Bisphosphorylated PEA-15 Sensitizes Ovarian Cancer Cells to Paclitaxel by Impairing the Microtubule-Destabilizing Effect of SCLIP Free

Paclitaxel is a chemotherapeutic drug commonly used for the treatment of ovarian, breast, and other malignancies (1, 2). Paclitaxel kills tumor cells by increasing microtubule polymerization via binding to the β-subunit in preformed microtubules (3) and by stabilizing the microtubule network via direct interaction with microtubules (4). Thus, paclitaxel interferes with microtubule function and disrupts the mitotic spindle, leading to G₂–M arrest and eventually programmed cell death (5, 6). Paclitaxel has limited effectiveness in the treatment of ovarian cancer because of intrinsic or acquired resistance of cancer cells to the drug (7–12). Identifying molecular determinants of sensitivity to paclitaxel may lead to improvement of the therapeutic efficacy of the drug in ovarian cancer.

Previous studies showed that high expression level of PEA-15 (phosphoprotein enriched in astrocytes—15 kDa) in cancer cells correlated with longer survival of patients with ovarian cancer (13) and lower proliferation and less invasiveness of breast cancer cells (14, 15). PEA-15 is a multifunctional protein regulating cell proliferation, apoptosis, autophagy, and glucose metabolism (16). PEA-15 controls cell survival by interfering with the apoptotic or autophagic pathways (13, 14) and controls cell proliferation by interfering with the extracellular signal–regulate kinase (ERK) pathway (17, 18). The functions of PEA-15 are highly dependent on its phosphorylation status (16). PEA-15 is phosphorylated at Ser104 by protein kinase C and at Ser116 by calmodulin-dependent protein kinase II and AKT/PKB (19, 20). Previously, PEA-15 was shown to mediate paclitaxel resistance in breast cancer cells and protect these cells from paclitaxel-induced apoptosis through inhibition of the stress kinase pathway (21). However, the effect of PEA-15 phosphorylation status on chemosensitivity of cancer cells remains unknown.

SCLIP (SCG10-like protein), also known as STMN3, belongs to the stathmin/OP18 family, which also includes stathmin, SCG10 (superior cervical ganglion-10 protein), and RB3 (stathmin-like protein B3; ref. 22). Like other members of the stathmin/OP18 family, SCLIP functions as a regulator of microtubule dynamics (22). The stathmin/OP18 family proteins prevent microtubule polymerization by sequestering soluble tubulin heterodimers (22, 23) and promote microtubule depolymerization by increasing microtubule catastrophe rate (22, 24–26). Stathmin has been shown to regulate resistance of cancer cells to microtubule-targeting drugs, including paclitaxel, by affecting their binding to microtubules and delaying G₂–M transition (27). To date, there is no direct evidence implicating SCLIP in chemosensitivity of cancer cells. Here, we showed for the first time that SCLIP is involved in PEA-15–mediated chemosensitization in ovarian cancer cells.

In the present study, we investigated the effect of PEA-15 phosphorylation status on the sensitivity of ovarian cancer cells to paclitaxel. We showed for the first time that PEA-15 phosphorylated at both Ser104 and Ser116 (pPEA-15) sensitized ovarian cancer cells to paclitaxel by impairing SCLIP-mediated microtubule destabilization, thereby enhancing induction of mitotic arrest and apoptotic cell death by paclitaxel. Our findings provide a rationale for future clinical studies using pPEA-15 as a biomarker to improve the efficacy of paclitaxel-based therapy in ovarian cancer.

OVCAR3 ovarian cancer cells were obtained from American Type Culture Collection. KOC-7c ovarian cancer cells were provided by Dr. Toru Sugiyama (Kurume University, Kurume, Japan); OVTOKO ovarian cancer cells by Dr. Hiroshi Minaguchi (Yokohama City University, Yokohama, Japan); HEY and OVCA 432 ovarian cancer cells by Dr. Robert Bast (The University of Texas MD Anderson Cancer Center, Houston, TX); and SKOV3.ip1 ovarian cancer cells by Dr. Mien-Chie Hung (The University of Texas MD Anderson Cancer Center). SKOV3.ip1-vector cells stably transfected with pcDNA3.1 empty vector, SKOV3.ip1-AA cells stably expressing nonphosphorylatable PEA-15 (AA; Ser104 and Ser116 of PEA-15 substituted with alanine), and SKOV3.ip1-DD cells stably expressing phosphomimetic PEA-15 (DD; Ser104 and Ser116 of PEA-15 substituted with aspartic acid) were generated previously in our laboratory (28). All cell lines used in this study except for KOC-7c were validated using a short tandem repeat method based on primer extension to detect single base deviations in February 2013 by the Characterized Cell Line Core Facility at The University of Texas MD Anderson Cancer Center (29). We are unable to provide the exact dates on which we obtained these cell lines. HEY, OVCA 432, and SKOV3.ip1 cells were maintained in Dulbecco's Modified Eagle Media (DMEM)/F12 (Invitrogen); and OVCAR3, KOC-7c, and OVTOKO cells in RPMI-1640 (Invitrogen).

Cells were seeded in 6-well plates and transfected the next day with an ON-TARGET SMART siPEA-15 pool (a mixture of 4 designed siRNAs; Dharmacon), an ON-TARGET SMART siSCLIP pool (a mixture of 4 designed siRNAs; Dharmacon), or a scrambled control siRNA pool (a mixture of 4 siControl nontargeting siRNAs; Dharmacon) at a final siRNA concentration of 100 nmol/L using Oligofectamine (Invitrogen) as described previously (14). The cells were used for chemosensitivity analyses 24 hours after transfection.

Cells were transfected with plasmids of interest using the Neon Transfection System (Invitrogen) according to the manufacturer's instructions. Plasmids used in this study included pCMV6-Entry vector (Origene), pCMV6-Entry vector carrying SCLIP (Origene), pcDNA3.1 empty vector, and pcDNA3.1 vector carrying PEA-15, nonphosphorylatable PEA-15 (AA; Ser104 and Ser116 of PEA-15 substituted with alanine), or phosphomimetic PEA-15 (DD; Ser104 and Ser116 of PEA-15 substituted with aspartic acid). All the pcDNA3.1 vectors were generous gifts from Dr. Mark H. Ginsberg (University of California, San Diego, CA; ref. 17). Cells were used for chemosensitivity analyses 24 hours after transfection.

Western blotting was done as described previously (14). Proteins of interest were probed using the following primary antibodies (1:1,000 dilution) purchased from Cell Signaling Technology or other suppliers as indicated: anti-PEA-15 (Santa Cruz Biotechnology), anti-pPEA-15 at Ser104, anti-pPEA-15 at Ser116 (Invitrogen), anti-SCLIP (Proteintech Group), anti-detyrosinated α-tubulin (Millipore), anti-acetylated α-tubulin (Sigma-Aldrich), anti-α-tubulin, anti-β-actin (Sigma-Aldrich), anti-caspase-3, anti-caspase-8, anti-caspase-9, and anti-PARP. The secondary antibodies used were horseradish peroxidase–conjugated IgG (Invitrogen) for chemiluminescent signal detection and Alexa Fluor–conjugated IgG (Invitrogen) for fluorescence signal detection. Expression levels of target proteins were quantified using NIH ImageJ software (http://rsb.info.nih.gov/ij/) and then normalized to the β-actin protein level.

Cell growth inhibition by paclitaxel (Supplementary Fig. S1; App Pharmaceuticals) was determined using the WST-1 cell proliferation assay or trypan blue exclusion assay as described previously (28, 30). For the WST-1 assay, cells were seeded in 96-well plates and treated the next day with paclitaxel (0–10 μmol/L). After 72 hours of paclitaxel treatment, optical density at 450 nm was determined, and median growth inhibition (IC₅₀), which corresponds to the concentration of paclitaxel that results in 50% inhibition of net cell growth, was calculated. For the trypan blue exclusion assay, cells were seeded in 6-well plates and treated the next day with paclitaxel. After 72 hours of paclitaxel treatment, cell viability was determined using Vi-CELL series cell viability analyzers (Beckman Coulter) according to the manufacturer's instructions.

Anchorage-independent growth assay was conducted as described previously (14). Briefly, SKOV3.ip1 stable cells were resuspended in 0.4% agarose growth medium with or without paclitaxel. The cell suspensions were then plated in 6-well plates containing solidified 0.8% agarose in growth medium. Three weeks later, colonies greater than 80 μm in diameter were counted using the GelCount System (Oxford Optronix).

Cells were grown overnight and then treated with paclitaxel for 24 or 36 hours. The cells were then harvested, fixed, treated with RNase/propidium iodide (PI), and subjected to cell-cycle analysis by flow cytometry as described previously (31).

Total RNA was extracted from SKOV3.ip1-S116A cells (stably expressing nonphosphorylatable PEA-15 at Ser116, which was substituted with alanine) and SKOV3.ip1-S116D cells (stably expressing phosphomimetic PEA-15 at Ser116, which was substituted with aspartic acid) using RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. The integrity of the purified RNA was assessed using an Agilent 2100 BioAnalyzer (Agilent Technologies). The Affymetrix HGU133 plus platform was used for hybridization, staining, and imaging of the arrays by following the manufacturer's instructions. Data obtained from the microarrays were normalized by the robust multiarray average method (32). ANOVA P values and fold changes for gene expression were calculated using R statistical software version 2.12.2. A threshold cutoff was set to false discovery rate less than 0.01 and at least a 2-fold geometric change in gene-level expression between SKOV3.ip1-S116A and SKOV3.ip1-S116D cells. The microarray data have been deposited into the Gene Expression Omnibus database under the accession number GSE37934.

Total RNA was extracted from SKOV3.ip1 stable cells using an RNA prep kit (Invitrogen) according to the manufacturer's instructions. First-strand cDNAs were reverse-transcribed using the ImProm-II Reverse Transcriptase System Kit (Promega) by following the manufacturer's protocol. The quantitative PCR reactions were carried out using the SYBR Green qPCR Kit (Bio-Rad) with a pair of SCLIP primers: 5′-GGAGCTGCAAAAGCGGCTGG-3′ (forward) and 5′-CTGCTTCAGCACCTGCGCCT-3′ (reverse). Primers for human β-actin mRNA control were 5′-GCGGGAAATCGTGCGTGACATT-3′ (forward) and 5′-AGACAGTCTCCACTCACCCAGGAAG-3′ (reverse). Human β-actin mRNA was used as a normalization control. The mRNA levels of SCLIP in SKOV3.ip1 stable cells were first normalized to the mRNA levels of the housekeeping gene β-actin and then the fold induction of SCLIP mRNA was calculated on the basis of the SCLIP mRNA level in SKOV3.ip1-vector cells.

SKOV3.ip1 stable cells were grown overnight and then treated with paclitaxel for 12 hours. The cells were harvested, fixed in ice-cold 70% ethanol, and permeabilized with 0.25% Triton X-100. The cells were then incubated with anti-phosphohistone H3 antibody (Cell Signaling) and subsequently with fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Millipore). The cells were treated with RNase/PI and then analyzed for mitotic index by flow cytometry as described previously (33).

SKOV3.ip1 stable cells grown in culture chamber slides were treated with paclitaxel for 6 or 12 hours. The cells were fixed with ice-cold methanol, permeabilized with 0.2% Triton X-100, and blocked with 3% bovine serum albumin in PBS. The cells were then incubated with the following primary antibodies: anti-α-tubulin (Cell Signaling), anti-phosphohistone H3 (Cell Signaling), anti-acetylated α-tubulin (Sigma-Aldrich), or anti-detyrosinated α-tubulin (Millipore), followed by incubation with FITC-conjugated secondary antibodies (Invitrogen). The slides were mounted with mounting solution containing 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen). The microtubule network and mitotic spindles were photographed under ×400 magnification using an Eclipse 80i fluorescence microscope (Nikon). Fluorescence intensity of the microtubule network was quantified using NIS-Elements BR3.1 software (Nikon).

SKOV3.ip1 stable cells were grown overnight and then treated with paclitaxel for 12 hours. Soluble and polymerized tubulin were separated from the cultured cells as described previously (34) and then analyzed by Western blotting.

Each experiment was carried out at least in duplicate with 3 repeats, and data were expressed as means ± SD. Statistical analyses were conducted using SAS 9.3 software (SAS Institute). ANOVA was used to evaluate the significance of differences in means among different groups. P < 0.05 was considered significant.

To assess whether PEA-15 mediates chemosensitization in ovarian cancer cells, we silenced PEA-15 expression in PEA-15–high expressing HEY and OVTOKO ovarian cancer cells using PEA-15–targeting siRNA (Fig. 1) and then tested sensitivity by cell-cycle analysis. PEA-15 silencing resulted in a 12% reduction in the sub-G₁ fraction in HEY cells (0.005 μmol/L paclitaxel; P < 0.001; Fig. 1B) and an 11.2% reduction in the sub-G₁ fraction in OVTOKO cells (0.05 μmol/L paclitaxel; P < 0.001; Fig. 1C) compared with the sub-G₁ fraction in control cells treated with scrambled siRNA after paclitaxel treatment, indicating that PEA-15 silencing reduces paclitaxel sensitivity in ovarian cancer cells. To verify that loss of PEA-15 expression was responsible for the reduced paclitaxel sensitivity, we re-expressed PEA-15 in siRNA PEA-15–treated cells and observed a 24% increase in the sub-G₁ fraction in HEY cells (0.005 μmol/L paclitaxel; P < 0.001; Fig. 1B) and a 15.2% increase in the sub-G₁ fraction in OVTOKO cells (0.05 μmol/L paclitaxel; P < 0.001; Fig. 1C) compared with the sub-G₁ fraction in control cells transfected with vector after paclitaxel treatment. These results show that PEA-15 mediates paclitaxel sensitization in ovarian cancer cells.

PEA-15 is known to mediate paclitaxel resistance in breast cancer (21). The functions of PEA-15 are phosphorylation-dependent (16). To date, the effect of PEA-15 phosphorylation status on chemosensitivity of cancer cells remains unclear. We hypothesized that PEA-15 phosphorylated at both Ser104 and Ser116 sensitizes ovarian cancer cells to paclitaxel. To test this hypothesis, we used SKOV3.ip1-vector cells, SKOV3.ip1-AA cells stably expressing nonphosphorylatable PEA-15, and SKOV3.ip1-DD cells stably expressing phosphomimetic PEA-15 (Fig. 2A). SKOV3.ip1-DD cells (IC₅₀ = 0.01 μmol/L) were more sensitive to paclitaxel than were SKOV3.ip1-vector cells (IC₅₀ = 0.02 μmol/L) and SKOV3.ip1-AA cells (IC₅₀ = 0.11 μmol/L; Fig. 2B). Furthermore, SKOV3.ip1-DD cells displayed a 92.2% reduction in colony formation compared with SKOV3.ip1-vector cells (P < 0.001) and an 88.9% reduction in colony formation compared with SKOV3.ip1-AA cells (P < 0.001) after paclitaxel treatment (Fig. 2C). To confirm that PEA-15–mediated paclitaxel sensitization was dependent on its phosphorylation status, we silenced PEA-15 expression using PEA-15–targeting siRNA and then transiently overexpressed AA and DD in HEY and OVTOKO ovarian cancer cells (Fig. 2D) and assessed paclitaxel sensitivity using trypan blue exclusion assay. Compared with cells transfected with vector, HEY cells transfected with DD showed a 12% reduction (P < 0.001) in viability, whereas HEY cells transfected with AA showed a 10.6% increase (P < 0.05) in viability after treatment with paclitaxel (0.005 μmol/L; Fig. 2D). Similarly, compared with cells transfected with vector, OVTOKO cells transfected with DD showed a 19% reduction (P < 0.001) in viability, whereas OVTOKO cells transfected with AA showed an 11% increase (P < 0.05) in viability after treatment with paclitaxel (0.05 μmol/L; Fig. 2D). These results indicate that PEA-15–mediated paclitaxel sensitization depends on PEA-15 phosphorylation status.

Paclitaxel induces G₂–M arrest and apoptotic cell death (5, 6, 8). Thus, we hypothesized that pPEA-15 increases paclitaxel sensitivity in ovarian cancer cells by enhancing mitotic arrest and apoptosis. As expected, flow cytometric analysis showed that paclitaxel exposure led to a robust mitotic arrest in a dose-dependent manner (data not shown). After exposure to paclitaxel (0.01 μmol/L), SKOV3.ip1-DD cells displayed a 13.2% increase in mitotic accumulation, whereas SKOV3.ip1-AA cells displayed a 10.9% decrease in mitotic accumulation, compared with SKOV3.ip1-vector control cells (Fig. 3A). This result suggests that pPEA-15 increases paclitaxel cytotoxicity by enhancing paclitaxel-induced mitotic arrest.

It is known that paclitaxel-induced mitotic arrest is a result of inhibition of spindle organization (35). We observed that in the absence of paclitaxel, in dividing SKOV3.ip1-vector and SKOV3.ip1-AA cells, mitotic spindles were normal, whereas in dividing SKOV3.ip1-DD cells, the majority of mitotic spindles were abnormal and monopolar and contained ball-shaped aggregations of condensed chromosomes (Fig. 3B). After paclitaxel treatment, cells that were blocked at mitosis had abnormal spindles, and the majority of metaphase spindles displayed either monopolar or bipolar organization with chromosomes scattered throughout the spindles (Fig. 3B). However, no significant differences in spindle organization were observed among SKOV3.ip1-vector, SKOV3.ip1-AA, and SKOV3.ip1-DD cells after paclitaxel treatment. These findings show that when PEA-15 is phosphorylated at both Ser104 and Ser116, it promotes the ability of paclitaxel to block cell-cycle progression at mitosis.

We next investigated the effect of PEA-15 phosphorylation status on apoptosis induction by paclitaxel by cell-cycle analysis. After exposure to paclitaxel (0.01 μmol/L), SKOV3.ip1-DD cells showed a 14% increase in the sub-G₁ fraction, whereas SKOV3.ip1-AA cells showed a 12% decrease in the sub-G₁ fraction, compared with SKOV3.ip1-vector control cells (Fig. 3C). Similarly, when siRNA PEA-15–treated HEY and OVTOKO ovarian cancer cells (Fig. 2D) and KOC-7c ovarian cancer cells were transiently transfected with DD, these cells displayed 10% to 15% increases in the sub-G₁ fraction compared to the cells transfected with vector after paclitaxel treatment (Supplementary Fig. S2A–S2C). We further confirmed the enhanced apoptosis in SKOV3.ip1-DD cells after paclitaxel treatment by Annexin-V and 7-AAD staining. Compared with SKOV3.ip1-vector control cells, SKOV3.ip1-DD cells showed 19.8% and 30.1% increases in apoptosis when treated with 0.005 and 0.01 μmol/L paclitaxel, respectively (Supplementary Fig. S2D). These results indicate that when PEA-15 is phosphorylated at both Ser104 and Ser116, it enhances paclitaxel-induced apoptosis. Moreover, after paclitaxel treatment, we observed significant cleavage of PARP and caspase-3, -8, and -9 in SKOV3.ip1-DD cells compared with SKOV3.ip1-vector and SKOV3.ip1-AA cells (Fig. 3D). The addition of caspase-8 and -9 inhibitors reduced paclitaxel-induced apoptosis by 28.5% and 4.8%, respectively, in SKOV3.ip1-AA cells and by 56.6% and 49.5%, respectively, in SKOV3.ip1-DD cells (Supplementary Fig. S3). These results indicate that the enhanced apoptosis resulting from PEA-15 phosphorylation is a consequence of cleavage of both caspase-8 and -9, suggesting that paclitaxel induces apoptosis, at least in part, through activation of both the death effector and mitochondrial pathways.

To dissect the molecular mechanisms of PEA-15–mediated chemosensitization of ovarian cancer cells, we analyzed our cDNA microarray data obtained previously using SKOV3.ip1-S116A cells (Ser116 of PEA-15 substituted with alanine) and SKOV3.ip1-S116D cells (Ser116 of PEA-15 substituted with aspartic acid). We found that SCLIP was highly expressed in SKOV3.ip1-S116D cells. On the basis of this finding, we assessed SCLIP expression in SKOV3.ip1 stable cells by reverse transcription (RT)-PCR and Western blotting. The expression of SCLIP was 4.5-fold higher at the mRNA level (P < 0.001; Fig. 4A and Supplementary Fig. S4) and 1.6-fold higher at the protein level in SKOV3.ip1-DD cells than in SKOV3.ip1-vector control cells (Fig. 4B).

Because SCLIP regulates cellular functions by modulating microtubule dynamics (23, 26) and stathmin, a protein in the same family as SCLIP, has been implicated in paclitaxel resistance (27), we hypothesized that SCLIP is involved in pPEA-15–mediated chemosensitization. We next investigated whether the high expression level of SCLIP was associated with the enhanced chemosensitivity in SKOV3.ip1-DD cells. We assessed SCLIP expression at both the mRNA and protein levels in SKOV3.ip1 stable cells after paclitaxel treatment. No changes were observed in SCLIP expression at the mRNA level after paclitaxel treatment (data not shown). Interestingly, the protein level of SCLIP in paclitaxel-treated SKOV3.ip1-DD cells was only 51.2% of that in untreated SKOV3.ip1-DD cells, whereas SCLIP expression in paclitaxel-treated SKOV3.ip1-AA cells was not substantially different from that in untreated SKOV3.ip1-AA cells (Fig. 4C). Furthermore, SCLIP silencing in SKOV3.ip1-DD cells using SCLIP-targeting siRNA resulted in a 9% increase (P < 0.001) in the sub-G₁ fraction compared with the sub-G₁ fractions in the control cells treated with scrambled siRNA after treatment with paclitaxel (0.005 μmol/L) for 24 hours (Supplementary Fig. S5A). In contrast, transient re-expression of SCLIP in siRNA SCLIP-treated SKOV3.ip1-DD cells led to a 21.9% reduction (P < 0.001) in the sub-G₁ fraction compared to cells transfected with vector after treatment with paclitaxel (0.005 μmol/L) for 24 hours (Supplementary Fig. S5A). Similar results were obtained when cells were treated with paclitaxel for 36 hours (Supplementary Fig. S5A). These results indicate that reduced SCLIP expression following paclitaxel exposure correlates with increased paclitaxel sensitivity of SKOV3.ip1-DD cells.

To elucidate whether SCLIP regulates paclitaxel sensitization in ovarian cancer cells, we first silenced SCLIP expression in SCLIP–high expressing OVTOKO and HEY cells and then analyzed cell viability by trypan blue exclusion assy. Compared with cells treated with scrambled siRNA, OVTOKO cells with SCLIP silencing displayed a 29.1% reduction in cell viability (0.05 μmol/L paclitaxel; P < 0.001; Fig. 4D) and HEY cells with SCLIP silencing displayed an 18.6% reduction in cell viability (0.005 μmol/L paclitaxel; P < 0.01; Supplementary Fig. S5B) after 72 hours of treatment with paclitaxel. In contrast, when SCLIP was overexpressed in SCLIP–low expressing OVCA 432 and OVCAR3 cells, we observed a 16% increase in cell viability in OVCA 432 cells (0.0025 μmol/L paclitaxel; P < 0.01; Fig. 4E) and a 17.7% increase in cell viability in OVCAR3 cells (0.0025 μmol/L paclitaxel; P < 0.05; Supplementary Fig. S5C) compared with cells transfected with vector after 72 hours of paclitaxel treatment. These results show that SCLIP promotes paclitaxel resistance in ovarian cancer cells.

Interestingly, a shift in SCLIP molecular weight was noted in SKOV3.ip1-DD cells but not in SKOV3.ip1-AA cells following paclitaxel treatment (Fig. 4C, top). Paclitaxel has been reported to induce serine phosphorylation of the 66-kDa Shc isoform (36). Thus, the shift in SCLIP molecular weight may be a result of phosphorylation of SCLIP in paclitaxel-treated SKOV3.ip1-DD cells. However, phosphatase treatment of cell lysates did not revise the paclitaxel-induced shift of SCLIP molecular weight but only partially reduced SCLIP intensity (data not shown). These results suggest that reduced expression and possibly posttranslational modification of SCLIP following paclitaxel treatment impair the inhibitory effect of SCLIP on paclitaxel, which in turn enhances the sensitivity of SKOV3.ip1-DD cells to paclitaxel.

Paclitaxel exerts its antitumor activity by inducing cell-cycle arrest at mitosis through enhancement of microtubule polymerization (35). In contrast, SCLIP sequesters soluble tubulin heterodimers and destabilizes microtubule polymers (22, 23). As both paclitaxel and SCLIP modulate microtubule dynamics, we analyzed microtubule organization in SKOV3.ip1 stable cells by immunofluorescence staining following paclitaxel exposure. In the absence of paclitaxel, SKOV3.ip1-vector and SKOV3.ip1-AA cells contained long microtubule polymers, whereas SKOV3.ip1-DD cells had short and diffuse microtubule polymers that formed a meshwork of microtubules (Fig. 5A, left). After a 12-hour exposure to paclitaxel, the microtubule network in SKOV3.ip1-AA cells bundled with parallel microtubules radiating out from the nucleus, whereas microtubules were more condensed and oriented in parallel fashion and reorganized into thick ring-like bundles near the nucleus in SKOV3.ip1-vector and SKOV3.ip-DD cells (Fig. 5A, left). Similar phenomena were observed when cells were treated with paclitaxel for 3 or 6 hours (data not shown). Furthermore, the fluorescence intensity of α-tubulin in untreated SKOV3.ip1-DD cells was significantly lower than that in untreated SKOV3.ip1-vector and SKOV3.ip1-AA cells (P < 0.01; Fig. 5A, right). Following paclitaxel treatment, α-tubulin intensity in SKOV3.ip1-DD cells dramatically increased and was similar to that in SKOV3.ip1-vector and SKOV3.ip1-AA cells.

To confirm immunofluorescence staining results, we further quantified soluble and polymerized α- and β-tubulin from cells after 12 hours of paclitaxel exposure. In the absence of paclitaxel, SKOV3.ip1-DD cells had more soluble α- and β-tubulin subunits than did SKOV3.ip1-vector and SKOV3.ip1-AA cells (Fig. 5B), which might be a result of SCLIP-dependent microtubule destabilization in SKOV3.ip1-DD cells. However, paclitaxel exposure reduced the content of both soluble α-tubulin and soluble β-tubulin subunits, indicating enhanced polymerization of tubulin into microtubules in SKOV3.ip1-DD cells after paclitaxel exposure (Fig. 5B). In contrast, no significant changes were observed in the ratio of soluble to polymerized tubulin in SKOV3.ip1-vector and SKOV3.ip1-AA cells following paclitaxel exposure (Fig. 5B). These results suggest that increased paclitaxel sensitivity of SKOV3.ip1-DD cells may be a result of enhanced microtubule polymerization.

Posttranslational modifications of microtubules are known to stabilize individual microtubules in response to various stimuli. These modifications include α-tubulin detyrosination at the C-terminus (37) and acetylation at Lys40 (38), which occur primarily on tubulin assembled into microtubules and have been correlated with microtubule stability (39, 40). To address whether increased paclitaxel sensitivity of SKOV3.ip1-DD cells correlates with enhanced microtubule stability, we assessed the assembled microtubules by immunofluorescence staining for detyrosinated (Glu) and acetylated (Ac) α-tubulin subunits. In the absence of paclitaxel, all SKOV3.ip1 stable cells showed weak staining for both Glu and Ac α-tubulin (Fig. 6A and B, top). SKOV3.ip1-DD cells had less intense staining for both Glu and Ac α-tubulin than did SKOV3.ip1-vector and SKOV3.ip1-AA cells (Fig. 6A and B, bottom), indicating that tubulin is less polymerized in untreated SKOV3.ip1-DD cells. As expected, a 1-hour paclitaxel exposure resulted in an increase in the intensity of both Glu and Ac α-tubulin staining in each cell type (Fig. 6A and B). Paclitaxel-treated SKOV3.ip1-DD cells displayed more intense staining for both Glu and Ac α-tubulin than did paclitaxel-treated SKOV3.ip1-vector and SKOV3.ip1-AA cells (Fig. 6A and B).

Immunofluorescence staining results were further confirmed by Western blot analysis. The expression levels of both Glu (Fig. 6C) and Ac (Fig. 6D) α-tubulin were low in all untreated cells. In SKOV3.ip1-DD cells, the levels of both Glu (Fig. 6C) and Ac (Fig. 6D) α-tubulin dramatically increased 6 hours after paclitaxel exposure and remained high up to 48 hours. In SKOV3.ip1-vector and SKOV3.ip1-AA cells, the levels of Glu α-tubulin increased slightly at 24 hours after paclitaxel exposure and were similar to the levels in SKOV3.ip1-DD cells at 48 hours (Fig. 6C), whereas the levels of Ac α-tubulin peaked at 24 hours in both cell lines and then remained high up to 48 hours in SKOV3.ip1-vector cells but decreased in SKOV3.ip1-AA cells (Fig. 6D). The elevated acetylation and detyrosination of α-tubulin in paclitaxel-treated SKOV3.ip1-DD cells reflected the enhancement of microtubule polymerization and stability. These results clearly show that pPEA-15 overexpression enhances paclitaxel-induced microtubule polymerization and stability, resulting in increased sensitivity of SKOV3.ip1-DD cells to paclitaxel.

Our results show that when PEA-15 is phosphorylated at both Ser104 and Ser116, it increases the sensitivity of ovarian cancer cells to paclitaxel by impairing SCLIP-mediated microtubule destabilization, which is a result of reduced expression and possibly posttranslational modification of SCLIP following paclitaxel treatment. Impairment of SCLIP-mediated microtubule destabilization promotes induction of microtubule polymerization and stabilization by paclitaxel, leading to enhanced mitotic arrest and apoptotic cell death.

Our results here show that pPEA-15 mediates paclitaxel sensitization in ovarian cancer cells by activating both the death effector and the mitochondrial pathways. In contrast to our findings, Stassi and colleagues reported that PEA-15 mediated resistance to paclitaxel-induced apoptosis in breast cancer cells (21). This discrepancy may be due to differences in the phosphorylation status of PEA-15 as well as differences in the cell types, that is, breast cancer cells versus ovarian cancer cells. The functions of PEA-15 are tightly regulated by its phosphorylation status. Studies have shown that phosphorylation at Ser104 inhibits binding of PEA-15 to ERK, whereas phosphorylation at Ser116 promotes binding of PEA-15 to FADD (41). Furthermore, PEA-15 phosphorylation at Ser116 is required for its anti-apoptotic activity in both 293 cells and U373MG glioma cells (20), whereas monophosphorylation at either Ser104 or Ser116 is essential for inhibition of Fas-mediated but not TNF-α–mediated apoptosis in NIH-3T3 cells (41). These studies clearly indicate that phosphorylation status of PEA-15 plays a crucial role in determining its biologic functions. In addition, our finding that pPEA-15 sensitizes ovarian cancer cells to paclitaxel strongly suggests that PEA-15 phosphorylation status is also critical for the responsiveness of cancer cells to chemotherapy.

Microtubule dynamics have been considered a key factor in determining the chemosensitivity of cancer cells (5). Subtle suppression of microtubule dynamics by paclitaxel inhibits the assembly and functioning of the mitotic spindle, leading to mitotic arrest and apoptotic cell death (5, 6). Thus, cellular molecules that enhance the effect of paclitaxel on microtubule polymerization and stabilization may improve its cytotoxicity (42). Our results here showed that overexpression of PEA-15 that is phosphorylated at both Ser104 and Ser116 markedly increased the ability of paclitaxel to induce mitotic arrest and apoptosis in ovarian cancer cells. Furthermore, dramatic enhancement in microtubule polymerization and in tubulin acetylation and detyrosination was observed in pPEA-15–overexpressing SKOV3.ip1-DD cells as early as 6 hours following paclitaxel exposure. It has been reported that the effect of paclitaxel on microtubule dynamics is at least partially due to its effect on tubulin acetylation and detyrosination, established markers of microtubule stability (39, 40). This information in combination with our results indicates that the mechanism underlying the high sensitivity of SKOV3.ip1-DD cells to paclitaxel is the enhanced microtubule polymerization and stabilization, which lead to mitotic arrest and apoptosis.

In the present study, we showed that overexpression of PEA-15 phosphorylated at both Ser104 and Ser116 induced high-level expression of SCLIP, whereas paclitaxel exposure resulted in a dramatic reduction in SCLIP expression at the protein level. SCLIP is able to inhibit microtubule polymerization by sequestering soluble tubulin (43) and promote microtubule depolymerization by increasing the microtubule catastrophe rate (24, 25). Stathmin has been implicated in modulating the sensitivity of cancer cells to microtubule-targeting drugs by interfering with interaction of the drugs with microtubules (31, 44). However, up to now, the biologic functions of SCLIP and its role in chemosensitization have been largely elusive, and these functions were only inferred, in part, by extrapolation from what is known about other stathmin/OP18 family members. Our study here provides direct evidence that SCLIP has a functional effect on chemosensitivity of paclitaxel through a microtubule-dependent mechanism. The reduced SCLIP expression following paclitaxel treatment impaired the microtubule-destabilizing effect of SCLIP as supported by high-level expression of Glu and Ac tubulin, thereby facilitating the promotion of microtubule assembly and apoptosis by paclitaxel.

In addition to reduced SCLIP expression, a shift in SCLIP molecular weight was noted after paclitaxel exposure, which may have resulted from enhanced phosphorylation of SCLIP. Phosphorylation of stathmin/OP18 family proteins at serine residues has been reported to impair their effects on microtubule dynamics and halt their antagonistic activity against microtubule-targeting drugs (22). As there are no antibodies available to detect the phosphorylation status of SCLIP, we conducted phosphatase treatment of cell lysates to verify whether the molecular weight shift of SCLIP was a result of phosphorylation. However, phosphatase treatment did not revise the shift of SCLIP molecular weight but only partially reduced the intensity of SCLIP (data not shown), indicating that phosphorylation is not the sole reason for SCLIP molecular weight change. Thus, other unknown modifications, possibly glycosylation, may also be responsible for the change in SCLIP molecular weight and influence the effect of SCLIP on microtubule instability. Further studies are needed to address the mechanisms that underlie the effect of pPEA-15 on SCLIP expression and the effect of paclitaxel on SCLIP functioning.

Our study defines an important role for pPEA-15 in regulating chemosensitization of paclitaxel in ovarian cancer cells: pPEA-15 impairs the microtubule-destabilizing activity of SCLIP, thereby enhancing the effects of paclitaxel on microtubule assembly, mitosis, and apoptosis. In addition, our results suggest that pPEA-15 could be used as a biomarker to predict the responsiveness of ovarian tumors to paclitaxel treatment and that patients with ovarian tumors expressing high levels of pPEA-15 would be more likely to benefit from paclitaxel treatment than other patients. In our future studies, we will examine the association between PEA-15 phosphorylation status and the pathologic grade of ovarian tumors and evaluate the in vivo impact of PEA-15 phosphorylation status on the effectiveness of paclitaxel in a human ovarian carcinoma xenograft.

No potential conflicts of interest were disclosed.

Conception and design: X. Xie, C. Bartholomeusz, N.T. Ueno

Development of methodology: A.A. Ahmed, N.T. Ueno

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Kazansky, N.T. Ueno

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Xie, C. Bartholomeusz, A.A. Ahmed, L. Diao, K.A. Baggerly, N.T. Ueno

Writing, review, and/or revision of the manuscript: X. Xie, C. Bartholomeusz, G.N. Hortobagyi, N.T. Ueno

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Xie, G.N. Hortobagyi, N.T. Ueno

Study supervision: G.N. Hortobagyi, N.T. Ueno

The authors thank Stephanie P. Deming of the Department of Scientific Publications for her editorial assistance; the Flow Cytometry and Cellular Imaging Core Facility at MD Anderson Cancer Center for flow cytometric analyses; the Genomics Core Facility at MD Anderson Cancer Center for cDNA microarray analysis; and Tamer S. Kaoud (The University of Texas at Austin, Austin, TX) for providing paclitaxel structure.

This work was supported by NIH/NCI grant (R01CA127562 03) awarded to N.T. Ueno and NIH/NCI Cancer Center Support Grant (CA016672) awarded to The University of Texas MD Anderson Cancer Center.

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.