-Glo 3/7 assay kit. Caspase activity is represented as fold induction relative to untreated cells, and data is means ± SEM from three independent determinations performed in duplicate. Statistically significant differences are indicated by **P < 0·01 and *P < 0·05 for JVM-2 cells compared with Jeko-1 cells. (D) Cell viability assay of MCL primary cultures after treatment with PsA. Patient cells were treated with the indicated concentrations of PsA and incubated for 48 h, followed by addition of Cell Titer-Glo Reagent and detection of luminescence with a multi-well scanning spectrophotometer. Cell viability is represented as a percentage relative to untreated cells. (E) and (F) Cell viability assay of MCL cells after treatment with PsA and doxorubicin. Jeko-1 and JVM-2 cells (E) or MCL primary cultures (F) were treated with the indicated concentrations of doxorubicin (Doxo, solid line) combined with PsA when indicated (dashed line, (E) 0·5 μmol/l, (F) 1 μmol/l) and incubated for 48 h, followed by addition of Cell Titer-Glo Reagent and detection of luminescence with a multi-well scanning spectrophotometer. Cell viability is represented as a percentage relative to untreated cells. For Jeko-1 and JVM-2 cells, data is means ± SEM from three independent determinations performed in duplicate.

We next studied the effect of PsA on cell viability of MCL cells, performing both a dose-response and a time-course analysis (Fig 5B). The results obtained indicated that treatment with PsA had a profound effect on Jeko-1 cells expressing FOXO3, while the response of JVM-2 cells was less pronounced. The differential effect of PsA between Jeko-1 and FOXO3-deficient JVM-2 cells was obvious at all the concentrations and time-points analysed.

In order to investigate whether the cellular effects of PsA on MCL cells were due to apoptosis induction, we measured caspase 3/7 activity upon PsA treatment. Jeko-1 cells showed a clear and sustained increase in caspase activity after PsA treatment, while only a low and transient induction of caspase activity was detected in JVM-2 cells (Fig 5C). Accordingly, pretreatment of MCL cells with the caspase inhibitor QVD (Q-VD-OPH non-O-methylated) partially and significantly reverted the effect of PsA on cell viability (data not shown).

These results were confirmed in clinical samples of MCL, treating primary cells derived from two different MCL patients with different doses of PsA (range, 1–5 μmol/l). Interestingly, cell viability analysis revealed that PsA dramatically reduced the number of viable cells in both patient samples, with a >95% inhibition in cell viability at 5 μmol/l (Fig 5D).

We finally assessed whether FOXO3 reactivation with PsA could enhance the efficacy of doxorubicin, the main component of the MCL chemotherapeutic regimen CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone/prednisolone). Using both Jeko-1 and JVM-2 cells and patient-derived primary cells, we analysed the effect on cell viability of doxorubicin treatment alone compared with the combination of doxorubicin together with PsA. Interestingly, in Jeko-1 cells and primary cultures a suboptimal concentration of the FOXO modulator showed a cooperative effect with all concentrations of doxorubicin tested (Figs 5E and 5F). The viability of cells treated with the combination of doxorubicin and PsA was lower than the viability of cells treated with either agent alone. In contrast, the addition of PsA had no effect on the response of JVM-2 cells to doxorubicin (Fig 5E). These results suggest that concomitant treatment with chemical modulators of FOXO3 activity can enhance the chemotherapeutic effect of conventional drugs used in the treatment of MCL.

FOXO3 subcellular localization in MCL cell lines and patients

As FOXO proteins are transcription factors and their activity is thus dependent on their subcellular localization, we investigated whether the different compounds used in our study were able to relocate FOXO3 to the nuclei of MCL cells as part of its activation process. To this end, Jeko-1 cells were treated with LY294002, TCB-P, Rapamycin and PsA, and the intracellular localization of FOXO3 was analysed by immunofluorescence. In untreated cells FOXO3 resided in the cytoplasm, with negligible nuclear staining (Fig 6A). Interestingly, LY294002, TCB-P and PsA clearly induced FOXO3 nuclear translocation, in agreement with our previous data regarding sensitivity to these inhibitors and FOXO3 activation. In sharp contrast, subcellular localization of FOXO3 remained cytoplasmic and unaltered in Rapamycin-treated cells (Fig 6A), clearly suggesting a connection between the inability of Rapamycin to reactivate FOXO3 and its modest impact on MCL cell viability. Similar results were obtained using Granta-519 cells (data not shown).

Figure 6.

 Subcellular localization of FOXO3 in MCL cells. (A) Subcellular localization of FOXO3 in Jeko-1 cells. Cells were treated for 16 h with LY294002 (10 μmol/l), Triciribine-phosphate (TCB-P, 10 μmol/l), Rapamycin (20 nmol/l) or PsA (2·5 μmol/l) before being fixed in 4% formaldehyde. FOXO3 was visualized with a rabbit polyclonal antibody (Millipore) followed by the addition of ALEX488 (green) labeled anti-rabbit antisera. Ethidium homodimer-1 was added to visualize the nuclei. (B) Subcellular localization of FOXO3 in MCL primary cultures. Primary MCL cells were treated for 16 h with PsA (2·5 μmol/l), before being fixed in 4% formaldehyde. FOXO3 was visualized with a rabbit polyclonal antibody (Millipore) followed by the addition of ALEX488 (green) labelled anti-rabbit antisera. DAPI was added to visualize the nuclei. Insert panels show the indicated white-boxed areas at higher magnification.

Furthermore, the analysis of FOXO3 subcellular localization in primary cultures confirmed that in patient-derived cells PsA is also able to induce FOXO3 nuclear retention (Fig 6B), which correlates with its observed effect on cell viability. Taken together, these results support our hypothesis that nuclear translocation and activation of FOXO3 is a necessary requisite for the effective induction of cell death in response to modulators of the PI3K pathway.

Discussion

The FOXO transcription factors regulate a wide variety of cellular responses directly related to oncogenesis, such as cell cycle arrest, induction of cell death and protection from stress stimuli. FOXOs were first identified in humans as the sites of chromosomal translocations responsible for rhabdomyosarcoma (FOXO1/FOXO3) and acute leukaemia (FOXO4) (Galili et al, 1993; Borkhardt et al, 1997). Further characterization of their biology (reviewed in Lam et al (2006)) has allowed the identification of these factors as tumour suppressors, including the recent generation of FOXO1/3a/4 triple knock-out mice (Paik et al, 2007; Tothova et al, 2007) in which deletion of all FOXO alleles generates thymic lymphomas and haemangiomas (Paik et al, 2007). Also, the expression of dominant-negative FOXO in Eμ-myc transgenic haematopoietic stem cells accelerates lymphoma development in recipient mice by attenuating apoptosis (Bouchard et al, 2007).

Regulation of these factors is influenced by signalling pathways that are crucial for malignant transformation, such as PI3K/AKT. Notably, gene expression profiling and proteomic studies have reported aberrant expression of PI3K/AKT pathway genes in MCL (Ghobrial et al, 2005; Rizzatti et al, 2005), including FOXO3 (Rizzatti et al, 2005). Also, cellular studies of MCL cell lines and primary cultures have established the importance of this pathway in MCL (Peponi et al, 2006; Rudelius et al, 2006; Dal Col et al, 2008). Recently, a gain of gene copy number of PIK3CA has been reported as a frequent genetic alteration that contributes to MCL progression (Psyrri et al, 2009). Despite this knowledge, little has been established regarding the role of FOXO proteins as regulators of cell proliferation and cell death in the context of MCL. The main goal of this study was to evaluate the contribution of FOXO3 to cell growth and survival of MCL cells.

For this purpose, we characterized the FOXO pathway in a panel of MCL cell lines and in MCL primary cultures. First, we analysed the expression and phosphorylation of FOXO3 showing that, when expressed, FOXO3 is constitutively phosphorylated and inactive in MCL cell lines. In line with this, other studies have reported the presence of phosphorylated FOXO3 in blastoid variants of MCL (Rudelius et al, 2006), and the phosphorylation and inactivation of PTEN, another relevant tumour suppressor and mediator of the PI3K pathway, in MCL cell lines and patient samples (Dal Col et al, 2008). Notably, in our study the JVM-2 cell line showed a dramatic reduction in FOXO3 levels. Also, our analysis of MCL cell lines and patient-derived cells revealed the cytoplasmic localization of FOXO3, which implies an inactive state. Interestingly, the expression of phosphorylated and cytoplasmic FOXO3 has also been reported in acute myeloid leukaemia (Kornblau et al, 2010) and breast cancer (Habashy et al, 2011) respectively, and both parameters have been linked to poor prognosis.

Further characterization of the MCL cell lines according to their response to modulators of FOXO function showed that sensitivity to the various inhibitors correlated with levels of FOXO expression, because FOXO3-defective JVM-2 cells are significantly resistant to LY294002 and TCB-P treatments. Apoptotic assays and analysis of apoptosis markers confirmed that, whereas PI3K/AKT inhibition had a moderate effect on JVM-2 cells, it clearly induced apoptosis in FOXO3-expressing cells. Interestingly, resistance to the pro-apoptotic effect of LY94002 has been reported for primary cells from a MCL patient (Dal Col et al, 2008). It would be interesting to investigate whether such refractory primary cells are also defective for FOXO3 expression. Our results also prove that inhibition of mTOR with Rapamycin does not alter the apoptotic index of MCL cells and can thus be considered a less effective therapeutic strategy. The observed differences in cell growth and apoptosis between PI3K and mTOR inhibitors correlate with published results that report a lack of apoptosis induction by Rapamycin (Hipp et al, 2005), and a decreased effect when compared with LY294002 (Peponi et al, 2006; Dal Col et al, 2008), in various MCL cell lines and primary cultures.

A detailed analysis of the molecular changes induced by treatment with the various inhibitors showed that, in sensitive cells, FOXO3 is dephosphorylated and activated by LY294002 and TCB-P, which correlates with upregulation of its target p27kip1. A similar effect on p27kip1 has been reported in Granta-519 cells treated with LY294002 (Rudelius et al, 2006; Dal Col et al, 2008) and anotherAKT inhibitor (Rudelius et al, 2006). The complete dephosphorylation of FOXO3 after treatment of Jeko-1 cells with LY294002 and TCB-P indicate that both drugs effectively inhibited the survival kinase AKT and induced the activity of the FOXO3 transcription factor. Interestingly, the failure of those inhibitors to transactivate p27kip1 expression in JVM-2 cells coincided with the defect in FOXO3 expression. Moreover, although prolonged exposure to Rapamycin affected AKT phosphorylation (in agreement with previous reports in other cellular systems, (Sarbassov et al, 2006)), it did not alter FOXO activity or expression of p27kip1. Taken together, our results suggest that the ability of PI3K pathway modulators to affect MCL cell proliferation and survival strongly correlates with their capacity to induce FOXO3 activation. Our data thus point to a plausible role for FOXO3 in determining the cellular response to PI3K pathway inhibition.

To further investigate the relevance of FOXO3 in the context of MCL, we used the specific FOXO nuclear-export inhibitor PsA (Kau et al, 2003; Schroeder et al, 2005). This compound has been shown to affect viability of colon cancer cells (Fernández de Mattos et al, 2008) and also to induce apoptosis of endometrial cancer and leiomyoma cells through FOXO activation (Berry et al, 2009; Hoekstra et al, 2009). Our study is the first to demonstrate the effects of PsA in MCL cells. The molecular analysis revealed that FOXO3 is accumulated in PsA-treated MCL cells, suggesting that its nuclear retention may be preventing its cytoplasmic proteasome-mediated degradation. Treatment with PsA had a profound effect on cell viability of Jeko-1 cells, and to a lesser extent of JVM-2. Also, PsA effectively induced caspase activity of Jeko-1 cells, but not of JVM-2 cells. Thus, the ability of PsA to alter cell viability and induce apoptosis correlates with the expression levels of FOXO3.

FOXO3 is the dominant isoform of the FOXO family in lymphocytes (Furuyama et al, 2000; Lin et al, 2004), but the fact that MCL cell lines also express FOXO1, although at lower levels (data not shown), could explain the effect of PsA on JVM-2 cell viability. Interestingly, the viability of patient-derived cells was seriously compromised after exposure of primary cultures to PsA, suggesting that reactivation of FOXO function could represent an effective therapy. Also, treatment of MCL cells and primary cultures with the combination of doxorubicin plus a suboptimal dose of PsA caused a more marked reduction of cell viability than treatment with either agent alone. Notably, the additive effect of PsA was only observed in FOXO3-expressing MCL cells, and not in the FOXO3-defective JVM-2 cell line. These results suggest that reactivation of FOXO function enhances the efficacy of doxorubicin, encouraging the development of treatment protocols for MCL through targeting the FOXO pathway. We have recently reported similar additive results for cisplatin and PsA in colon cancer cells (Fernández de Mattos et al, 2008). Interestingly, a small-molecule inhibitor of the transcriptional repressor BCL6 has been shown to induce cell death in diffuse large B-cell lymphoma (Cerchietti et al, 2010), reinforcing the fact that transcription factors can be therapeutically targeted, and represent a valid approach for the treatment of lymphomas.

Finally, our immunocytochemistry studies of MCL cells showed that LY294002, TCB-P and PsA induced nuclear translocation of FOXO3, which coincided with the dephosphorylation, activation and accumulation observed in the molecular analysis. Notably, treatment with Rapamycin fails to alter FOXO3 cytoplasmic localization. Furthermore, the analysis of FOXO3 localization in primary cultures confirmed that PsA effectively induces FOXO3 nuclear retention in clinical samples of MCL. Thus, our results suggest a clear correlation between the ability of the inhibitors to promote FOXO3 nuclear translocation and their capacity to affect cell viability and induce apoptosis.

In summary, our findings clearly indicate that FOXO3 is a relevant inductor of apoptosis in MCL. On the basis of our observations, we propose that expression studies of FOXO3 will help to identify a subset of patients that may benefit from PI3K inhibition and/or FOXO3 reactivation therapies, thus allowing the design of innovative therapeutic strategies that could open new avenues for the treatment of patients with MCL.

Acknowledgements

This work was supported by grants from Instituto de Salud Carlos III, Ministerio de Ciencia e Innovación (PI060371 and PI09/00058), Direcció General de R+D+I, Govern Balear (PROGECIB-12A) and Junta de Balears-AECC. A.O.-H. is the recipient of a postdoctoral fellowship from Fundació Internacional Josep Carreras-Fundación Caja Madrid (FIJC-08/ESP-FCAJAMADRID). P.V. is a Ramón y Cajal Fellow (Ministerio de Ciencia e Innovación, Spain). We thank the Haematology Services at Hospital Son Llàtzer and Hospital de Manacor (Illes Balears, Spain) for providing patient samples, Dr. Catalina Crespí (Research Unit, Hospital de Son Dureta, Spain) for help with flow cytometry analysis, Dr. J. Clardy (Harvard Medical School, USA) for providing the Psammaplysene A reagent, Dr. Dolors Colomer (Hospital Clínic, Barcelona, Spain) and Dr. Beatriz Martínez (CNIO, Madrid, Spain) for cell lines, and Dr. María Calvo (Serveis Cientificotècnics, Universitat de Barcelona, Spain) for advice and reagents for confocal microscopy analysis. We are indebted to Prof. Eric Lam (Imperial College London, UK) for valuable advice, reagents and continuous support.

Authorship and disclosures

SFdM, PV and JR participated in the conception of the study. AO-H and MS-S performed the laboratory work. SFdM and PV coordinated the research. SFdM, PV and AO-H wrote the paper, with input from JR. SFdM, PV, JR and AO-H participated in data interpretation and discussion. The authors declare no potential conflict of interest.

Ancillary