Counteracting Autophagy Overcomes Resistance to Everolimus in Mantle Cell Lymphoma

Mantle cell lymphoma (MCL) is an aggressive lymphoid neoplasm that accounts for 5% to 10% of all B-cell non-Hodgkin's lymphomas. It is genetically characterized by the chromosomal translocation t(11;14)(q13;q32) resulting in overexpression of cyclin D1. Moreover, high levels of chromosomal instability related to the disruption of the DNA damage response pathway and activation of cell survival mechanisms may confer an aggressive clinical course to the disease (1). Standard chemotherapy approaches are frequently used, but long-term remissions are rare. After failure of first- or second-line treatments, various single agents are used despite limited response rates (2). Thus, there is still a strong unmet medical need for new treatment options in MCL.

The constitutive activation of the phosphatidylinositol 3-kinase (PI3K) and the serine/threonine kinases Akt and mTOR are known to confer drug resistance to many types of cancer, including MCL (2). Thus, this pathway has emerged as a promising therapeutic target. mTOR is an evolutionarily conserved kinase that integrates signals from growth factors, nutrients, and stresses to regulate multiple processes, including mRNA translation, cell-cycle progression, autophagy, and cell survival. mTOR resides in 2 distinct multiprotein complexes referred to as mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2). By phosphorylating S6 kinase 1 and the eIF4E-binding protein 1 (4E-BP1), mTORC1 controls the translation of key regulatory proteins involved in cell proliferation (3). mTORC2 modulates cell survival in response to growth factors by phosphorylation of its downstream effectors Akt and serum/glucocorticoid regulated kinase 1 (4). In addition to directly activate Akt as part of mTORC2, mTOR, as part of mTORC1, also negatively regulates Akt by suppressing growth factor–driven pathways (5).

The macrocyclic lactone antibiotic rapamycin (sirolimus), an allosteric mTORC1 inhibitor isolated from Streptomyces hygroscopicus, was found to possess immunosuppressive and antiproliferative properties. More recently, 3 rapamycin derivatives (termed rapalogs)—temsirolimus (CCI-779), ridaforolimus (AP23573, MK-8669), and everolimus (RAD001)—with improved properties were developed (6). Temsirolimus was approved by the European Medicines Agency for the treatment of relapsed and refractory MCL, while ridaforolimus and everolimus have been studied in various clinical trials (7). These rapalogs have reported a moderate success as monotherapy with mild toxic effects, with an observed half-life of about 30 hours (8, 9). A recent phase II study of everolimus reported a 32% overall response rate in MCL patients (10), which is similar to that found in other trials using temsirolimus (11–13). However, the effectiveness of these agents as single agent therapies is stifled in part by strong mTORC1-dependent negative feedback loops that become inactive on mTORC1 inhibition, paradoxically leading to survival promoting events such as the activation of Akt (5, 14). Thus, these compounds may represent good candidates in combination with chemotherapy and targeted agents capable of counteracting these mechanisms of resistance (15).

In this context, our purpose was to describe the molecular bases of MCL cell response to everolimus, with the aim to validate new possible combination strategies. Herein, we show that everolimus induces the rephosphorylation of Akt, mediated by a feedback loop within the PI3K/Akt/mTOR axis that can be blocked by the Akt inhibitor VIII isozyme-selective Akti-1/2. We also show the protective effect of autophagy to actively counteract everolimus and Akti-1/2 activity in MCL cell lines and primary samples. Thus, we propose the triple targeting of mTOR, Akt, and autophagy as a requisite for effective antitumoral therapy in MCL.

The MCL cell lines REC-1, JEKO-1, GRANTA-519, UPN-1, HBL-2, JVM-2, MAVER-1, and Z-138 (Table 1) used in this study were cultured in RPMI 1640 or Dulbecco's Modified Eagle's Medium, supplemented with 10% to 20% heat-inactivated FBS, 2 mmol/L glutamine and 50 μg/mL penicillin-streptomycin (Life Technologies). All cultures were routinely tested for Mycoplasma contamination by PCR and the identity of all cell lines was verified by using AmpFISTR identifier kit (Life Technologies).

Cells from 11 patients diagnosed with MCL according to the World Health Organization classification criteria (16), who had not received treatment for the 3 previous months, were used. The clinical characteristics of these patients are listed in Table 1. Informed consent was obtained from each patient in accordance with the guidelines of the Ethical Committee of the Hospital Clinic in Barcelona, Spain, and the Declaration of Helsinki. For all samples, cyclin D1 overexpression was determined by immunohistochemistry or real-time PCR. P53 mutational status was assessed by FISH and direct sequencing. Mononuclear cells from peripheral blood samples (PBMC) were isolated by Ficoll/hypaque sedimentation (GE Healthcare), and conserved within the Hematopathology Biobank of our institution (CDB Biobank/IDIBAPS-Hospital Clínic Biobank). Cells were either used directly or cryopreserved in liquid nitrogen in the presence of 10% dimethyl sulfoxide, 60% FBS, and 30% RPMI 1640. Freezing/thawing manipulations did not influence cell response (17).

Cells received as indicated a single treatment of everolimus (kindly provided by Novartis), Akt inhibitor VIII isozyme-selective Akti-1/2 (Calbiochem), hydroxychloroquine sulfate (Sigma), bafilomycin A1 (Sigma), and tamoxifen (Enzo Life Sciences). Cell viability was quantified by staining of external exposure of phosphatidylserine (PS) residues with Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI; Bender Medsystems). For the analysis of apoptosis in CD3⁺ and CD19⁺ subpopulations, PBMCs were labeled simultaneously with anti-CD3-FITC, anti-CD19-Phycoerythrin (PE; Becton Dickinson) antibodies, and Annexin V-allophycocyanin (APC; Bender Medsystems). Changes in mitochondrial transmembrane potential (Δψm) and reactive oxygen species (ROS) production were evaluated by staining cells with 20 nmol/L 3,3′-diexyloxacarbocyanine iodide (DiOC₆[3]; Life Technologies) and 2 μmol/L dihydroethidine (DHE; Life Technologies), respectively. For the quantification of caspase-3/7 activity, cells were labeled with 2 μmol/L CellEvent caspase-3/7 green detection reagent (Life Technologies) for 30 minutes at 37°C. A total of 10,000 stained cells per sample were acquired and analyzed in a FACScan or FACSCalibur flow cytometer by using Cellquest and Paint-A-Gate softwares (Becton Dickinson). Lethal dose 50 (LD₅₀) was defined as the concentration of drug required to reduce cell viability by 50%. In drug combination studies, combination index (CI) values were calculated according to the Chou–Talalay method by means of the Calcusyn software version 2.0 (Biosoft), where CI < 1 indicated synergistic effect between 2 drugs.

MCL cell lines (5 × 10⁴) were incubated for 72 hours with everolimus. MTT [3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide] reagent (Sigma) was added for 2 to 5 additional hours before spectrophotometric measurement. Each measurement was made in triplicate, and the mean value was calculated. For each cell line, values were represented using untreated control cells as a reference. The growth inhibitory activity 50 (GI₅₀) was calculated as the dose that produced 50% growth inhibition.

Cells were incubated for 72 hours with everolimus, washed in PBS and fixed with 70% ethanol. After incubation with PI for 30 minutes at 37°C, cell-cycle fractions were determined by flow cytometry (FACScan). The analysis was conducted by applying the ModFit LT software (Verity Software House).

Whole-cell protein extracts were obtained by lysing cells in Triton buffer (20 mmol/L Tris-HCL pH 7.6, 150 mmol/L NaCl, 1 mmol/L EDTA, and 1% Triton X-100) supplemented with protease and phosphatase inhibitors (10 μg/mL leupeptin, 10 μg/mL aprotinin, 1 mmol/L phenylmethanesulfonyl fluoride, 5 mmol/L NaF, and 2 mmol/L Na₃VO₄). Solubilized proteins were quantified by Bradford protein assay and 50 μg of cell lysates were loaded onto 12% to 15% SDS-PAGE and transferred to an Immobilon-P membrane (Millipore). Membranes were blocked in TBS-Tween 20 containing 5% phosphoBlocker Blocking Reagent (Cell Biolabs) and probed with antibodies against: phospho-mTOR (Ser2448), phospho-Akt (Ser473), phospho-S6 ribosomal protein (S6RP; Ser235/236), phospho-4E-BP1 (Thr37/46), mTOR, Akt, S6RP, 4E-BP1, cleaved caspase-3 (Asp175, clone 5A1E), LC3B, ATG5 (D1G9) and ATG3 (Cell Signaling Technology), PARP (Roche), and ATG7 (clone EP1759Y; Abcam). Membranes were then incubated with horseradish peroxidase-labeled anti-mouse or anti-rabbit (Sigma and Cell Signaling Technologies) secondary antibodies. Chemiluminiscence detection was done by using ECL system (Pierce) in a mini-LAS4000 Fujifilm device, and relative protein quantification of LC3B was conducted with Image Gauge software (Fujifilm). Equal protein loading was confirmed by analyzing α-tubulin expression (Sigma).

Samples containing 5 × 10⁵ to 1 × 10⁶ cells were stained for autophagolysosome contents with Cyto-ID Green Detection Reagent (Enzo Life Sciences) as described (18) for 30 minutes at 37°C, and the intensity of the green fluorescence was measured by flow cytometry (FACScan). Results were represented as histogram overlays. Hoechst 33342 Nuclear Stain was added to the Cyto-ID Green Detection Reagent stained cell suspension and applied to a glass microscope slide, covered with a cover slip and visualized on a Olympus BX41 microscopy (×40 objective) with the use of Cell⁁ID Imagine Software (Olympus).

Z-138 cells (5 × 10⁶) were cultured for 24 hours without antibiotics and washed with FBS-free RPMI medium. Cells were then electroporated with a Nucleofector II device (Lonza) in 100 μL of Ingenio Electroporation Solution (Mirus) containing 10 μmol/L of a mix of 6 different Silencer Select Predesigned siRNAs targeting ATG7, ATG5, and ATG3, and a nonsilencing negative control (Ambion). C-005 Nucleofector program was used. After transfection, cells were transferred to culture plates for 6 hours before experiments were set up.

Data are represented as mean ± SD or SEM of 3 independent experiments. Statistical analysis was conducted with the use of GraphPad Prism 4.0 software (GraphPad Software). Two-way ANOVA was used to determine how response is affected by 2 factors. Results were considered statistically significant when P value < 0.05 (*, P < 0.05, **, P < 0.01, ***, P < 0.001).

To explore the antitumoral effect of everolimus, a panel of 8 MCL cell lines and 11 MCL primary samples were exposed for 48 hours (primary cells) or 72 hours (cell lines) to increasing doses of the drug (0.05, 0.5, and 5 μmol/L). Apoptosis induction and drug cytostatic effect were determined simultaneously by flow cytometry and MTT proliferation assay, respectively. As shown on Fig. 1A, everolimus mainly induced a cytostatic effect in MCL cell lines in a dose-dependent manner with minor (< 11%) cytotoxic activity (Table 1). In the high-sensitive MCL cell lines (UPN-1, GRANTA-519, JEKO-1, and REC-1), most of this effect (>50% antiproliferative activity) was reached at the lowest dose tested (0.05 μmol/L), with a GI₅₀ < 130 nmol/L. In contrast, the low-sensitive cell lines (Z-138, MAVER-1, JVM-2, and HBL-2) did not undergo a marked proliferation decrease after everolimus treatment at doses below 5 μmol/L. In these cell lines, the GI₅₀ at 72 hours could not be reached (Table 1). Consistently, cell-cycle analysis of 3 representative cell lines revealed a blockade in G₀–G₁ phase in the sensitive cells REC-1 and GRANTA-519, whereas the low sensitive Z-138 cells showed only a slight increase in the S phase cell fraction (Fig. 1B).

MCL primary cells showed a high sensitivity to the compound as a cytotoxic effect was detected in the majority of the samples as soon as 48 hours of treatment, contrasting with the low cell death rates observed in the cell lines, even after a 72-hour exposure to everolimus (Fig. 1C and Table 1). Five cases (MCL no. 7, 8, 9, 10, and 11) were sensitive to everolimus, with a mean LD₅₀ of 3.84 ± 0.15 μmol/L, whereas 6 cases (MCL no. 1, 2, 3, 4, 5, and 6) showed a lower response to the drug (LD₅₀ > 5 μmol/L; Table 1). This cytotoxic effect was shown to be selective as neither normal B (CD19⁺) nor normal T lymphocytes (CD3⁺) from healthy donors were notably affected by 5 μmol/L everolimus (Fig. 1C).

Of note, there was no association between MCL sensitivity to everolimus and common cytogenetic alterations such as deletions of P53, ATM, or P16, either in MCL cell lines or primary cells (Table 1).

Thus, these results show that everolimus induces apoptosis in the majority of MCL primary samples, at physiologically achievable doses, sparing normal B and T cells.

To ensure that everolimus activity in MCL cell lines and primary MCL cells was linked to mTOR pathway inhibition, REC-1, GRANTA-519, Z-138 cells, and a representative MCL primary sample were treated with everolimus (doses ranging from 0.05 to 5 μmol/L) for up to 72 hours. Fig. 2A and B shows that short time incubation (24 hours) with everolimus in REC-1 and GRANTA-519 cells caused a dose-dependent decrease in the phosphorylation levels of Akt, mTOR, and its downstream targets, S6RP and 4E-BP1. About 4E-BP1, while the hyperphosphorylated isoforms decreased, the hypophosphorylated form accumulated after everolimus treatment. Interestingly, part of these effects was not sustained and after prolonged exposure to everolimus (48–72 hours), we observed a rephosphorylation of Akt and mTOR. In contrast, in the low-sensitive Z-138 cell line, everolimus only induced a transient downregulation of p-S6RP and p-4E-BP1 proteins, with no substantial changes in the levels of phospho-mTOR and a slight decrease in Akt phosphorylation at the higher dose and time of exposure (Fig. 2C). Importantly, everolimus efficiently inhibited mTOR activity in a representative sensitive MCL primary sample (Fig. 2D), as established by lower levels of phosphorylated mTOR, Akt, S6RP, and 4E-BP1, without observing compensatory reactivation of Akt after 48 hours of incubation.

Altogether, these data indicate that although everolimus inhibits mTOR pathway at short time treatment, Akt rephosphorylation after prolonged exposure could limit its efficacy, consistent with the existence of feedback loops within the mTOR signaling pathway, as described previously (14, 19).

To investigate the impact of the Akt rephosphorylation in MCL cells exposed to everolimus, REC-1, GRANTA-519, and Z-138 cell lines were pretreated with or without 2 or 5 μmol/L of the Akt inhibitor Akti-1/2 for 3 hours and incubated with everolimus at 0.5 and 5 μmol/L for 24 additional hours. As shown on Fig. 3A, MCL cell viability was almost unaffected by Akti-1/2, whereas its addition to everolimus allowed to reduce cell viability more efficiently than single drug treatment. We found the combination of everolimus 5 μmol/L with Akti-1/2 5 μmol/L to be synergistic in REC-1 and GRANTA-519 cells, with respective CI values of 0.691 and 0.602. In contrast, this combination failed to induce cytotoxicity in the everolimus-low sensitive cell line Z-138 (Fig. 3A), although it allowed a synergistic cytostatic effect, as monitored by MTT assay (Supplementary Fig. S1). Of note, the combination of both agents was synergistic in all the MCL primary cells tested (n = 8), although a significant increase (P < 0.001) in cell death induction was only observed in the subset of MCL cases sensitive to everolimus single agent, with a mean CI value of 0.229 (Fig. 3B).

In an attempt to analyze the mechanisms underlying the combinatory effect of Akti-1/2 and everolimus, both REC-1 and Z-138 cells were treated for 24 and 48 hours with everolimus and/or Akti-1/2, followed by Western blot of Akt/mTOR pathway status. Figure 3C shows that Akti-1/2 efficiently blocked Akt phosphorylation in both cell lines, including when added to everolimus. This effect, combined with everolimus inhibitor activity against the mTOR targets p-S6RP and p-4E-BP1, led to the activating processing of the effector caspase-3 and degradation of the caspase substrate PARP in REC-1, but not in Z-138 cells, indicating a defect in apoptosis induction after Akt inhibition in Z-138 cells (Fig. 3C).

These results show a synergistic interaction between everolimus and the Akt inhibitor Akti-1/2 in MCL cells, suggesting a role for Akt signaling in MCL resistance to mTOR targeting.

The Akt/mTOR pathway has been implicated in the regulation of autophagy in several models of cancer (20, 21). To explain the heterogeneous efficacy of everolimus/Akti-1/2 combination in MCL cells, we assessed the degree of autophagy induction in Z-138 and REC-1 cells. Western blot analysis of the autophagy-initiating protein LC3B showed that everolimus/Akti-1/2 treatment led to the processing of LC3B-I to LC3B-II, indicative of an increase in the autophagic activity (Fig. 4A). Importantly, this phenomenon was mainly observed in the everolimus/Akti-1/2 low-responsive Z-138 cells (ratio LC3B-II/α-tubulin of 5.8), compared with the sensitive cells REC-1 (ratio LC3B-II/α-tubulin of 2.4). As LC3B-II is known to specifically associate with the autophagosome membrane, we then studied the production of autophagic vacuoles in REC-1 and Z-138 cells. Both cell lines were pretreated with Akti-1/2 (5 μmol/L, 3 hours) and incubated with everolimus. After 24 hours of treatment, cells were stained with the autophagy Cyto-ID Green dye, and then analyzed by flow cytometry. Tamoxifen (22) and bafilomycin A1-treated cells (23, 24) were used as a positive and negative control for autophagolysosome detection, respectively (Fig. 4B). In REC-1 cells, everolimus/Akti-1/2 treatment provoked a slight green fluorescence increase, indicative of some levels of autophagy induction. In contrast, in Z-138 cells, everolimus and Akti-1/2 alone, but especially the combination of both drugs, caused a 5-fold increase in mean fluorescence intensity, revealing an enhanced autophagolysosome formation by dual Akt/mTOR targeting.

To provide further evidence of the autophagy stimulation after everolimus/Akti-1/2 treatment, we studied the accumulation of Cyto-ID Green stained autophagolysosomes by fluorescence microscopy (Fig. 4C). Untreated control cells harbored a diffuse distribution of green fluorescence throughout the cytoplasm, whereas treated cells presented punctuate structures corresponding to the autophagic vacuoles. As shown in Fig. 4C, we observed a substantial autophagy induction after Akti-1/2 treatment in both REC-1 and Z-138 cells (arrows), but the addition of everolimus notably increased the autophagic flux (arrows). Consistently, this effect was more pronounced in the everolimus/Akti-1/2 low-responsive Z-138 cells than in the sensitive cells REC-1.

To ascertain if this increase in autophagy could promote MCL cell survival and drug resistance, we used a siRNA-mediated approach to simultaneously knockdown the ubiquitin-like activating enzyme homolog ATG7, the ubiquitin folds-containing protein ATG5, and the ubiquitin-conjugating enzyme analog ATG3, all involved in LC3 activation. ATG7/ATG5/ATG3 triple silencing reduced the basal protein levels of these 3 factors, thereby decreasing LC3B-II expression and significantly enhancing cell death after everolimus/Akti-1/2 treatment in Z-138 cell line (P < 0.001, Fig. 4D). A sensitizing effect, albeit reduced, was also observed in REC-1 cells (data not shown), emphasizing the role of autophagy in Z-138 resistance to everolimus/Akti-1/2 combination.

These findings highlight the contribution of autophagy to the resistance to everolimus/Akti-1/2 treatment in MCL cells.

In an attempt to counteract the protective effect of autophagy toward everolimus/Akti-1/2 treatment, we assessed the effect of adding an autophagy inhibitor to increase the antitumoral effect of the combination. We used hydroxychloroquine, an antimalarial drug that blocks lysosome acidification and degradation of autophagosomes, causing an accumulation of autophagic vacuoles (25). REC-1 and Z-138 cells were preincubated with hydroxychloroquine 10 μmol/L for 1 hour, then pretreated with 5 μmol/L Akti-1/2 for 3 hours and finally incubated with 5 μmol/L everolimus for 24 hours. Although this triple combination led to a significant induction of apoptosis in both MCL cell lines, this effect was highly significant (P < 0.001) in the everolimus/Akti-1/2 low sensitive Z-138 cells (Fig. 5A). As expected, hydroxychloroquine addition to the everolimus/Akti-1/2 treatment provoked an increase of LC3B-II expression in both cell lines (Fig. 5B). Moreover, the triple combination increased PARP cleavage in REC-1 cells and also in Z-138 cells, where the everolimus/Akti-1/2 treatment did not exert any cytotoxic effect (Fig. 5B). Similar results were obtained with bafilomycin A1, another autophagy inhibitor (data not shown).

We then tested the effect of combining hydroxychloroquine with the Akt/mTOR inhibitors in MCL primary cases. Although hydroxychloroquine addition synergized with Akti-1/2 and everolimus in all MCL cases tested, only those cases with poor response to the everolimus/Akti-1/2 combination had a significant increase in cell death (P < 0.01; Fig. 5C).

Accordingly, while combination of everolimus and Akti-1/2 slightly activated the mitochondrial apoptotic pathway in low responsive MCL cells (Z-138 and a representative MCL primary sample), the addition of hydroxychloroquine enhanced the typical mitochondrial hallmarks of apoptosis, including mitochondrial depolarization, ROS production, caspase-3/7 activity, and PS exposure (Fig. 5D).

Taken together, these results show that the addition of an autophagy inhibitor overcomes the resistance of MCL cells to Akt/mTOR inhibitors, leading to efficient apoptosis induction and suggesting a prosurvival role of autophagy.

The PI3K/Akt/mTOR axis is known to be constitutively activated in the majority of B-cell lymphomas. In these cancers, mTOR-activating events may include loss of PTEN function, leading to constitutive activation of Akt, constitutional or growth factor–induced stimulation of receptor tyrosine kinases, or overexpression of eIF4E (26). Specifically, in the blastoid variant of MCL, high levels of phosphorylation on Akt at Ser473 are frequently observed (27).

To target this pathway, several rapamycin analogs have shown activity against lymphoma cells both in vitro and in vivo (28). Here, using an extended panel of MCL cell lines, we confirm previous reports showing that the rapalog everolimus exerts an antiproliferative effect in MCL cell lines, mediated by cell-cycle blockade at G₁ phase (29). We also show for the first time that this agent induces a tumor-selective, dose-dependent cytotoxicity in the majority of primary MCL cases, in the range of doses achievable in vivo. As expected, we show that everolimus efficiently inhibits mTOR activity in sensitive cell lines and primary samples, as attested by a decrease in the phosphorylation levels of the mTOR downstream targets S6RP and 4E-BP1 at short-time treatment, and a transient dephosphorylation of Akt at Ser473. Importantly, we observe that prolonged mTORC1 inactivation leads to Akt rephosphorylation at this residue, a phenomenon described as a crucial determinant of malignant cell response to rapamycin-like drugs (14, 19). Suggesting that this late reactivation of Akt may counteract everolimus activity, we report that the combination of the rapalog to an isoselective Akt inhibitor exerts synergistic antitumoral activity in all the samples tested, especially in those cases with high sensitivity to everolimus single agent. In these cells, we show the effect of the combination to be mediated by activation of the intrinsic apoptotic program, whereas everolimus alone appears to be a weak apoptosis inducer, as previously reported for rapamycin (30).

Several evidences suggest that the synergism between Akt and mTOR dual targeting may rely on the blockade of the mTORC1-dependent response loop after mTOR inhibition, which has been shown to involve reactivation of upstream receptor kinase signaling within the IGFR pathway, as well as Akt itself, in both in vitro and in vivo models of human cancers (5, 31, 32). Emerging from this observation, the use of dual PI3K/mTOR inhibitors has shown to be useful not only to downregulate the mTOR targets 4E-BP1 or S6RP, but also to prevent Akt rephosphorylation at Ser473, independent of PI3K mutation status (33). In line with our results, the complete abolition of Akt/mTOR signaling by means of these dual inhibitors has been shown to exert increased cytostatic effect and to lead to apoptotic cell death in PI3K/Akt/mTOR-addicted lymphomas (34, 35), thus highlighting the requirement of a full inhibition of the pathway for improved antitumoral activity. However, we observed that those MCL samples that are weak responders to everolimus single agent still harbor a high viability rate despite the complete Akt/mTOR axis inhibition, similar to a previous observation in follicular lymphoma (35). Among the resistance mechanisms that may account for this defective apoptosis initiation, accumulating evidences suggest that autophagy is one of the major process functionally involved in cancer cell survival after Akt and mTOR inhibitor exposure (36).

Autophagy is an evolutionarily conserved intracellular self-defense mechanism, which serves to maintain cellular metabolism through recycling of cellular components when the availability of external nutrient sources is limited (36). Although constitutive autophagy is a homeostatic mechanism for metabolic regulation, it is also stress responsive, and through the removal of damaged proteins and organelles, it confers stress tolerance and sustains viability under adverse conditions (37). mTOR kinase inhibition has been shown to induce a prosurvival autophagy that can counteract the effect of common chemotherapeutics in colon carcinoma (38). In MCL cells, the rapalog temsirolimus has also been shown to activate autophagic processing, although the contribution of this phenomenon to the activity of the mTOR inhibitor was not clearly discussed (39). Our results show that 2 hallmarks of autophagy, LC3B processing and autophagic vacuoles formation, are enhanced in MCL cells resistant to everolimus/Akti-1/2 combination, when compared with sensitive cells. More important, we show that autophagy controls MCL response to mTOR/Akt inhibitors, as knockdown of ATG7, ATG5, and ATG3, all 3 proteins required for the progression of autophagy, allows MCL cells to undergo apoptosis upon exposure to the combination. Accordingly, it has been described that dual inhibitors of PI3K/mTOR kinases may induce autophagy as a central survival signal (40), pointing out that effective cell death in malignant cells with constitutive Akt activation would require the blockade of the 3 targets we describe herein: mTOR, Akt, and autophagy.

As autophagy is generally a survival pathway used by tumor cells to tolerate metabolic stress (22, 41), autophagy inhibitors, in combination with other agents, are expected to efficiently target therapy-resistant tumor cells in hypoxic tumor regions (37). Compounds such as hydroxychloroquine or chloroquine that block lysosome acidification and consequent autophagosome fusion, have been entered in a number of clinical trials in combination with standard or experimental agents (42). In hematologic malignancies, these agents are being tested in combination with the proteasome inhibitor bortezomib in multiple myeloma (NCT00568880; ref. 36) or with the histone deacetylase inhibitor vorinostat in chronic myeloid leukemia (43). These latest studies sustain the concept that autophagy is a mechanism of therapeutic resistance, and that hydroxychloroquine can increase cytoxicity by abrogation of autophagy. In accordance with this, we found that the addition of this agent to everolimus/Akti-1/2 combination fully activates the intrinsic apoptotic program in MCL cells primarily resistant to Akt/mTOR dual targeting.

In summary, we show for the first time that the use of an autophagy inhibitor may overcome resistance to the combination of everolimus and an isoselective Akt inhibitor in MCL cell lines and primary samples. The proposed prosurvival role of autophagy in Akt/mTOR compromised cells points out some potential opportunities and warrant further clinical activity of this triple combinational strategy in MCL patients.

No potential conflicts of interest were disclosed.

Conception and design: L. Rosich, D. Colomer, G. Roué

Development of methodology: L. Rosich, M. Lopez-Guerra, D. Colomer, G. Roué

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Rosich, D. Colomer, G. Roué

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L. Rosich, S. Xargay-Torrent, E. Campo, D. Colomer, G. Roué

Writing, review, and/or revision of the manuscript: L. Rosich, M. Lopez-Guerra, E. Campo, D. Colomer, G. Roué

Study supervision: D. Colomer, G. Roué

Revision of the final version of the manuscript: S. Xargay-Torrent, G. Roué

The authors thank Jocabed Roldán, Laura Jiménez, and Sandra Cabezas for expert technical assistance and Anna Bellmunt for her help in in vitro studies. Everolimus was kindly provided by Novartis. This work was carried out, in part, at the Esther Koplowitz Center, Barcelona.

Fondo de Investigación Sanitaria (PI09/0060; to G. Roué), Ministerio de Ciencia e Innovación (SAF 09/9503; to D. Colomer), Redes Temáticas de Investigación Cooperativa de Cáncer from the Instituto de Salud Carlos III (RED 2006-20-014 to D. Colomer and 2006-20-039 to E. Campo), and Generalitat de Catalunya (2009SGR967 to D. Colomer). L. Rosich and S. Xargay-Torrent are recipients of predoctoral fellowships from IDIBAPS and Ministerio Ciencia e Innovación (FPU), respectively. M. López-Guerra has a contract from RED 2006-20-014.

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.