Phosphoinositide 3-Hydroxide Kinase Blockade Enhances Apoptosis in the Ewing’s Sarcoma Family of Tumors¹

ESFTs⁴ affects patients between the ages of 3 and 40 years, with most cases occurring in the second decade of life. Patients who present with metastatic ESFTs have a poor prognosis, despite high-dose chemotherapy (1), whereas patients who present with localized tumors have ∼70% disease-free survival (2). These clinical response rates have persisted for the past decade, despite attempts to dose-intensify chemotherapy, which has included the antitumor agent doxorubicin (3).

Ninety-five % of ESFTs contain a characteristic translocation that joins the EWS gene located on chromosome 22 to an ets family gene: either FLI-1 located on chromosome 11, t(11;22) (4), or ERG located on chromosome 21, t(21;22) (5, 6). The chimeric fusion protein, EWS/FLI-1 or EWS/ERG, generated as a result of either of these translocations contains two primary domains. The EWS domain is a potent transcriptional activator, whereas the FLI-1 or ERG domains contain a highly conserved DNA binding domain (7). The EWS/FLI-1 transforms mouse fibroblasts, and this transformation requires intact EWS and FLI-1 domains (7). We have shown that the IGF-IR is required for the EWS/FLI-1 transformation of fibroblasts (8). IGF-I also functions in an autocrine loop in ESFT, e.g., some ESFT cell lines synthesize IGF-I, which then activates the IGF-IR (9, 10).

The signaling pathway initiated by the binding of IGF-I to the IGF-IR is of pivotal importance in human carcinogenesis (11). Recently, a series of studies have shown that elevated plasma levels of IGF-I are associated with an increased risk of prostate and breast carcinoma in selected patients (12, 13). The molecular mechanisms by which IGF-I/IGF-IR interactions moderate normal and neoplastic growth are only partly understood. In normal cells, the IGF-IR pathway is modulated to achieve homeostatic growth, whereas in malignant cells, the pathways are dysregulated, resulting in increased transformation and diminished apoptosis (11).

IGF-IR inhibition of apoptosis is in part mediated through PI 3-K activation (14, 15). PI 3-K produces PIP₃, a molecule that binds a serine-threonine protein kinase, termed Akt, alters the conformation of Akt, and causes its translocation to the cytoplasmic membrane (16, 17). Akt kinase, PDK1, then activates Akt, which then modulates cell survival (18, 19). Because in vivo blockade of IGF-IR signaling in tumor cells results in decreased growth, we investigated whether this decrease in growth was attributable to inhibition of the downstream target, Akt (20).

Unfortunately, no molecules that specifically inhibit Akt are presently available; therefore, we targeted PI 3-K, a molecule upstream from Akt in the IGF-IR pathway. Specific inhibitors of PI 3-K have been reported including wortmannin and LY294002. Prior studies have demonstrated that PI 3-K blockade with these inhibitors prevents the activation of Akt in cultures of primary neurons (21), fibroblasts (14, 16, 22), and lymphoid progenitors (17).

Our goal was to determine whether the IGF-IR-initiated pathway from PI 3-K through Akt activates a survival function in ESFT cell lines. Our data demonstrate that IGF-I, indeed, can act as a survival factor for ESFT cells. Survival signaling can occur through the PI 3-K pathway, and the survival is attributable to reduced apoptosis. Inhibition of PI 3-K increases doxorubicin-induced apoptosis in ESFT cells. To establish whether Akt plays a role in ESFT survival signaling, we transfected an ESFT cell line to express constitutively activated Akt. This constitutively activated Akt renders the cells more resistant to doxorubicin-induced apoptosis. These data support the hypothesis that the IGF-IR/PI 3-K/Akt pathway drives survival signaling in ESFT cells.

TC-32 cells have been described previously (23), and TC-71 cells were a generous gift of Dr. Maria Tsokos (National Cancer Institute, Bethesda, MD). The plasmid pCEFL contains the elongation factor-1 promoter. pCEFL-HA-Akt-MYR contains the cAkt cDNA fused to a myristoylation signal (24). Hemagglutinin antibody was purchased from Boehringer Mannheim Corp. (Indianapolis, IN). Wortmannin and LY294002 were purchased from Sigma Chemical Co. (St. Louis, MO). Doxorubicin was purchased from Pharmacia (Piscataway, NJ).

Cells were grown at 37°C in 5% CO₂ in RPMI 1640 and 10% fetal bovine serum, except for 6-h apoptosis studies, which were performed in serum-free media to avoid serum binding of study molecules. Cells were transfected using a Life Technologies Electroporator at 1180 μF and 350 V. Polyclonal and monoclonal populations were selected in 500 μg/ml G418.

Growth studies were performed by assessing cell viability with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide reagent (Sigma) at 500 μg/ml for 3.0 h. Precipitated dye was dissolved in 100% isopropanol, and absorbance was read at 570 nm with background subtraction at 690 nm. Microtiter plate experiments were performed in triplicate, and averaged results were reported as a percentage of vehicle-treated cells.

For DNA laddering, log-phase cells were washed and exposed to combinations of agents for 6 h, following which all cells, both floating and adherent, were collected, washed in PBS, and then lysed in a buffer containing 10 mm Tris (pH 8.0), 100 μm EDTA (pH 8.0), and 0.5% SDS. Lysates were treated with proteinase K overnight at 56°C, followed by DNA extraction using phenol:chloroform and ethanol precipitation. Ten μg of DNA were loaded into wells of a 1% agarose gel. After electrophoretic separation, the gel was stained with ethidium bromide and photographed.

Cells were treated for 6 h, as described in “Results,” and all cells in a dish were harvested, fixed, and stained according to the manufacturer’s instructions (Phoenix Flow Systems, Inc., San Diego, CA). After staining, cells were analyzed by flow cytometry for FITC staining to demonstrate apoptotic DNA fragments and propidium iodide counterstain to evaluate cell cycle.

Cells were treated with agents described for the times indicated in the specific figures. After the incubation, cells were lysed with a previously reported buffer that protects phosphorylated proteins with the protease inhibitors leupeptin, aprotinin, and phenylmethylsulfonyl fluoride (8). Cell lysates were clarified with centrifugation, and total protein was determined using copper reduction (BCA; Pierce). Twenty μg of protein were incubated with 20 mm HEPES (pH 7.5), 2 mm DTT, and 10% glycerol buffer containing 0.08 mm DEVD-AMC (PharMingen, San Diego, CA). Reactions were excited at 360 nm, and emission at 460 nm was measured in a fluorescence plate reader (Cytofluor II; Biosearch, Inc., Bedford, MA) at hours 2–6. Relative caspase activity was determined by dividing the fluorescence of the experimental point by that of the control the control point.

IGF-I is a well-described survival factor in nonmalignant cell types as well as some malignancy models. Whereas the IGF-IR has been shown to be critical for ESFT cell growth and transformation, signaling through this receptor has not been considered as a mechanism of drug resistance. We investigated whether IGF-I, as principal ligand for IGF-IR in the ESFTs (10), would render cells more resistant to doxorubicin toxicity. Because doxorubicin is known to kill cells by apoptosis (25), we treated two ESFT cell lines with doses of doxorubicin that rapidly induce caspase-3 activity, an early marker of apoptosis. Relative levels of caspase-3 were increased 23-fold in the ESFT cell line TC-32, after a 6-h incubation with doxorubicin (1 μg/ml), whereas concomitant treatment with 10 nm IGF-I reduced the caspase-3 level to only 3-fold, an 8-fold drop (Fig. 1,A). A second ESFT cell line, TC-71, was treated with a dose escalation of doxorubicin, and as expected, a dose escalation in caspase-3 activity is seen (Fig. 1 B). When treated with IGF-I, caspase levels are reduced an average of 4-fold. A threshold effect of IGF-I is seen with 100 nm IGF-I, not reducing caspase-3 activity significantly >10 nm IGF-I. IGF-I, therefore, acts as a potent suppresser of doxorubicin-induced caspase-3 activation.

PI 3-K is activated after IGF-IR stimulation by way of the adaptor molecules IRS-1/2. Many additional cytokine and interleukin receptors also activate PI 3-K to enhance cell survival (26, 27, 28). We evaluated the role of PI 3-K and Akt, downstream effectors of IGF-IR which have been implicated in cell survival (15, 27). Activated PI 3-K produces PIP₃, which interacts with the pleckstrin homology domain of Akt, promoting translocation of Akt to the cell membrane. We used wortmannin, a well-characterized inhibitor of PI 3-K, because no specific Akt inhibitors are presently known. Wortmannin (500 nm) demonstrated complete inhibition of IGF-I-stimulated Akt activity in TC-32 cells, using an in vitro kinase assay (data not shown). Cells were incubated with doxorubicin (0.1 μg/ml) for 4 days. In addition, wortmannin (5 μm) or an equal amount of DMSO vehicle was added to wells on day 0 after the cells were plated. The ESFT cell line TC-32 achieved only 28% of control cell growth when treated with the combination of wortmannin and doxorubicin, whereas each agent alone showed 79 and 63% growth compared with control, respectively (Fig. 2). The ESFT cell line TC-71 achieved 55% of control growth when treated with the combination of wortmannin and doxorubicin, whereas growth with each agent alone was 94 and 84%, respectively (Fig. 2). Thus, wortmannin, presumably through inhibition of PI 3-K, enhances the toxicity of doxorubicin upon growing cell ESFT cell lines.

We sought to find out whether inhibition of PI 3-K led to increased apoptotic cell death. The connection between PI 3-K and apoptosis has been shown in lymphoid (26) and neuronal (21, 29) precursors. We investigated whether inhibition of PI 3-K decreased cell number by increasing apoptosis rather than by retarding cell growth. We evaluated apoptosis both by assaying caspase-3 activity and DNA fragmentation assays. We demonstrated DNA fragmentation both qualitatively (DNA ladder and fluorescence microscopy) and quantitatively using TdT end-labeling, followed by flow cytometry.

TC-71 cells were treated in the presence or absence of doxorubicin (0.5 μg/ml) for 6 h in the presence or absence of the PI 3-K inhibitor LY294002 (40 and 80 μm). Caspase activity was measured in a fluorometer after cleavage of the fluorescent marker AMC from a quadrapeptide target of caspase-3 (DEVD) after a 4-h incubation. Relative caspase activity is determined in each experiment by dividing arbitrary fluorescence units of the experimental condition by the fluorescence in untreated cells. We show an 11-fold increase in caspase-3 activity when TC-71 cells are treated with doxorubicin and 80 μm LY294002 compared with a 4-fold increase with LY294002 alone (Fig. 3). Increasing the concentration of LY294002 caused an increase in caspase-3 activity; thus, a dose effect of LY294002 occurred when combined with doxorubicin. In TC-71 cells, caspase activity was not increased with doxorubicin alone at this concentration. We also studied DNA fragmentation, an end point of apoptosis, to confirm our findings. TC-71 cells were treated as above and harvested as intact cells. Apoptotic nuclei are demonstrated by the characteristic DNA fragmentation pattern in cells treated with the combination of doxorubicin and LY294002 (Fig. 4, lower panel C). Cells were fixed for the TUNEL assay, which consists of TdT-FITC labeling followed by propidium iodide counterstain. FITC staining indicates the number of apoptotic cells on the Y axis, whereas propidium iodide staining shows DNA content on the X axis. Flow cytometric analysis of cells treated with the combination of doxorubicin and LY294002 demonstrate apoptosis in 36% of the cells (Fig. 4, upper panel C). Cells treated with either LY294002 (Fig. 4,A) or doxorubicin (Fig. 4 B) demonstrated 6.3 and 3.6% apoptosis, respectively, and lack nuclear changes. These data show enhanced apoptosis at both the enzymatic level, caspase-3 activation, and terminal DNA fragmentation, TUNEL assay, in cells treated with a PI 3-K inhibitor along with doxorubicin.

TC-32 cells were tested for the ability to undergo apoptosis by inducing caspase-3 and DNA fragmentation in similar fashion to TC-71. Cells were established in log phase growth and washed prior to the addition of doxorubicin with or without the PI 3-K inhibitors wortmannin or LY294002. We show dose responses to doxorubicin alone (0.05, 0.125, and 0.25 μg/ml), doxorubicin plus wortmannin (0.5 μm), and doxorubicin plus LY294002 (40 μm). The data in Fig. 5 demonstrate that little caspase activation is seen with doxorubicin alone. However, when wortmannin is added, a doxorubicin dose-dependent increase in apoptosis occurred with up to a 5-fold increase in caspase-3 activity at 0.25 μg/ml doxorubicin (Fig. 5, middle set of columns). Combining LY294002 with doxorubicin also enhanced caspase-3 activity 4-fold (Fig. 5, right set of columns). Relative caspase activity with wortmannin alone was 1.4 and LY294002 was 1.2 relative caspase units (data not shown). DNA fragmentation was used to determine whether inhibition of PI 3-K affected apoptosis. TC-32 cells were treated as above and harvested as intact cells or nucleic acid lysates, followed by DNA extraction. Apoptosis is demonstrated here by characteristic DNA ladder after treatment with doxorubicin and wortmannin (Fig. 6,A). Flow cytometric analysis demonstrated that the combination of doxorubicin plus wortmannin produced 43% apoptosis, doxorubicin alone produced 18% apoptosis, whereas apoptosis occurred in only 3% with either wortmannin alone or medium alone (Fig. 6, B–E). Here, five independent experiments with different measurements of apoptosis demonstrated that inhibition of PI 3-K, with two different molecules, enhanced doxorubicin-induced apoptosis in two different ESFT cell lines.

We evaluated Akt as a possible mediator of ESFT cell survival that occurs downstream of PI 3-K. Akt-MYR is a constitutively active Akt attributable to the addition of a myristoylation signal that causes membrane targeting (24). TC-32 cells were stably transfected with Akt-MYR, which also has a hemagglutinin tag, whereas control cells were transfected with the empty vector (Neo). Expression of the Akt-MYR gene is demonstrated with PAGE separated proteins immunoblotted with a hemagglutinin antibody (Fig. 7, insets). This series of TC-32 stably transfected Akt-MYR clones were tested for caspase-3 activity after doxorubicin treatment. Three clones that expressed stable levels of Akt-MYR are shown: D4, A2, and B2. Clone N5 is representative of the three Neo clones studied. At low doses of doxorubicin (0.1 μg/ml), all three Akt-MYR clones show reduced relative caspase activity (2–4 fold), with N5 showing a 10-fold increase in relative caspase activity (Fig. 7). With a 2.5-fold higher dose of doxorubicin (0.25 μg/ml), only the highest Akt-MYR expressing clone, B2, demonstrated a reduction in relative caspase activity compared with the other clones (Fig. 7). In each case, expression of Akt-MYR rendered TC-32 cells more resistant to apoptosis compared with Neo vector-transfected cells. Polyclonal transfection of Akt-MYR into TC-32 cells evaluated by TUNEL, followed by flow cytometry, demonstrated similar reductions in apoptosis after doxorubicin treatment compared with control transfected cells (data not shown). Thus, constitutive expression of activated Akt reduced doxorubicin-induced apoptosis.

Survival of patients with the ESFT has not improved significantly over the past 10–15 years, despite dose intensification of chemotherapeutic regimens. The IGF-IR has been shown to play a major role in the growth (9, 10, 30) and transformation (8) of ESFTs. The IGF-IR initiates a cytokine signaling pathway, implicated in promoting cell survival by inhibiting apoptosis. We sought to determine whether ESFTs could use this pathway as a mechanism of resistance to chemotherapy-induced apoptosis.

IGF-I protects many types of cells from toxic insults, including osmotic stress (31), UV irradiation (14), serum withdrawal (15, 16), and amyloid (32). Recently, IGF-I was shown to protect retinoblastoma cells from chemotherapy-induced apoptosis (33). Our studies show that IGF-I potently inhibits doxorubicin-induced caspase-3. Caspase-3 activation is seen prior to morphological apoptosis in the ESFT cells.⁵ Our data, therefore, implies that IGF-I treatment of ESFT rendered the cells more resistant to doxorubicin-induced apoptosis. This finding might represent a mechanism of ESFT resistance to chemotherapy.

PI 3-K is a downstream target of the IGF-IR that has been shown to mediate survival after exposure to UV irradiation (14) or serum withdrawal (15, 16). Our data demonstrate that growth of ESFT cells is inhibited by a 4-day incubation with a PI 3-K inhibitor along with doxorubicin. To determine whether the decreased cell number was due to apoptosis, both caspase-3, an enzyme in the apoptotic cascade, and terminal DNA fragmentation was studied. These assays show that ESFT cells exhibit enhanced apoptosis when treated with doxorubicin in the presence of PI 3-K blockade. The enhanced apoptosis in ESFT cells infers that IGF-I is acting, at least in part, through PI 3-K. The IGF-I induced activation of PI 3-K in ESFT could come direct via IRS-1, which moves from a low to a high phosphorylation state in fibroblasts transformed with the EWS/FLI-1 fusion protein (8). Alternatively, IGF-I is known to activate Ras, which can then activate PI 3-K (34). PI 3-K can also be activated via the heterotrimeric G-proteins (24). Because multiple pathways lead to activation of PI 3-K, tumor cells might exploit a variety of pathways to maintain survival.

The PI 3-K product PIP₃ directs the translocation of Akt, a serine-threonine kinase, to the membrane where it is activated by phosphorylation. Akt-enhanced cell survival occurs via phosphorylation of the bcl family member Bad (16, 17) and/or pro-caspase-9 (35). The phosphorylated Bad is then sequestered by 14-3-3, allowing Bcl-x_L to homodimerize (36). The homodimerization of Bcl-x_L blocks the progression of apoptosis. Phosphorylation of pro-caspase-9 prevents the activation of caspase-3 and thus prevents the progression of the caspase cascade. We demonstrate that constitutively active Akt results in decreased doxorubicin-induced caspase-3 activation. Increased resistance to doxorubicin paralleled increased levels of activated Akt expression in these clones, further implying a mechanistic connection between ESFT cell survival and Akt activation. It is likely that caspase-9 is a target of Akt in ESFT because caspase-9 is the activator of caspase-3 (35).

Because the IGF-IR activates many different pathways, protection from cell death needs to be considered in a broad sense. Our results confirm the importance of PI 3-K and Akt in ESFT cell survival. However, this may not be the only IGF-IR-mediated survival pathway that ESFT cells might use to avoid toxin-induced apoptotic death. Human embryonal kidney cells demonstrate IGF-I inhibition of the stress-activated protein kinase c-Jun NH₂-terminal kinase (37). This inhibition is thought to progress in part through Akt; however, part of the suppression of apoptosis is not accounted for by using a dominant-negative Akt construct (37). Rat fibroblasts have an alternate survival pathway when the IGF-IR is overexpressed that is not yet defined (27). These alternate pathways may play a role in ESFT cell survival as well and require additional study.

Akt is pivotal in the decision to enter the apoptotic pathway in many nonmalignant models (14, 16, 17, 21). Our work implicates PI 3-K and Akt as important survival molecules in ESFTs. It will be important to further dissect the pathway proximal and distal to Akt to develop an understanding of its full role as a survival protein in ESFT. As survival pathways and their redundancies are understood, novel therapeutic targets will emerge. These novel targets will hopefully allow for improved antineoplastic agents to increase survival in patients with ESFT.

J. A. T. would like to thank Drs. Jonathan Finlay, Bruce Bostrom, Harold Maurer, Gita Massey, Len Neckers, Narayan Bhat, Lee Helman, and Phillip Pizzo for inspiring his commitment to excellence in the care of children with cancer and cancer research. We thank Dr. Silvio Gutkind for Akt plasmids and helpful discussions. We also thank Drs. Bob Fenton and Anne Hamburger for critical review of the manuscript.