Skip to main content
NCBI home page
Endocrinology logoLink to Endocrinology
. 2008 May 1;149(8):3809–3816. doi: 10.1210/en.2007-1584

The Role of Activin A and Akt/GSK Signaling in Ovarian Tumor Biology

Thuy-Vy Do 1, Lena A Kubba 1, Monica Antenos 1, Alfred W Rademaker 1, Charles D Sturgis 1, Teresa K Woodruff 1
PMCID: PMC2488253  PMID: 18450971

Abstract

Elevated activin A levels in serum, cyst fluid, and peritoneal fluid of ovarian cancer patients suggest a role for this peptide hormone in disease development. We hypothesize that activin A plays a role in ovarian tumor biology, and analyzed activin-mediated pro-oncogenic signaling in vitro and the expression of activin signaling pathway molecules in vivo. Activin A regulation of Akt and GSK, and the effects of repressing the activities of these molecules (with pharmacological inhibitors) on cellular proliferation were assessed in the cell line, OVCA429. Activin A activated Akt, which phosphorylated GSK, repressing GSK activity in vitro. Activin A stimulated cellular proliferation and repression of GSK augmented activin-regulated proliferation. To validate in vitro observations, immunostaining of the βA-subunit of activin A and phospho-GSKα/β (Ser9/21) was performed, and the correlation between immunoreactivity levels of these markers and survival was evaluated in benign serous cystadenoma, borderline tumor, and cystadenocarcinoma microarrays. Analysis of tissue microarrays revealed that βA expression in epithelia did not correlate with survival or malignancy, but expression was elevated in stromal cells from carcinomas when compared with benign tumors. Phospho-GSKα/β (Ser9/21) staining was more intense in mitotically active carcinoma cells and exhibited a polarized localization in benign neoplasms that was absent in carcinomas. Notably, lower phospho-GSKα/β (Ser9/21) immunoreactivity correlated with better survival for carcinoma patients (P = 0.046). Our data are consistent with a model in which activin A may mediate ovarian oncogenesis by activating Akt and repressing GSK to stimulate cellular proliferation.

AMONG WOMEN, ovarian cancer is the fifth most common cancer in the United States and ranks fifth in cancer-related deaths in developed countries (1,2). In the United States, one in 2500 postmenopausal women are diagnosed with ovarian cancer, posing a challenge for the establishment of improved technologies for screening and clinical management (3). Because ovarian cancer is asymptomatic during early stages and, therefore, usually detected late, early detection and controlling disease dissemination are major hurdles to overcome. During early stages when cancer is still confined to the ovary, the 5-yr survival rate is over 90%. However, survival rates plummet to 20–30% if diagnosis occurs during later stages (4,5). The dismal survival rates reflect mortalities that result from recurring disease that does not respond to currently available therapies.

Approximately 90% of ovarian cancers are epithelial in origin, and despite numerous published studies, the etiology and epidemiology of this cancer are poorly understood. Ovarian carcinoma is a heterogeneous group of malignancies that is subdivided into histological subtypes–serous (Fallopiantube-like), mucinous (endocervical-like), endometrioid (proliferative endometrium-like), and clear cell (gestational endometrium-like)–according to the type of epithelia, which make up the organs of the reproductive tract, that the carcinoma resembles (6,7). Ovarian carcinomas are thought to be derived from the ovarian surface epithelia (OSE), which have a more uncommitted phenotype (possessing both epithelial and mesenchymal characteristics) than their malignant counterpart (8). Normal OSE seldom express the epithelial marker, E-cadherin, but do express the mesenchymal marker, N-cadherin. In fact, malignant ovarian epithelia acquire E-cadherin expression (9,10). Furthermore, expression of E-cadherin in SV40 T-antigen-immortalized OSE cells induces a mesenchymal-to-epithelial transition and the secretion of the tumor antigen, CA125, which is often produced by metaplastic and neoplastic OSE (11).

The peptide hormone, activin, is a member of the TGF-β superfamily, and initiates signal transduction pathways critical for reproductive functions and development. Activins exist as homo- or heterodimers of the βA- or βB-subunits to make up activin A (βA-βA), activin AB (βA-βB), and activin B (βB-βB). Activins bind to heteromeric receptor complexes consisting of a type I (ActRIA and ActRIB) and a type II receptor (ActRIIA and ActRIIB). Like TGF-β, activin signals through Smad-dependent pathways mediated by the transcription factors, Smad2 and Smad3, or through Smad-independent pathways, mediated by MAPK family members and phosphatidylinositol 3-kinase (PI3K), to name a few (12,13,14).

In the ovary, activin A regulates follicle development and the expression of the FSH β-gene (FSHβ). Inhibin, another TGF-β superfamily member, and the activin-binding protein, follistatin (Fst), antagonize activin signaling (15,16). Activin A treatment has differential effects on normal OSE cells and their malignant counterpart, ovarian carcinoma cells. For example, activin A has no effect on the proliferation of normal OSE but stimulates proliferation of ovarian cancer cells (17).

Activin A promotes migratory and invasive potential in normal cells during epithelial-to-mesenchymal transitions (EMTs) that occur in normal human and mouse epithelial cells, and promotes migration in keratinocytes, monocytes, and mast cells (14,18,19,20). Furthermore, activin A induces matrix metalloproteinase (MMP) expression in macrophages (21) and in endometrial cells during decidualization and trophoblast invasion (22).

Activin A also regulates the metastatic phenotype in ovarian cancer cells. Steller et al. (23) demonstrated that activin A stimulated invasion in the SKOV3 and OCC1 cell lines. The same study also proposed that inhibin A could repress activin-stimulated invasion in SKOV3 cells, and not OCC1 cells, because activin receptors were overexpressed in OCC1 cells compared with SKOV3 cells.

The role of activin A in regulating ovarian cancer progression has remained relatively unexplored, despite clinical studies that report the elevation of activin levels in patient serum, cyst fluid, and peritoneal fluid (24,25,26). These observations suggest that up-regulation of activin signaling plays a role in ovarian oncogenesis.

In vitro studies reveal different roles for activin A in the regulation of ovarian cancer cell proliferation. Activin A stimulated proliferation in cell lines derived from ovarian serous carcinomas, but not normal OSE (17,23). However, other reports indicate that activin A has no effect on proliferation in a clear cell adenocarcinoma cell line (27) and a negative effect on proliferation in an endometrioid adenocarcinoma cell line (28). Another study using three endometrioid adenocarcinoma cell lines reports that activin A has no effect on proliferation but can reverse the growth inhibitory effects of TGF-β (29).

In the current study, the role of Smad-independent activin A signaling in serous ovarian oncogenesis was investigated. Activin A stimulated cell proliferation, and attenuation of GSK activity enhanced proliferation in the serous adenocarcinoma cell line, OVCA429. Activation of Akt resulted in the phosphorylation of GSK-β at Ser9/21, repressing GSK activity. Because the serous subtype occurs most frequently, comprising 60% of the ovarian carcinomas in developed countries (30), we analyzed tissue microarrays consisting of benign serous cystadenomas, borderline tumors, and cystadenocarcinomas. Our analyses revealed that phospho-GSKα/β (Ser9/21) levels (P < 0.0001), but not βA-subunit expression (P = 0.40), were different among benign neoplasms, borderline tumors, and cystadenocarcinomas. Although βA expression did not correlate with survival, βA was detected in stromal cells from carcinomas, but not benign tumors. Lower phospho-GSK levels in carcinoma patients correlated with shorter survival (P = 0.046). Furthermore, phospho-GSK levels were elevated in mitotically active epithelial cells in carcinoma tissues, consistent with in vitro results. Polarized phospho-GSK immunostaining (at terminal bar structures) was present in benign epithelia but absent in carcinomas. Collectively, these data suggest that activin A may contribute to ovarian oncogenesis by regulating Akt and GSK activities to mediate cellular proliferation.

Materials and Methods

Patient characteristics and ovarian tissue microarrays

Human tissues were used in experiments after approval by an institutional review board. All tissues were retrospectively obtained from surgical specimens and rereviewed and confirmed by a pathologist (C.D.S.). Tissue microarrays were created from these archival, paraffin-embedded, human ovarian tissues from Evanston Northwestern Healthcare (Evanston, IL) and classified into three categories: benign serous cystadenoma (n = 44), borderline serous tumor (n = 36), and malignant serous cystadenocarcinoma (n = 40). Core tissue punches (0.6 μm) were used to assemble the microarray, with three core punches for each patient in the borderline and carcinoma category, and five punches for each patient in the benign category. The mean ages for patients included in the benign tumor, borderline tumor, and cystadenocarcinoma microarrays were 58.0 ± 13.7 yr (range 27–86), 51.6 ± 15.7 yr (range 27–85), and 62.4 ± 11.8 yr (range 37–82), respectively. Tumors were analyzed and staged by a pathologist according to the criteria of the Federation Internationale de Gynecologie et d’Obstetrique. The borderline microarray consisted of two stage I, 13 stage IA, four stage IB, nine stage IC, one stage IIB, three stage IIC, three stage IIIA, and one stage IIIB. The carcinoma microarray consisted of one stage IA, one stage IIC, three stage III, two stage IIIB, 31 stage IIIC, and two stage IV patients. For all protein markers analyzed by immunohistochemistry (IHC), positive staining was defined as staining present in greater than or equal to 25% of epithelia within a tissue core. Protein expression was scored for each tissue core on a scale: zero (no staining), one (weak to moderate staining), and two (strong staining). IHC results were evaluated and scored by two investigators, and then the scores for each patient’s tissue cores were averaged, so that each patient was assigned one average score for each marker. Only those patients who had at least two tissue cores, which contained epithelia that could be scored for immunostaining, were used for data analysis. Because the number of tissue cores that could be scored varied from slide to slide for each microarray, the total number of patients included in data analysis for a specific IHC marker could be less than the total number of patients included in each microarray.

Statistical analyses

Statistical analyses were performed using SAS 9.1 (SAS Institute Inc., Cary, NC) and Prism 4 (GraphPad Software Inc., San Diego, CA). For comparison of average scores among the three microarrays, Fisher’s exact test or χ2 tests were performed. Survival curves and analyses were generated for serous cystadenocarcinoma patients using the Kaplan-Meier estimate and log-rank tests.

Cell culture and treatments

The OVCA429 cell line, a generous gift from Dr. M. Sharon Stack (Northwestern University, Chicago, IL), was cultured as previously described (31). This cell line was isolated from the ascites of a patient with late-stage serous adenocarcinoma (32). For all experiments, cells were cultured to 80% confluency and serum-starved for 24 h before treatment. Cells were pretreated with the inhibitors LY294002 (Calbiochem, San Diego, CA) and SB216763 (Sigma-Aldrich, St. Louis, MO) for 1 h before culturing cells in the presence of inhibitor and human activin A (purified in our laboratory) for 72 h. For experiments using human Fst-288 (a gracious gift from T. Lerch and Dr. T. Jardetzky, Stanford University, Los Altos Hills, CA), Fst-288 was added at the same time as activin A.

Immunofluorescence and imaging techniques

Cells were cultured on glass coverslips, rinsed with PBS, fixed in methanol for 2 min, blocked in 1% BSA/PBS for 1 h, and then incubated at room temperature for 1 h with anti-E-cadherin (Invitrogen Corp., Carlsbad, CA) at 1:100 in blocking solution. Coverslips were then rinsed in PBS, incubated with donkey-antimouse-Cy3 secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), and mounted in Vectashield media (Vector Laboratories, Burlingame, CA). Images were acquired on a Nikon Eclipse TE2000-U microscope (Nikon Corp., Melville, NY) using the SPOT digital camera (Diagnostic Instruments, Sterling Heights, MI) and Metamorph 5.0 software (Molecular Devices, Downington, PA).

Immunoblotting

Whole cell lysates were prepared by lysing cells in 50 mmol/liter Tris-HCl (pH 7.5), 150 mmol/liter NaCl, 1% Triton X-100, and 0.1% sodium dodecyl sulfate supplemented with phosphatase (Sigma-Aldrich) and protease (Roche Diagnostics, Mannheim, Germany) inhibitor cocktail. Lysates were separated on 4–12% polyacrylamide gradient gels (Invitrogen) and then transferred to nitrocellulose. Blots were blocked for 1 h in 5% milk/0.1% Tris-buffered saline (TBS) with 0.005% Tween 20 (TBS-T), and then incubated with primary antibody in 3% BSA/TBS overnight at 4 C. Blots were then washed thrice for 15 min in TBS-T, incubated in secondary antibody in 5% milk/TBS-T for 1 h, washed thrice for 15 min in TBS-T, then twice for 10 min in TBS, and developed with ECL Plus Reagent (Amersham, Buckinghamshire, UK). Anti-phospho-Akt, anti-phospho-GSKα/β (Ser9/21), anti-Akt, and anti-GSKα/β (Cell Signaling Technologies, Beverly, MA) were used at 1:1000 dilutions.

Cell proliferation assay

Cells were cultured in 96-well plates (8000 cells per well) overnight in complete media before being rinsed with PBS and replaced with Opti-MEM (Invitrogen). Proliferation was assayed using Promega’s MTS assay (Madison, WI), according to the manufacturer’s instructions over a period of 3 d, except cells were incubated for 3 h before reading plates on an EL312e Biokinetics microplate reader at 490 nm (Bio-Tek Instruments, Winooski, VT).

IHC

IHC was performed as previously described (33). Briefly, slides were deparaffinized and rehydrated in a graded series of ethanol. Antigen retrieval was performed in a 10 mm sodium citrate buffer (pH 6.0) and permeabilized in TBS-T. Endogenous peroxidase activity was quenched in a 3% H2O2 solution before blocking with the Avidin-Biotin Kit (Vector Laboratories). Slides were incubated for 1 h at room temperature in blocking solution and then incubated, overnight at 4 C, with primary antibody in blocking solution. Slides were then washed in TBS-T and incubated with 2.5 μg/ml biotin-labeled secondary antibody (Vector Laboratories) for 30 min at room temperature. Slides were washed again in TBS-T and incubated in ABC reagent (Vector Laboratories) for 30 min. Horseradish peroxidase was detected with the diaminobenzidine reagent kit (Vector Laboratories), and counterstained with Harris-modified hematoxylin (Sigma-Aldrich). Anti-βA-subunit (a generous gift of Dr. Wylie Vale, Columbia, MO) and anti-phospho-GSKα/β (Ser9/21) (Cell Signaling Technologies) were used at 1:100 dilutions.

Results

Activin A activates Akt resulting in GSK-β inhibition

We have previously shown that TGF-β1, -β2, and -β3 induce an EMT and MMP-2 and MMP-9 secretion in the ovarian cancer cell line, OVCA429 (49). To determine if activin A exerts similar or different effects, we analyzed the effects of activin A on cell morphology in the human serous adenocarcinoma cell line, OVCA429. Cell morphology was evaluated by analyzing the localization of the adherens junction protein, E-cadherin, by indirect immunofluorescence. TGF-β1 induced an EMT (i.e. a transition to a fibroblastic morphology), resulting in the loss of E-cadherin staining at sites of cell-cell junctions, as demonstrated by the absence of E-cadherin staining (Fig. 1A). In contrast, activin-treated cells resembled vehicle-treated cells, retaining E-cadherin staining at cell-cell junctions. Activin A also selectively induced pro-MMP-9 secretion, which was dependent upon Akt activity (data not shown). Thus, activin and TGF-β1 both induce MMP secretion, but activin A does not regulate the transition to a fibroblastic morphology.

Figure 1.

Figure 1

Open in a new tab

Effects of activin A treatment on cell morphology and Akt/GSK-β activity. A, Anti-E-cadherin indirect immunofluorescence of cells treated with vehicle (Veh), activin A (Act) (10 ng/ml), and TGF-β1 (β1) (10 ng/ml) for 72 h showing adherens junctions. B, Activin A activates Akt, leading to suppression of GSK activity. Quiescent cells were treated with activin A (10 ng/ml) for 0, 10, 20, 30, 45, and 60 min, and then probed with anti-phospho-Akt and anti-phospho-GSKα/β (Ser9/21) antibodies. The same blots were reprobed with either anti-Akt or anti-GSKβ as a loading control. C, Fst-288 inhibits activin A stimulation of Akt phosphorylation in a dose-dependent manner. Lysates from cells treated with increasing concentrations of Fst-288 (10, 20, 40, and 60 ng/ml) and activin A (10 ng/ml) for 30 min were immunoblotted with anti-phospho-Akt. The same blot was reprobed with anti-Akt as a loading control. D, Inhibition of the PI3K/Akt pathway with LY294002 (LY) (1 and 10 μm) represses phosphorylation of GSKβ in a dose-dependent manner. Anti-phospho-GSK immunoblot of lysates from cells treated with activin A and LY294002 for 30 min. The same blot was reprobed with anti-GSK as a loading control.

PI3K/Akt signaling is often up-regulated in ovarian cancer cells, and plays an important role in ovarian oncogenesis (34,35,36,37,38,39). Therefore, the activation of Akt, a downstream mediator of PI3K, by activin A was analyzed. When ovarian cancer cells were treated with activin A over a time course of 60 min, phosphorylation of Akt was observed starting at 10 min, and continued to increase over the time course (Fig. 1B).

Akt is known to phosphorylate GSK-α/β at serine residues 9 and 21 to inhibit GSK activity (40,41). Therefore, the phosphorylation of GSK-α/β at these inhibitory residues was evaluated over a time course of 60 min. A very low level of serine phosphorylation of GSK-α was detected (Fig. 1B). However, GSK-β phosphorylation, which was detectable after 10-min activin treatment, was robust by 30 min, and persisted for up to 60 min. Down-regulation of GSK-β activity contributes to EMTs in other cell types, but activin A did not mediate an EMT in OVCA429 cells, as described previously (Fig. 1A).

To demonstrate the specificity of activin A mediating these biological responses, Fst-288, an activin A antagonist, was used to inhibit activin signaling. Fst-288 was able to down-regulate activin A-induced phosphorylation of Akt, in a dose-dependent manner, when compared with activin treatment alone (Fig. 1C). Furthermore, inhibition of the PI3K/Akt pathway with the PI3K inhibitor, LY294002, attenuated serine phosphorylation of GSK in a dose-dependent manner (Fig. 1D).

Collectively, these results demonstrate the specificity of the effects of activin A on the activation of Akt to repress GSK.

βA-subunit and phospho-GSK-α/β (Ser9/21) immunoreactivity in epithelia from serous benign tumors, borderline tumors, and cystadenocarcinomas

To determine whether activin A played a role in disease development in vivo, IHC was performed to assay βA-subunit and phospho-GSK-α/β (Ser9/21) levels in ovarian benign serous cystadenoma, borderline tumor, and cystadenocarcinoma tissue microarrays. IHC using anti-phospho-Akt antibody, concurrent with the anti-βA and anti-phospho-GSK-α/β (Ser9/21) antibodies, did not yield a detectable signal in any of the tissue microarrays; therefore, this marker could not be analyzed. For the βA and phospho-GSK biomarkers, immunostaining for each tissue core was classified as zero (no staining), one (weak to moderate staining), or two (strong staining). Only those patients with at least two tissue cores that displayed scorable immunostaining were included in our data analyses, and the scores for all tissue cores for each patient were averaged. The average scores for benign tumors, borderline tumors, and carcinomas were then divided into three categories: 1) no expression (average score is zero); 2) low expression (average score < 1.5); and 3) high expression (average score ≥ 1.5).

Analysis of βA-subunit expression in benign tumors, borderline tumors, and carcinomas revealed that there was not a statistically significant difference among the three tissue types (P = 0.40), and pairwise comparison yielded similar results (Table 1). Low βA-subunit expression was detected in 37% of benign tumors, 22% borderline tumors, and 25% carcinomas. High βA-subunit expression was detected in 63% of benign tumors, 78% borderline tumors, and 75% carcinomas (Table 1). Representative bright-field images of βA-subunit expression in benign tumor (Fig. 2A), borderline tumor (Fig. 2B), and cystadenocarcinoma (Fig. 2C) show the βA-subunit exhibits punctate cytoplasmic staining in all epithelia. Furthermore, borderline tumors and carcinomas exhibited staining in stromal cells (Fig. 2, B and C).

Table 1.

βA and phospho-GSKα/β (Ser9/21) expression in benign serous cystadenomas, borderline tumors, and cystadenocarcinomas

Frequency of Average Scores P Total No. Patients
βA-subunit 0 <1.5 ≥1.5
Benign 0 10 17 0.26 (Benign/borderline) 27
Borderline 0 8 28 0.79 (Borderline/carcinoma) 36
Cystadenocarcinoma 0 10 30 0.41 (Benign/carcinoma) 40
Phospho-GSKα/β
Benign 0 12 16 0.0011 (Benign/borderline) 28
Borderline 1 30 5 <0.0001 (Borderline/carcinoma) 36
Cystadenocarcinoma 16 24 0 <0.0001 (Benign/carcinoma) 40
Open in a new tab

Fisher’s exact and χ-square tests were used to analyze differences in βA subunit and phospho-GSK expression, respectively, between tissue categories. 

Figure 2.

Figure 2

Open in a new tab

βA-subunit expression in benign tumors, borderline tumors, and cystadenocarcinomas. The majority of benign tumors (A), borderline tumors (B), and cystadenocarcinomas (C) exhibit high levels of βA expression. Stromal cells in borderline tumors and carcinomas also express βA. Representative bright-field images at ×10 (left; bar, 100 μm) and ×40 (right; bar, 20 μm) magnifications are shown.

In contrast, phospho-GSK-α/β (Ser9/21) immunostaining was significantly different among the three tissue categories, with benign neoplasias exhibiting higher phospho-GSK-α/β (Ser9/21) immunostaining than carcinomas (P < 0.0001; Table 1). Pairwise comparison revealed that there was a statistically significant difference in immunostaining between benign and borderline tumors (P = 0.001), benign tumors and carcinomas (P < 0.0001), and borderline tumors and carcinomas (P < 0.0001). No phospho-GSK-α/β immunoreactivity was detected in 0% of benign neoplasias, 2.7% borderline tumors, and 40% carcinomas (Table 1). Low phospho-GSK levels were found in 42.9% of benign neoplasias, 83.8% borderline tumors, and 60% carcinomas. High phospho-GSK levels were detected in 57.1% of benign neoplasias, 13.5% borderline tumors, and 0% carcinomas (Table 1).

Bright-field images showing phospho-GSK-α/β immunostaining in the three categories of tissue are presented in Fig. 3A. In benign epithelia, and less frequently in borderline epithelia, phospho-GSK-α/β immunoreactivity is present at the terminal bar, the apical tight junction structure in polarized epithelia observed by light microscopy (Fig. 3B, Polarized panel, and Table 2). Polarized phospho-GSK staining was present in 25.8% of benign tumor tissue cores and 3.92% borderline tumor tissue cores (Table 2). Benign and borderline epithelia also displayed mixed phospho-GSK-α/β staining, in which both diffuse cytoplasmic and polarized localization at the terminal bar was observed (Fig. 3B, Mixed panel, and Table 2). Mixed phospho-GSK localization was exhibited in 12.4% of benign tumor tissue cores and 8.82% borderline tumor tissue cores (Table 2). In contrast, phospho-GSK-α/β immunoreactivity in carcinomas exhibited diffuse, cytoplasmic staining, but never the polarized localization (Fig. 3B, Diffuse panel). Diffuse phospho-GSK localization was present in 100, 87.3, and 61.8% of carcinoma, borderline tumor, and benign tumor tissue cores, respectively (Table 2).

Figure 3.

Figure 3

Open in a new tab

Activin regulation of proliferation and phospho-GSKα/β (Ser9/21) immunoreactivity in benign tumors, borderline tumors, and cystadenocarcinomas. A, Representative bright-field images (×10) of phospho-GSK (Ser9/21) immunoreactivity in benign tumors, borderline tumors, and cystadenocarcinomas (bar, 100 μm). B, Phospho-GSK (Ser9/21) immunostaining (×40) exhibits polarized, mixed, and diffuse localization. Polarized localization (black arrow) at the terminal bar occurs more frequently in benign tumors than borderline tumors or cystadenocarcinomas. Carcinomas exhibit diffuse (black arrowhead), cytoplasmic phospho-GSK staining. Benign and borderline tumors exhibit mixed localization (a combination of diffuse and polarized staining; bar, 20 μm). C, Phospho-GSK (Ser9/21) levels (×40) are often elevated in mitotically active carcinoma cells (black arrowheads). D, Activin A treatment stimulates cellular proliferation in ovarian cancer cells in vitro. MTS assay in cells treated with vehicle (Veh) or activin A (ActA) (10 ng/ml) over a period of 3 d. Columns, means relative to d 0; bars, se (n = 3). E, Inhibition of GSK activity is sufficient to promote cellular proliferation and increases activin-stimulated proliferation. MTS assay for proliferation in cells treated with activin A (10 ng/ml) and the GSK inhibitor, SB216763 (SB2) (3 μm), over a period of 3 d. Columns, means relative to vehicle treatment; bars, se (n = 3).

Table 2.

Phospho-GSKα/β (Ser9/21) localization in benign cystadenomas, borderline tumors, and cystadenocarcinomas

Phospho-GSKα/β Localization
P Total No. Cores
Polarized Diffuse Mixed
Benign 23 55 11 <0.0001 (Benign/borderline) 89
Borderline 4 89 9 0.002 (Borderline/carcinoma) 102
Cystadenocarcinoma 0 91 0 <0.0001 (Benign/carcinoma) 91
Open in a new tab

Localization of phospho-GSK immunostaining was scored for each tissue core. Polarized, staining predominantly at the terminal bar. Diffuse, diffuse cytoplasmic staining. Mixed, combination of diffuse and polarized staining. Differences in phospho-GSK localization between tissue categories were analyzed using χ-square test. 

Carcinomas also displayed phospho-GSK-α/β staining that was frequently more intense in mitotically active cells when compared with surrounding epithelia (Fig. 3C). Therefore, we analyzed the effects of activin A treatment and the effects of inhibiting GSK activity on cellular proliferation. Activin A stimulated proliferation in ovarian cancer cells in a statistically significant manner by the second and third days of treatment (Fig. 3D). Proliferation increased 30.9% (P < 0.001) and 17.5% (P < 0.001) by d 2 and 3 of treatment, respectively. In addition, inhibition of GSK alone was sufficient to up-regulate proliferation as efficiently as activin A treatment alone (Fig. 3E). Treatment with the GSK inhibitor, SB216763, was able to modestly enhance activin A-regulated proliferation in a statistically significant manner.

In summary, our in vitro and in vivo observations suggest that repression of GSK activity positively regulates cellular proliferation. Furthermore, the differential immunoreactivity of phospho-GSK-α/β (Ser9/21) in benign cystadenomas, borderline tumors, and carcinomas raises the possibility that inactivation of GSK may play a role in ovarian neoplastic and malignant transformation.

Phospho-GSK, but not βA, levels correlate with survival in carcinoma patients

The relationships between average βA-subunit or phospho-GSK-α/β (Ser9/21) levels and survival were analyzed to see if these markers correlate with prognosis. For the βA-subunit, patients exhibiting low expression (average score < 1.5) had a median survival time of 40.6 months, and patients exhibiting high expression (average score ≥ 1.5) had a median survival time of 50.9 months. The log-rank test revealed no difference (P = 0.06) in survival between patients expressing low vs. high levels of the βA-subunit (Fig. 4A). For the phospho-GSK marker, there was no difference (P = 0.12) in survival between patients that did not express this marker (score = 0) compared with patients that expressed the marker (score > 0; P = 0.77). Because no cancer patients expressed high levels of phospho-GSK, we compared patients expressing very low levels of phospho-GSK (average scores < 1) and patients expressing low to moderate levels (average scores ≥ 1). Patients expressing very low phospho-GSK levels had longer median survival times (50.1 months) than patients expressing moderate levels (40.6 months) of phospho-GSK (P = 0.046; Fig. 4B). Collectively, these observations suggest that phospho-GSK may prove useful as both a diagnostic and prognostic marker for serous ovarian cancer.

Figure 4.

Figure 4

Open in a new tab

Relationship between βA or phospho-GSK (Ser9/21) levels and survival in cystadenocarcinoma patients. A, Curves comparing survival in patients with low of βA expression (average score < 1.5) and high βA expression (average score ≥ 1.5). B, Curves comparing survival in patients exhibiting very low phospho-GSK (Ser9/21) levels (average score < 1) and low to moderate levels (average score ≥ 1). For both markers, scores for all tissue cores from each patient were averaged and then subjected to Kaplan-Meier estimate and log-rank test.

Discussion

Because ovarian carcinomas are clinically heterogeneous, it is critical to elucidate the molecular pathways that contribute to disease development and outcome so that more effective individualized therapies can be implemented to improve patient survival. It is equally important to identify the pathological consequences of activating these molecular pathways as an integral part of identifying novel therapeutic targets.

Clinical studies reporting elevated activin A levels in the ascites and serum of ovarian cancer patients suggest that this cytokine mediates pro-oncogenic functions (24,25,26). The current study investigated the role of activin A in ovarian tumor biology by analyzing activin-regulated pathways in serous benign tumors, borderline tumors, and cystadenocarcinomas. The pro-oncogenic effects of activin/Akt/GSK signaling were assessed by assaying cellular proliferation in vitro.

Evaluation of ovarian tissue microarrays demonstrates, for the first time, that phospho-GSKα/β (Ser9/21) levels may serve as a diagnostic and prognostic marker for serous ovarian cancer. Phospho-GSK levels were statistically different among benign cystadenoma, borderline tumor, and cystadenocarcinoma. The dramatic loss and down-regulation of phospho-GSK staining in carcinoma compared with benign neoplasia may reflect the fact that carcinoma cells possess more epithelial characteristics than the putative precursor OSE, which have a more uncommitted phenotype. Normal OSE rarely express the epithelial marker, E-cadherin, whereas malignant epithelia acquire expression of E-cadherin during tumorigenesis (8). Loss of phospho-GSK immunoreactivity indicates up-regulation of GSK activity, which would promote the expression of the epithelial marker, E-cadherin (42).

Patients with carcinomas displaying higher levels of phospho-GSK-α/β (Ser9/21) had shorter median survival times than those displaying lower levels (P = 0.046). Enhanced inactivation of GSK in mitotically active cancer epithelia was evident by elevated phospho-GSK levels in these cells compared with surrounding epithelia. Down-regulation of GSK activity also modestly promoted activin-regulated ovarian cancer cell proliferation in vitro. This raises the intriguing possibility that poor survival for patients with carcinomas expressing higher levels of phospho-GSK may be due to increased tumor cell proliferation. Activation of Akt and repression of GSK-β promote both survival and proliferation in lung cancer, lymphoblastic leukemia, and multiple myeloma cells in vitro (43,44,45).

Phospho-GSK staining in carcinoma cells was completely devoid of the polar localization at terminal bar structures observed in benign epithelia. It will be important to determine if the loss of polarized localization of GSK-β may play a role in the loss of epithelial cell polarity in ovarian cancer, which would promote invasion and metastasis. In both Xenopus and Drosophila embryos, Wnt inactivation of GSK-β signaling determines polarity during gastrulation. In addition, the polarity of neurons is established and maintained by Akt and PTEN regulation of GSK-β activity (46). Localized PI3K activation at the growth cone inactivates nearby GSK-β molecules to modulate the function of adenomatous polyposis coli, a microtubule plus end binding protein, and initiates directed axon growth (47). These observations suggest that localized repression of GSK-β plays a critical role in mediating neuronal cell polarity. Regulation of GSK in ovarian cancer may be complex because it is possible that cross talk between Wnt and activin signaling cascades mediates GSK-β activity.

In contrast to phospho-GSK, βA-subunit expression was similar among benign tumors, borderline tumors, and cystadenocarcinoma, and did not correlate with survival. Clinical studies that report elevated activin expression in patient ascites and serum may be due, in part, to cancer cell proliferation, which increases the number of epithelial cells producing βA. In addition, expression of βA observed in stromal cells from carcinoma tissue, but not from benign neoplasias, could also be a source of aberrant activin secretion that may exert effects on surrounding epithelia.

Di Simone et al. (48) published a study analyzing the expression of activin, Fst, and ActRII in six ovarian cancer cell lines. All cell lines analyzed in this study expressed the ActRII receptor. Furthermore, ovarian cancer cell lines that secreted activin, but not Fst, proliferated in response to exogenous activin (up to 100 ng/ml). Cell lines that secreted Fst did not respond to exogenous activin, presumably due to endogenous Fst binding to exogenous activin. However, this study did not determine activin protein secretion in the cell lines that secreted Fst. Based on these observations, Di Simone et al. (48) proposed that activin may mediate autocrine signaling cascades to promote ovarian cancer cell proliferation. OVCA429 cells secrete the βA-subunit of activin as assayed by Western blot analysis of conditioned media (unpublished data), and respond to exogenous activin, suggesting that this cell line does not secrete endogenous Fst.

Additional studies will need to be performed to define clearly the role of activin A in mediating ovarian tumor development, and studies using larger cohorts will be necessary to confirm and extend our observations. Furthermore, it will be important to evaluate the expression of these biomarkers in other histological subtypes of ovarian cancer to determine if these molecular determinants are unique to serous cancers.

In conclusion, the current study implicates a role for activin A signaling in promoting ovarian oncogenesis by mediating cellular proliferation. The degree of GSK-β phosphorylation at Ser9/21 is different among epithelia of different pathobiological status, and may serve as an important diagnostic and prognostic biomarker.

Footnotes

This work was supported by National Institutes of Health Grant HD044464 (to T.K.W.).

Disclosure Statement: The authors have nothing to disclose.

First Published Online May 1, 2008

Abbreviations: EMT, Epithelial-to-mesenchymal transition; Fst, follistatin; IHC, immunohistochemistry; MMP, matrix metalloproteinase; OSE, ovarian surface epithelia; PI3K, phosphatidylinositol 3-kinase; TBS, Tris-buffered saline; TBS-T, Tris-buffered saline in Tween 20.

References

  1. Leung PC, Choi JH 2007 Endocrine signaling in ovarian surface epithelium and cancer. Hum Reprod Update 13:143–162 [DOI] [PubMed] [Google Scholar]
  2. Mok SC, Elias KM, Wong KK, Ho K, Bonome T, Birrer MJ 2007 Biomarker discovery in epithelial ovarian cancer by genomic approaches. Adv Cancer Res 96:1–22 [DOI] [PubMed] [Google Scholar]
  3. Bast Jr RC, Brewer M, Zou C, Hernandez MA, Daley M, Ozols R, Lu K, Lu Z, Badgwell D, Mills GB, Skates S, Zhang Z, Chan D, Lokshin A, Yu Y 2007 Prevention and early detection of ovarian cancer: mission impossible? Recent Results Cancer Res 174:91–100 [DOI] [PubMed] [Google Scholar]
  4. Cvetkovic D 2003 Early events in ovarian oncogenesis. Reprod Biol Endocrinol 1:68 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bast Jr RC 2003 Status of tumor markers in ovarian cancer screening. J Clin Oncol 21(Suppl):200s–205s [DOI] [PubMed] [Google Scholar]
  6. Seidman JD, Russel P, Kurman RJ 2002 Surface epithelial tumors of the ovary. In: Kurman RJ, ed. Blaustein’s pathology of the female genital tract. 5th ed. New York: Springer-Verlag; 791–904 [Google Scholar]
  7. Scully RE 1999 World Health Organization international histological classification of tumors. 2nd ed. New York: Springer [Google Scholar]
  8. Auersperg N, Wong AS, Choi KC, Kang SK, Leung PC 2001 Ovarian surface epithelium: biology, endocrinology, and pathology. Endocr Rev 22:255–288 [DOI] [PubMed] [Google Scholar]
  9. Peralta Soler A, Knudsen KA, Tecson-Miguel A, McBrearty FX, Han AC, Salazar H 1997 Expression of E-cadherin and N-cadherin in surface epithelial-stromal tumors of the ovary distinguishes mucinous from serous and endometrioid tumors. Hum Pathol 28:734–739 [DOI] [PubMed] [Google Scholar]
  10. Wong AS, Maines-Bandiera SL, Rosen B, Wheelock MJ, Johnson KR, Leung PC, Roskelley CD, Auersperg N 1999 Constitutive and conditional cadherin expression in cultured human ovarian surface epithelium: influence of family history of ovarian cancer. Int J Cancer 81:180–188 [DOI] [PubMed] [Google Scholar]
  11. Auersperg N, Pan J, Grove BD, Peterson T, Fisher J, Maines-Bandiera S, Somasiri A, Roskelley CD 1999 E-cadherin induces mesenchymal-to-epithelial transition in human ovarian surface epithelium. Proc Natl Acad Sci USA 96:6249–6254 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Dupont J, McNeilly J, Vaiman A, Canepa S, Combarnous Y, Taragnat C 2003 Activin signaling pathways in ovine pituitary and LβT2 gonadotrope cells. Biol Reprod 68:1877–1887 [DOI] [PubMed] [Google Scholar]
  13. Peng C 2003 The TGF-β superfamily and its roles in the human ovary and placenta. J Obstet Gynaecol Can 25:834–844 [DOI] [PubMed] [Google Scholar]
  14. Zhang L, Deng M, Parthasarathy R, Wang L, Mongan M, Molkentin JD, Zheng Y, Xia Y 2005 MEKK1 transduces activin signals in keratinocytes to induce actin stress fiber formation and migration. Mol Cell Biol 25:60–65 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Thompson TB, Cook RW, Chapman SC, Jardetzky TS, Woodruff TK 2004 βA versus βB: is it merely a matter of expression? Mol Cell Endocrinol 225:9–17 [DOI] [PubMed] [Google Scholar]
  16. Pangas SA, Woodruff TK 2000 Activin signal transduction pathways. Trends Endocrinol Metab 11:309–314 [DOI] [PubMed] [Google Scholar]
  17. Choi KC, Kang SK, Nathwani PS, Cheng KW, Auersperg N, Leung PC 2001 Differential expression of activin/inhibin subunit and activin receptor mRNAs in normal and neoplastic ovarian surface epithelium (OSE). Mol Cell Endocrinol 174:99–110 [DOI] [PubMed] [Google Scholar]
  18. Petraglia F, Sacerdote P, Cossarizza A, Angioni S, Genazzani AD, Franceschi C, Muscettola M, Grasso G 1991 Inhibin and activin modulate human monocyte chemotaxis and human lymphocyte interferon-γ production. J Clin Endocrinol Metab 72:496–502 [DOI] [PubMed] [Google Scholar]
  19. Olsson N, Piek E, ten Dijke P, Nilsson G 2000 Human mast cell migration in response to members of the transforming growth factor-β family. J Leukoc Biol 67:350–356 [DOI] [PubMed] [Google Scholar]
  20. Hyuga S, Kawasaki N, Hashimoto O, Hyuga M, Ohta M, Yamagata S, Yamagata T, Hayakawa T 2000 Possible role of hepatocyte growth factor/scatter factor and activin A produced by the target organ in liver metastasis. Cancer Lett 153:137–143 [DOI] [PubMed] [Google Scholar]
  21. Ogawa K, Funaba M, Mathews LS, Mizutani T 2000 Activin A stimulates type IV collagenase (matrix metalloproteinase-2) production in mouse peritoneal macrophages. J Immunol 165:2997–3003 [DOI] [PubMed] [Google Scholar]
  22. Jones RL, Findlay JK, Farnworth PG, Robertson DM, Wallace E, Salamonsen LA 2006 Activin A and inhibin A differentially regulate human uterine matrix metalloproteinases: potential interactions during decidualization and trophoblast invasion. Endocrinology 147:724–732 [DOI] [PubMed] [Google Scholar]
  23. Steller MD, Shaw TJ, Vanderhyden BC, Ethier JF 2005 Inhibin resistance is associated with aggressive tumorigenicity of ovarian cancer cells. Mol Cancer Res 3:50–61 [PubMed] [Google Scholar]
  24. Menon U, Riley SC, Thomas J, Bose C, Dawnay A, Evans LW, Groome NP, Jacobs IJ 2000 Serum inhibin, activin and follistatin in postmenopausal women with epithelial ovarian carcinoma. BJOG 107:1069–1074 [DOI] [PubMed] [Google Scholar]
  25. Chudecka-Glaz A, Rzepka-Gorska I 2005 Activin A levels in serum and cyst fluid in epithelial tumors of the ovary. Int J Gynaecol Obstet 89:160–162 [DOI] [PubMed] [Google Scholar]
  26. Cobellis L, Reis FM, Luisi S, Danero S, Pirtoli L, Scambia G, Petraglia F 2004 High concentrations of activin A in the peritoneal fluid of women with epithelial ovarian cancer. J Soc Gynecol Investig 11:203–206 [DOI] [PubMed] [Google Scholar]
  27. Mabuchi Y, Yamoto M, Minami S, Umesaki N 2006 The autocrine effect of activin A on human ovarian clear cell adenocarcinoma cells. Oncol Rep 16:373–379 [PubMed] [Google Scholar]
  28. Tanaka T, Toujima S, Umesaki N 2004 Growth-inhibitory signals by activin A do not affect anticancer drug-sensitivity and acquired multi-drug-resistance in human ovarian endometrioid adenocarcinoma OVK-18 cells. Oncol Rep 11:667–671 [PubMed] [Google Scholar]
  29. Tanaka T, Toujima S, Umesaki N 2004 Activin A inhibits growth-inhibitory signals by TGF-β1 in differentiated human endometrial adenocarcinoma cells. Oncol Rep 11:875–879 [PubMed] [Google Scholar]
  30. Shih IeM, Kurman RJ 2005 Molecular pathogenesis of ovarian borderline tumors: new insights and old challenges. Clin Cancer Res 11:7273–7279 [DOI] [PubMed] [Google Scholar]
  31. Moser TL, Young TN, Rodriguez GC, Pizzo SV, Bast Jr RC, Stack MS 1994 Secretion of extracellular matrix-degrading proteinases is increased in epithelial ovarian carcinoma. Int J Cancer 56:552–559 [DOI] [PubMed] [Google Scholar]
  32. Lau KM, Mok SC, Ho SM 1999 Expression of human estrogen receptor-α and -β, progesterone receptor, and androgen receptor mRNA in normal and malignant ovarian epithelial cells. Proc Natl Acad Sci USA 96:5722–5727 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pangas SA, Rademaker AW, Fishman DA, Woodruff TK 2002 Localization of the activin signal transduction components in normal human ovarian follicles: implications for autocrine and paracrine signaling in the ovary. J Clin Endocrinol Metab 87:2644–2657 [DOI] [PubMed] [Google Scholar]
  34. Cheng JQ, Godwin AK, Bellacosa A, Taguchi T, Franke TF, Hamilton TC, Tsichlis PN, Testa JR 1992 AKT2, a putative oncogene encoding a member of a subfamily of protein-serine/threonine kinases, is amplified in human ovarian carcinomas. Proc Natl Acad Sci USA 89:9267–9271 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Wang Y, Helland A, Holm R, Kristensen GB, Borresen-Dale AL 2005 PIK3CA mutations in advanced ovarian carcinomas. Hum Mut 25:322 [DOI] [PubMed] [Google Scholar]
  36. Bellacosa A, de Feo D, Godwin AK, Bell DW, Cheng JQ, Altomare DA, Wan M, Dubeau L, Scambia G, Masciullo V, Ferrandina G, Benedetti Panici P, Mancuso S, Neri G, Testa JR 1995 Molecular alterations of the AKT2 oncogene in ovarian and breast carcinomas. Int J Cancer 64:280–285 [DOI] [PubMed] [Google Scholar]
  37. Shayesteh L, Lu Y, Kuo WL, Baldocchi R, Godfrey T, Collins C, Pinkel D, Powell B, Mills GB, Gray JW 1999 PIK3CA is implicated as an oncogene in ovarian cancer. Nat Genet 21:99–102 [DOI] [PubMed] [Google Scholar]
  38. Philp AJ, Campbell IG, Leet C, Vincan E, Rockman SP, Whitehead RH, Thomas RJ, Phillips WA 2001 The phosphatidylinositol 3′-kinase p85α gene is an oncogene in human ovarian and colon tumors. Cancer Res 61:7426–7429 [PubMed] [Google Scholar]
  39. Levine DA, Bogomolniy F, Yee CJ, Lash A, Barakat RR, Borgen PI, Boyd J 2005 Frequent mutation of the PIK3CA gene in ovarian and breast cancers. Clin Cancer Res 11:2875–2878 [DOI] [PubMed] [Google Scholar]
  40. van Weeren PC, de Bruyn KM, de Vries-Smits AM, van Lint J, Burgering BM 1998 Essential role for protein kinase B (PKB) in insulin-induced glycogen synthase kinase 3 inactivation. Characterization of dominant-negative mutant of PKB. J Biol Chem 273:13150–13156 [DOI] [PubMed] [Google Scholar]
  41. Bakin AV, Tomlinson AK, Bhowmick NA, Moses HL, Arteaga CL 2000 Phosphatidylinositol 3-kinase function is required for transforming growth factor β-mediated epithelial to mesenchymal transition and cell migration. J Biol Chem 275:36803–36810 [DOI] [PubMed] [Google Scholar]
  42. Rosano L, Spinella F, Di Castro V, Nicotra MR, Dedhar S, de Herreros AG, Natali PG, Bagnato A 2005 Endothelin-1 promotes epithelial-to-mesenchymal transition in human ovarian cancer cells. Cancer Res 65:11649–11657 [DOI] [PubMed] [Google Scholar]
  43. Barata JT, Silva A, Brandao JG, Nadler LM, Cardoso AA, Boussiotis VA 2004 Activation of PI3K is indispensable for interleukin 7-mediated viability, proliferation, glucose use, and growth of T cell acute lymphoblastic leukemia cells. J Exp Med 200:659–669 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Kim YG, Kim MJ, Lim JS, Lee MS, Kim JS, Yoo YD 2006 ICAM-3-induced cancer cell proliferation through the PI3K/Akt pathway. Cancer Lett 239:103–110 [DOI] [PubMed] [Google Scholar]
  45. Hideshima T, Nakamura N, Chauhan D, Anderson KC 2001 Biologic sequelae of interleukin-6 induced PI3-K/Akt signaling in multiple myeloma. Oncogene 20:5991–6000 [DOI] [PubMed] [Google Scholar]
  46. Jiang H, Guo W, Liang X, Rao Y 2005 Both the establishment and the maintenance of neuronal polarity require active mechanisms: critical roles of GSK-3β and its upstream regulators. Cell 120:123–135 [DOI] [PubMed] [Google Scholar]
  47. Zhou FQ, Zhou J, Dedhar S, Wu YH, Snider WD 2004 NGF-induced axon growth is mediated by localized inactivation of GSK-3β and functions of the microtubule plus end binding protein APC. Neuron 42:877–879 [DOI] [PubMed] [Google Scholar]
  48. Di Simone N, Crowley Jr WF, Wang QF, Sluss PM, Schneyer AL 1996 Characterization of inhibin/activin subunit, follistatin, and activin type II receptors in human ovarian cancer cell lines: a potential role in autocrine growth regulation. Endocrinology 137:486–494 [DOI] [PubMed] [Google Scholar]
  49. Do TV, Kubba LA, Du H, Sturgis CD, Woodruff TK 2008 Transforming growth factor-β1, transforming growth factor-β2, and transforming growth factor-β3 enhance ovarian cancer metastatic potential by inducing a Smad3-dependent epithelial-to-mesenchymal transition. Mol Cancer Res 6:695–705 [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Endocrinology are provided here courtesy of The Endocrine Society

RESOURCES