PTEN posttranslational inactivation and hyperactivation of the PI3K/Akt pathway sustain primary T cell leukemia viability

The PI3K/Akt pathway is constitutively hyperactivated in primary T-ALL cells.

Based on evidence from T-ALL cell lines, we hypothesized that PI3K/Akt signaling is hyperactivated in primary disease. To establish how frequently PI3K/Akt constitutive activation actually occurs, we evaluated the integrity of the PTEN/PI3K/Akt axis in T-ALL patient samples collected at diagnosis. Constitutive hyperactivation of the PI3K/Akt pathway was detected in most T-ALL specimens (87.5%; 21 of 24), as evaluated by increased phosphorylation of Akt and/or at least one of its downstream targets GSK-3β and FOXO3a in comparison to normal T cell precursors (Figure 1, A and B, Table 1, and data not shown). Hyperactivation of the PI3K/Akt pathway was not associated with overexpression of total Akt (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI34616DS1) and was consistently observed independently of the stage of developmental arrest of the T-ALL samples (Table 1). Moreover, primary T-ALL cells presented higher PIP3 levels in the plasma membrane, as determined by both confocal microscopy and flow cytometry analysis (Figure 1, C–E). Consistent with previous reports (15, 33) we also observed PI3K/Akt hyperactivation in leukemia T cell lines (Supplemental Figure 2).

Each T-ALL sample essentially consists of a clonal population of leukemia blasts arrested at a particular stage of thymic T cell development (34, 35), whereas normal thymocytes are immunophenotypically and functionally heterogeneous, covering the spectrum of distinct developmental subpopulations. To ensure that our data were not biased by the prevalence of a particular subpopulation in the normal control samples, we next analyzed patient specimens arrested at 2 particular stages of T cell development and compared them with their purified normal equivalents. We selected patients with CD3⁺CD4⁺CD8⁺ (triple positive; TP) and CD3⁺CD4^–CD8⁺ (CD8 single positive; SP8) immunophenotypes because they were the most prevalent among primary T-ALL samples (Table 1 and Supplemental Figure 3A). Additionally, normal TP and SP8 thymocytes represent 2 distinct subpopulations: TP cells are mostly quiescent, whereas SP8 cells are actively proliferating or have recently done so (36). Both TP and SP8 T-ALL cells revealed clearly higher levels of phosphorylated Akt than their normal immunophenotypic counterparts (Supplemental Figure 3B). Likewise, both CD3^– and CD3⁺ T-ALLs displayed hyperphosphorylation of Akt compared with immunophenotype-matched normal thymocytes (data not shown). Also, PIP3 levels were similar among the different normal thymic subsets and significantly and consistently lower than those of primary T-ALL cells (Supplemental Figure 4). Our data indicate that constitutive hyperactivation of the PI3K/Akt pathway is a common event in T-ALL, not only in immortalized cell lines, but more importantly in primary tumors.

PTEN gene alterations are infrequent, whereas protein expression is commonly deregulated, in primary T-ALL.

PTEN dephosphorylates PIP3 and is the main negative regulator of the PI3K/Akt pathway. Thus, we next analyzed the expression of PTEN mRNA and protein in primary T-ALL samples. Whereas all 23 samples analyzed expressed PTEN mRNA, a minority of the cases (20%; 6 of 30) lacked PTEN protein (Figure 2A and Table 1). Notably, not only the PTEN protein–negative samples (n = 5), but also most of the PTEN protein–positive samples (84.2%; 16 of 19), showed PI3K/Akt pathway hyperactivation (Table 1). Because point mutations may render PTEN protein inactive with unnoticeable effects on protein expression, we performed mutation analysis of all PTEN exons and flanking intron sequences in 3 PTEN protein–negative and 15 PTEN protein–positive T-ALL cases. Interestingly, all 3 PTEN protein–null samples analyzed showed PTEN gene alterations involving exon 7 (Figure 2C and Table 2), which is mutated in Jurkat cells and is essential for PTEN protein stability (15, 37). One of the samples was also affected in exon 1, which appears to be involved in PTEN membrane binding and activation (38). None of the 15 PTEN protein–positive T-ALLs analyzed presented PTEN mutations in exons 1–9 or in the flanking intronic splice site sequences. Our data indicate that relatively few primary T-ALL samples harbor PTEN gene alterations, in accordance with recent reports that PTEN mutations are rare in primary T cell leukemia specimens (33, 39). In addition, despite some heterogeneity, average PTEN mRNA expression in primary T-ALL cells was comparable to that in normal T cell precursors, as assessed by quantitative RT-PCR (Figure 2B). This observation suggests that PTEN transcription is largely normal in T-ALL, arguing against the possibility that PTEN promoter hypermethylation (40) or Notch-dependent PTEN transcriptional repression (39) plays a broad role in primary disease. Finally, we assessed whether hyperactivation of the PI3K/Akt pathway in PTEN-positive T-ALL cells is caused by decreased PTEN protein abundance. Unexpectedly, we found that PTEN-positive T-ALL samples expressed even higher levels of PTEN protein than did normal controls (Figure 2D), despite displaying constitutive activation of the PI3K/Akt pathway. Similar results were found for PTEN-positive T-ALL cell lines (Supplemental Figure 2). Analysis of PTEN protein expression by flow cytometry revealed no significant differences among the 4 major normal thymic subpopulations (Supplemental Figure 5, A–C). Furthermore, each subpopulation had lower PTEN levels than those of the T-ALL cell line TAIL7 (41), a good representative of PTEN-positive primary T-ALL T cells (Supplemental Figure 5, D–F, and Figure 2A). In addition, immunoblot analysis of TP and SP8 T-ALL cells and normal thymocytes isolated by fluorescence-activated cell sorting (FACS) showed once again that the T-ALL cells expressed higher PTEN protein levels than did their normal counterparts (Supplemental Figure 5G).

CK2 regulates PTEN expression and activity in PTEN protein–positive T-ALL cells.

The apparent contradiction between PTEN protein abundance and PI3K/Akt pathway activation in T-ALL samples prompted us to analyze PTEN phosphatase activity. We found that high PTEN expression was not associated with increased activity, but rather with PTEN inactivation, since T-ALL samples showed diminished PTEN in vitro lipid phosphatase activity compared with normal thymocytes (Figure 3A). Furthermore, most malignant specimens showed higher PTEN phosphorylation at different residues in the C-terminal tail (S380, S370, and the cluster including T382/T383/S385) than did normal controls (Table 1, Figure 3B, and data not shown). Thus, most T-ALL cells expressed higher levels of both PTEN (Figure 2A) and phosphorylated PTEN (Figure 3B). Importantly, the ratio of phosphorylated PTEN to total PTEN was significantly higher in phosphorylated PTEN–high T-ALL cases than in normal thymocytes (Supplemental Figure 6). This finding indicates that the higher amounts of phosphorylated PTEN in T-ALL cells are not simply the result of increased PTEN protein availability. Because CK2-mediated phosphorylation of PTEN at the C-terminal tail was previously proposed to downmodulate PTEN activity while increasing its stability (20, 21), we next examined expression and activation of CK2. T-ALL cells expressed higher levels of both the catalytic CK2α and regulatory CK2β subunits compared with normal thymocytes (Table 1, Figure 3, C–E, and Supplemental Figure 7) and showed clearly increased constitutive CK2 kinase activity (Figure 3F). Treatment of TAIL7, HPB-ALL, or primary T-ALL cells with the CK2-specific inhibitor tetrabromobenzotriazole (TBB) resulted in abrogation of C-terminal phosphorylation of PTEN, downregulation of PTEN expression, and simultaneous inhibition of Akt and GSK-3β phosphorylation (Figure 3G). These results suggest that PTEN phosphorylation — leading to PTEN protein stabilization and inactivation — and consequent PI3K/Akt pathway hyperactivation might be secondary to constitutive increase of CK2 activity in T-ALL cells. Because it was previously shown that CK2 may stimulate Akt activity directly (42), we next analyzed the effect of TBB on phosphorylation of Akt in PTEN-null Jurkat T-ALL cells stably expressing PTEN in a Tet-inducible manner (16). CK2 inhibition clearly diminished the levels of Akt S473 phosphorylation in Jurkat cells that expressed PTEN upon doxycycline treatment, whereas it induced minor downregulation of Akt phosphorylation in Jurkat cells that did not express PTEN (Supplemental Figure 8). These data suggest that the effect of CK2 on PI3K/Akt pathway activity in T-ALL cells is largely dependent on the ability of CK2 to regulate PTEN activity. To confirm this, we analyzed the effect of TBB directly on PTEN phosphatase activity. CK2 inhibition dramatically upregulated PTEN activity in TAIL7 (Figure 3H) and HPB-ALL cells (data not shown), suggesting that constitutive aberrant CK2 activity contributed to PI3K/Akt pathway activation, at least in part, by nondeletional posttranslational inactivation of PTEN. Consistent with this hypothesis, patient samples without PI3K/Akt pathway hyperactivation displayed levels of PTEN phosphorylation similar to those of normal thymocytes (e.g., T-ALL patient 14, Figure 3B). Moreover, high CK2 expression in PTEN-positive T-ALL samples was associated with increased phosphorylation of PTEN (Supplemental Figure 9 and Table 1).

High ROS levels downregulate PTEN activity in PTEN protein–positive T-ALL cells.

Cancer cells frequently express high levels of ROS, including superoxide anion (30) and H₂O₂ (29). By using the redox-sensitive dye DCF-DA, we found that the majority of primary T-ALL cells had higher intracellular ROS levels than did normal thymocytes (Table 1, Figure 4, A and B, and Supplemental Figure 10). Because H₂O₂ was previously shown to oxidize and thereby inactivate PTEN (18, 19, 32), we sought to determine whether diminished PTEN activity in T-ALL cells is also associated with increased ROS. Although most PTEN present in TAIL7 cells was in the reduced form, there were constitutively detectable levels of oxidized PTEN, which were further upregulated by addition of exogenous H₂O₂ and abrogated by in vitro treatment with the reducing agent dithiothreitol (DTT; Figure 4C). Treatment of PTEN-positive T-ALL cells with another antioxidant, β-mercaptoethanol (β-ME), downregulated constitutive phosphorylation of Akt and GSK-3β (Figure 4D and Supplemental Figure 11A), whereas H₂O₂ induced the opposite effect (Figure 4E and Supplemental Figure 11B). In contrast, Akt and GSK-3β phosphorylation levels were not significantly affected by modulation of ROS in PTEN-null T-ALL Jurkat cells (Figure 4, D and E), which suggests that oxidation-dependent activation of the PI3K/Akt pathway largely relied on regulation of PTEN. In accordance, the PI3K-specific chemical inhibitor LY294002 abrogated not only constitutive, but also H₂O₂-promoted, Akt phosphorylation, which indicates that ROS upregulated phosphorylated Akt in a PTEN/PI3K/PIP3-dependent manner and not via alternative mechanisms acting directly on Akt (Figure 4F). To further support the evidence that PTEN activity contributes to ROS-mediated activation of the PI3K/Akt pathway, we performed an indirect PTEN redox assay (19), in which lysis buffers alkylate reduced cysteine residues, thereby irreversibly inactivating PTEN. Oxidized cysteine residues are protected from alkylation, and, because oxidation is reversible, they can be recovered afterward by treatment with a reducing agent for assessment of PTEN activity. The resultant in vitro phosphatase activity reflects the proportion of PTEN that was oxidized and inactivated at the time of lysis. Using this approach, we determined that β-ME (Figure 4G) and H₂O₂ (Figure 4H) modulated endogenous PTEN redox state and activity in opposite directions. Importantly, addition of β-ME to T-ALL cells reverted constitutive oxidation–induced inactivation of PTEN. These data suggest that high levels of intracellular ROS contributed to PTEN inactivation by oxidation with subsequent PI3K/Akt pathway hyperactivation. Furthermore, combination of low doses of the antioxidant N-acetyl-cysteine (NAC) and the CK2 inhibitor TBB acted synergistically in inhibiting the PI3K/Akt pathway (Supplemental Figure 12). This effect did not appear to result from possible upregulation of CK2 by increased ROS levels in T-ALL cells, since treatment of TAIL7 and HPB-ALL cell lines with β-ME did not significantly affect the expression of CK2α/β (Supplemental Figure 13).

Targeting PI3K/Akt pathway hyperactivation in T-ALL.

Malignant cell lines lacking PTEN activity as a result of PTEN gene and protein deletion are particularly sensitive to PI3K/Akt pathway inhibition (43). Our present data demonstrated that although they generally expressed PTEN, primary T-ALL cells frequently showed functional inactivation of PTEN phosphatase activity and constitutive activation of the PI3K/Akt pathway. Hence, we next analyzed the effect of the PI3K inhibitor LY294002 on the viability of T-ALL cells. In contrast to normal thymocytes, primary T-ALL cells underwent significant death upon LY294002 treatment in vitro (Figure 5, A and B, and Supplemental Figure 14). The effect of LY294002 on T-ALL cells did not appear to be solely dependent on the expression of PTEN, because it affected not only PTEN-negative but also PTEN-positive patient samples (Figure 5C) and cell lines (Supplemental Figure 15). Moreover, PTEN-positive patient samples with hyperactivation of the PI3K/Akt pathway (e.g., T-ALL patient 19; Figure 5D) were clearly sensitive to PI3K inhibition (Figure 5, F and G). In contrast, PTEN-positive samples without constitutive activation of the PI3K/Akt pathway (e.g., T-ALL patient 08; Figure 5E) were unresponsive to LY294002 treatment (Figure 5, F and G). As expected, these samples showed low levels of PTEN phosphorylation (e.g., T-ALL patient 08, Figure 5E; and T-ALL patient 14, Figure 3B), indicative of normal PTEN activity, in contrast to the samples with hyperactivation of the PI3K/Akt pathway (e.g., T-ALL patients 4, 19, and 22, Figure 5, D and E) that displayed hyperphosphorylation of PTEN. These results suggest that the sensitivity of T-ALL cells to PI3K inhibition is determined by PTEN functional integrity and consequent activation status of the PI3K/Akt pathway, rather than the level of PTEN expression per se.

Because abrogation of CK2 activity leads to PTEN activation and PI3K/Akt pathway inhibition, we next treated T-ALL cells with 2 distinct CK2 inhibitors to assess their impact on cell viability. Both TBB and dichlororibofuranosylbenzimidazole (DRB) induced significant cell death in primary T-ALL cells but not in normal thymocytes (Figure 5H). Because we showed that CK2-mediated activation of the PI3K/Akt pathway was largely dependent on regulation of PTEN activity, we next compared the effect of CK2 inhibition on the viability of PTEN-positive versus PTEN-negative Jurkat T-ALL cells. Although TBB and DRB promoted death in PTEN-null T-ALL cells, the effect of both CK2 inhibitors was significantly higher in the cells that expressed PTEN (Supplemental Figure 16). This suggests that PTEN-positive T-ALL cells with hyperactivation of the PI3K/Akt pathway may especially benefit from therapies that include pharmacological inhibitors of CK2. Similarly, the viability of PTEN-expressing Jurkat cells was strikingly diminished upon treatment with β-ME, whereas PTEN-null Jurkat cells were insensitive to the reducing agent (Supplemental Figure 16).

Finally, to evaluate potential cooperative antileukemic effects, we performed experiments combining CK2 inhibition with ROS scavenging. CK2 blockade by TBB cooperated with the antioxidants β-ME and NAC to markedly decrease T-ALL T cell viability, with more than 95% cell death (Figure 6, A and B). We then used each of these agents together with the PI3K antagonist LY294002. Addition of LY294002 markedly increased the antileukemic activity of CK2 inhibitors TBB and DRB (Figure 6, C and D). Finally, PI3K blockade also potentiated the inhibitory effects of ROS scavenging by either β-ME or NAC (Figure 6, E and F). Similar cooperative effects were observed using HPB-ALL and primary T-ALL cells (Supplemental Figures 17 and 18). Together, these data suggest that combinatorial strategies targeting multiple elements related to PTEN regulation and PI3K/Akt activation are promising therapeutic approaches in T-ALL.

FACS of thymic subpopulations, PTEN intracellular staining, and levels of PIP3 and intracellular ROS in surface-stained cells.

See Supplemental Methods.

Primary samples and T-ALL cell lines.

Institutional review board approval was obtained for all primary leukemia and thymic sample collections from Comitê de ωtica em Pesquisa, Centro Infantil Boldrini (Campinas, São Paulo, Brazil), Gabinete de Investigação Clínica, Instituto Português de Oncologia (Lisbon, Portugal), Comissão de ωtica, Centro Hospitalar de Lisboa Occidental (Carnaxide, Portugal), and Comissão de ωtica, Faculdade de Medicina da Universidade de Lisboa (Lisbon, Portugal). All subjects provided informed consent. T-ALL cells were obtained from the peripheral blood and/or the bone marrow of patients with high leukemia involvement of 85%–100%. Samples were enriched by density centrifugation over Ficoll-Paque (GE Healthcare), washed twice in culture medium (RPMI-1640 supplemented with 10% FBS), subjected to immunophenotypic analysis by flow cytometry, and classified according to their maturation stage (Table 1). Normal thymocytes were isolated from thymic tissue obtained from children undergoing cardiac surgery. The tissue was gently minced in medium, the resulting cell suspension was filtered through a cell strainer, and thymocytes were enriched by density centrifugation to greater than 95% purity. The PTEN-positive T-ALL cell line TAIL7 shares significant similarities with primary leukemia samples (41). The cell lines Jurkat, Molt4, CEM (PTEN-negative), and HPB-ALL (PTEN-positive) have been extensively described (60–63). Jurkat PTEN Tet-inducible clones were provided by A. Weiss (UCSF, San Francisco, California, USA).

Immunoblotting.

Cell lysates were resolved by 10% SDS-PAGE, transferred onto nitrocellulose membranes, and immunoblotted with the following antibodies: ZAP70, p-PTEN (S370), p-PTEN (S380/T382/T383/S385) (Upstate), actin, CK2α, CKβ (Santa Cruz Biotechnology Inc.), Akt, p-Akt (S473), p–GSK-3β (S9), p-FOXO3a (T32), PTEN, and p-PTEN (S380) (Cell Signaling Technology). Where indicated, densitometry analysis was performed using Image Quant software (version 5.2). Each band was analyzed with a constant frame and normalized to the respective loading control.

Detection of PIP3 levels.

Cells were washed in PBS and fixed in 2% paraformaldehyde, blocked with 5% normal goat serum in PBS, and incubated overnight at 4°C with mouse anti-PIP3 monoclonal antibody (Echelon) at 1:10 dilution. Primary antibody was detected with Alexa Fluor 488–conjugated goat anti-mouse antibody (Invitrogen). Samples were split and (a) imaged by confocal microscopy using a Zeiss LSM 510 Meta microscope, using identical acquisition parameters in 1 imaging session, or (b) analyzed by flow cytometry to quantify mean fluorescence intensity using FACSCanto flow cytometer (BD Biosciences) and FlowJo software (Tree Star).

RT-PCR and quantitative RT-PCR.

Total RNA (1 μg) was reverse transcribed using the ImProm II Reverse Transcriptase enzyme (Promega) and random hexamers. PTEN expression was evaluated by RT-PCR using the following primers: forward, CCAAGTCCAGAGCCATTTC; reverse, GTGGGTCCTGAATTGGAGG. The quantitative assessment of PTEN transcripts was made by quantitative RT-PCR on an ABI PRISM 7500 (Applied Biosystems) with the following primers: ABL, TGGAGATAACACTCTAAGCATAACTAAAGGT and GATGTAGTTGCTTGGGACCCA; PTEN, GCTACCTGTTAAAGAATCATCTGG and CATGAACTTGTCTTCCCGT. PTEN primers were designed to provide the most possible difference to the PTEN pseudogene transcript. PCR products were cloned into the pGEM-T Easy vector (Promega), and standard curves were obtained by serial dilutions of uncut plasmid. PTEN transcript values were normalized with respect to the number of ABL transcripts.

PTEN sequencing and mutational analysis.

The entire PTEN coding and flanking intronic sequences were amplified by PCR. Primers (provided by A. Vettore, Ludwig Institute for Cancer Research, São Paulo, Brazil) are listed in Supplemental Table 1. PCR reactions were done using 100 ng DNA and 0.5 U TripleMaster Polymerase (Eppendorf). The PCR products were sequenced directly, on both strands, using the BigDye kit (Applied Biosystems). Sequences were evaluated using Chromas Lite (version 2.0; Technelysium) and Mutation Explorer (version 2.6.1; SoftGenetics LLC) software. All mutations were confirmed by sequencing a newly amplified product. When necessary, PCR amplicons were cloned into pGEM-T vetor (Promega), and several individual clones were sequenced.

Endogenous PTEN in vitro lipid phosphatase assay.

Immunoprecipitations were carried out with an anti-PTEN antibody (Santa Cruz Biotechnology Inc.) overnight and a secondary agarose-conjugate antibody for 3 hours at 4°C. Immunoprecipitated protein was washed, resuspended in enzyme reaction buffer (50 mM Tris, pH 8; 50 mM NaCl; 10 mM DTT; and 10 mM MgCl₂), and incubated with 10 μM PIP3 (Echelon) for 30 minutes at 37°C, after which phosphatase reaction was stopped with 100 μl malachite green reagent (Echelon). Free phosphate levels were measured in an ELISA plate reader at 620 nm. Absorbance was converted into pmol phosphate using a phosphate standard curve.

Endogenous CK2 kinase activity assay.

CK2 activity in cell lysates was measured using an appropriate kit from Upstate Biotechnology following the manufacturer’s instructions.

Intracellular ROS levels.

Cell samples were washed in PBS, incubated with 10 μM DCF-DA (Sigma-Aldrich) for 30 minutes in PBS at 37°C, washed, acquired on a FACSCalibur flow cytometer (BD Biosciences), and analyzed with FlowJo software.

Detection of PTEN oxidized form.

Cells were treated with or without 1mM H₂O₂ in PBS for 30 minutes at 37°C with 5% CO₂. Whole cell lysates were prepared in nonreducing lysis buffer (100 mM Tris-Hcl, pH 6.8; 2% SDS; and 50 mM iodoacetamide) and ultracentrifuged at 125,000 g for 10 minutes at 4°C. Lysates were split and incubated with or without 100 mM DTT (Sigma-Aldrich) for 5 minutes at room temperature. Protein was analyzed in denaturing, nonreducing conditions by SDS-PAGE, transferred onto nitrocellulose membranes, and immunoblotted.

Assay of PTEN redox state and activity.

To measure PTEN activity modulation by oxidation the indirect method relying upon the oxidative protection from alkylation was performed as described previously (19).

Assessment of cell viability.

Cells were cultured in 24-well plates as 2 × 10⁶ cells/ml at 37°C with 5% CO₂ in control medium or with the indicated concentrations of LY294002 (Calbiochem), TBB, DRB, β-ME, NAC (Sigma-Aldrich), or combinations thereof. At 24, 48, and/or 72 hours, cells were harvested and quantitative determination of viability was performed using an Annexin V–based apoptosis detection kit (R&D Systems), as previously described (3). Alternatively, viability was assessed by forward scatter–side scatter flow cytometry analysis. Viability index was calculated as the ratio of experimental to control conditions.

Statistics.

Differences between populations were calculated using unpaired 2-tailed Student’s t test or Mann-Whitney test, as appropriate. A P value less than 0.05 was considered significant.