This article is mentioned in:

Introduction

Ovarian cancer is the most frequent cause of cancer-related death among all gynecological cancers. Approximately 70% of patients with ovarian cancer are diagnosed at an advanced stage (1). The degree of peritoneal dissemination and chemotherapy-resistant tumors is related to the prognosis of patients with advanced-stage ovarian cancer. The molecular mechanisms that allow ovarian cancer cells to detach from the primary tumor, invade the peritoneal surface and regrow at this site are not yet well understood. Thus, obtaining a better understanding of the molecular events that contribute to tumor invasion and metastasis is crucial for developing novel treatment strategies for ovarian cancer.

The epithelial-mesenchymal-transition (EMT) is a developmental process in which polarized epithelial cells undergo multiple biochemical changes and assume a mesenchymal phenotype, with increased migratory capacity, invasiveness, resistance to apoptosis and the expression of extracellular matrix components (2–6). The EMT is regulated through the downregulation of the epithelial marker E-cadherin by transcriptional repressors such as Snail and Zeb (5). Importantly, the EMT is involved in cell invasion, resistance to chemotherapy, and the formation of side populations of cancer stem-like cells (CSCs) (7). Such CSCs or tumor initiating cells comprise a sub-population of cancer cells within each tumor that have tumorigenic potential in vivo and which are able to recreate the full heterogeneity of the parental tumor (8). However, the molecular mechanism underlying the regeneration of CSCs through EMT has not yet been clarified.

Cluster of differentiation (CD) 24 is a small, heavily glycosylated mucin-like cell surface protein (27–30 amino acids in length), which binds to the membrane via a glycosyl-phosphatidylinositol anchor (9). Under physiological conditions, CD24 was initially identified as a B cell marker (10). Later, it was found to be expressed not only in developing or regenerating tissue, but also in granulocytes, Pre-B cells, keratinocytes and renal tubules (11). CD24 has also been reported to be a ligand for P-selectin, an adhesion receptor on activated endothelial cells and platelets (12,13). These findings suggest that the molecule functionally enhances the metastatic potential of cancer cells. Various studies have identified CD24 as a marker of a poor prognosis in patients with hematological malignancies (14), and various solid tumors, including glioma (15), laryngeal squamous cell carcinoma (16), non-small cell lung cancer (17), breast cancer (18), renal cell carcinoma (19), colorectal cancer (20), bladder carcinoma (21) and epithelial ovarian cancer (22). However, the underlying mechanism has remained elusive. Recent findings have described CD24 as a marker for tumor stem cells in pancreatic and ovarian carcinomas (23,24).

In this study, we demonstrated that CD24 was markedly associated with a more aggressive disease course. The data on CD24 demonstrated that CD24-positive ovarian cancer cells are susceptible to the induction of EMT, and this study provides new knowledge on the function of CD24 in tumor metastasis and its suitability as a therapeutic target in patients with advanced ovarian cancer.

Materials and methods

Tissue samples

Tissue samples were obtained from 174 Japanese patients who underwent surgical resection for primary ovarian carcinomas at Osaka Medical College between January 2001 and December 2009. All patients underwent a primary staging operation, such as total abdominal hysterectomy, bilateral salpingo-oophorectomy, omentectomy, and/or lymphadenectomy. Furthermore, all patients received 6-course conventional TC (paclitaxel and carboplatin) therapies as the adjuvant chemotherapy. The median follow-up duration was 29.2 months (range, 2–122 months). The institutional review board approved this study (no. 0156) and informed consent was obtained from all of the patients. We performed a tissue microarray of the assembly of 5-mm cores retrieved from microscopically selected representative tissues of the primary tumors. The specimens were fixed in 10% formalin and embedded in paraffin. Serial sections cut out from paraffin-embedded blocks were used for routine histopathology. A 4-µm section was cut from a tissue microarray block and was immunohistochemically analyzed to detect the expression of CD24.

Immunohistochemistry

Tumor samples were formalin-fixed and embedded in paraffin. Deparaffinized and rehydrated sections (4-µm) were autoclaved in 0.01 mol/l citrate buffer (pH 6.0) for 15 min at 121°C for antigen retrieval. Endogenous peroxidase activity was blocked with 0.3% solution hydrogen peroxidase in methanol for 30 min. Tumor sections were incubated at 4°C for 12 h with a CD24-specific antibody (1:50 dilution; Thermo Fisher Scientific, Yokohama, Japan). The sections were washed with 1X phosphate-buffered saline (PBS) and incubated with Histofine Simple Stain MAX PO (Multi; Nichirei) for 30 min at room temperature. Finally, the sections were washed with 1X PBS and the signals were visualized by incubation with H2O2/diaminobenzidine substrate solution for 5 min. The sections were counterstained with hematoxylin prior to dehydration and mounting. The evaluation of the immunohistochemical data was performed by two independent pathologists who were blinded to the clinicopathological data. For the immunohistochemistry score, the expression levels of CD24 were assessed using a semiquantitative system based on the staining intensity (0, negative; 1+, weak; 2+, moderate; 3+, strong) and the percentage of cells positively stained cells (0–100%), with the positive expression of CD24 defined as ≥50% of tumor cells having a membranous or cytoplasmic staining intensity of ≥2+, as described by Kim et al (25). Scoring was performed three times per slide for three distinct fields, and the three scores were averaged.

Reagents/antibodies

Cisplatin was also purchased from Sigma-Aldrich (St. Louis, MO, USA). The anti-phospho-Akt (ser473), anti-Akt antibodies, anti-phospho-ERK (Thr202/Tyr204), anti-ERK antibodies, anti-phospho-NF-κB (Ser536), and anti-NF-κB antibodies purchased from Cell Signaling Technology (Beverly, MA, USA).

Cell culture

One human ovarian mucinous adenocarcinoma cancer cell line, Caov-3, which was obtained from the American Type Culture Collection (Rockville, MD, USA), was grown in phenol DMEM containing 10% dextran-coated, charcoal-treated fetal calf serum in a humidified atmosphere of 5% CO2 with 95% air at 37°C.

The expression plasmids and cDNA transfection

We purchased pEZ-M02-CD24 (Gene Copoeia, Rockville, MD, USA) which used Homo sapiens CD24 signal transducer mRNA. For the transfection of each sample, oligomer-Lipofectamine plus complexes were prepared as follows: 4.0 µg of cDNA oligomer were diluted in 250 µl of Opti-MEM (Invitrogen, CA, USA). The oligomer-Lipofectamine plus complexes were added to each of the wells, which contained cells and medium; gentle mixing was then performed by rocking the plate back and forth. The cells were incubated at 37°C in a CO2 incubator for 18 h. The cells were then prepared for each assay.

CD24 mRNA depletion by the plasmid transfection of siRNA

We purchased the Trilencer-27 siRNA plasmid for human CD24 (OriGene, Rockville, MD, USA), as well as scrambled sequence negative control plasmid (OriGene). Caov-3 cells were transfected using the CD24 siRNA Lipofectin method, according to the manufacturer's protocol, which was the same method that was used for the expression of CD24.

The reverse transcription polymerase chain reaction (RT-PCR)

Caov-3 cells were treated with PBS, TGF-β for 24 h. Total RNA was isolated from cells using an RNeasy Protect Mini kit (Qiagen) according to the manufacturer's protocol. The Super Script II Reverse Transcriptase kit (Invitrogen) was used to synthesize cDNA. Polymerase chain reactions were performed with a Platinum PCR SuperMix (Invitrogen) system according to the manufacturer's protocol. The following primers were used for the PCR: CD24 forward, 5′-ACCCACGCAGATTTATTCCA-3′; reverse, 5′-ACCACGAAGAGACTGGCTGT-3′. Snail forward, 5′-GCCTTCAACTGCAAATACTGC-3′; reverse, 5′-CTTCTTGACATCTGAGTGGGTC-3′. Ecad forward, 5′-GGTGGGTGACTACAAAATCAATCTC-3′; reverse, 5′-GGGGCAGTAAGGGCTCTTT-3′. β-actin forward, 5′-GAATTCATGTTTGAGACCTTCAA-3′; reverse, 5′-CCGGATCCATCTCTTGCTCGAAGTCCA-3′. The expression levels of β-actin were assessed as an internal control in all reactions.

Flow cytometry

The cultures from the Caov-3 cell lines were washed with PBS. For some experiments, single cells dissociated from tumor spheres were analyzed by the following method. One million trysinized cells were incubated with 7-amino-actinomycin D (BD Biosciences, NJ, USA) only or 7-amino-actinomycin D and a mouse anti-human monoclonal antibody of CD24-FITC (BD Biosciences) with stain buffer (BD Pharmingen™) for 20 min at room temperature in the dark. After washing, the cells were analyzed using a BD FACS Aria™. The fluorescence intensity was analyzed using the BD FACS Diva software program (BD Biosciences).

Magnetic activated cell sorting (MACS)

The CD24-positive/CD24-negative Caov-3 cells were obtained using the MACS kit according to the manufacturer's instructions (Miltenyi Biotec, Nordrhein-Westfalen, Germany). Briefly, total populations of adherent cells were enzymatically dissociated into a single cell suspension and counted to confirm the quantity of the whole cells. The cells were incubated with 40 µl buffer per 107 total cells, which were directly conjugated to 10 µl of CD24 Biotin at 4°C for 15 min. After washing, the cells were incubated with 80 µl buffer directly conjugated to 20 µl of Anti-Biotin MicroBeads at 4°C for 15 min. Subsequently, the suspended cells were added to a MACS column that was placed in the magnetic field of a MACS separator (Miltenyi Biotech). The labeled CD24-positive cells were retained on the column and the unlabeled cells were eluted; when the column was removed from the magnetic field, the magnetically retained CD24-positive cells were collected as positively selected cells for further investigation. We also confirmed the expression of CD24 mRNA in CD24-positive and CD24-negative Caov-3 cells using conventional PCR. All experiments were performed three times.

The soft agar assay

The Caov-3 cells sorted CD24-positive/CD24-negative by MACS. A 6-well culture plate was coated with 2 ml of bottom agar mixture (DMEM/F12 with 10% FBS, 0.6% agar). After the bottom layer solidified, 2 ml top agar-medium mixture (DMEM/F12 with 10% FBS, 0.3% agar) containing 2×104 CD24-positive or CD24-negative cells was added, and the plate was incubated at 37°C for 4 weeks. The medium was replenished every 3 days. The plate was stained with 0.5% crystal violet, and the colonies with a diameter of >40 µm were counted.

The cell proliferation assay

CD24-positive/CD24-negative Caov-3 cells were obtained using a MACS kit. CD24-positive and CD24-negative cells (1×105 cells/well) were seeded in triplicate in a 6-well plate with serum-free DMEM to achieve cellular starvation. After 24 h, the medium was replaced with 100 ml complete growth medium and incubated for 1 more week. The cell numbers were counted manually with an optical microscope after 1, 3, 5 and 7 days. This experiment was carried out in quadruplicate.

The cell invasion assays

Cell invasion was assessed using 24-well-Transwell chambers (Corning, NY, USA) with a pore size of 8 µm (Kurabo, Osaka, Japan). Transwell upper chambers were pre-coated with 5 µg/chamber of solubilized basement membrane (Matrigel; BD Biosciences). Forty percent fibronectin was added to the bottom chamber. Ovarian cancer cells (2.5×105 per chamber) were seeded in the serum-containing media in the upper well of the Transwell chambers and incubated for 16 h at 37°C in 5% CO2. The non-invasive cells in the lower chamber (below the filter surface) were fixed in 100% methanol after 3.7% formaldehyde, stained with Giemsa staining solution, and counted under a microscope (×40 magnification). Cells from at least five fields were counted per chamber.

The chemotherapeutic sensitivity assay

Cells cultured in 10% growth medium were seeded in 96-well plates at a density of 2×104 cells per well. After 16 h of starvation, the cells were incubated for 48 h before being exposed to 0, 10, 50 or 100 µM cisplatin, which was purchased from Sigma-Aldrich. Then, CellTiter 96 AQueous (MTS) One Solution reagent (Promega, Tokyo, Japan) was added to each well, and the absorbance was recorded at 490 nm using a Corona SH-1000 lab absorbance microplate reader (Corona Electric Co. Inc., Ibaraki, Japan). The cell numbers were then calculated using a standard curve of the correlation of absorbance with the cell number counted under a microscope. All of the experiments were carried out in quadruplicate, and cell viability was expressed as the ratio of the number of viable cells with cisplatin treatment to that of untreated cells.

Western blotting

The cells were serum-starved and stimulated with PBS or 10 ng/ml TGF-β for 10 min. The cells were then washed three times in ice-cold PBS and lysed using Pierce RIPA buffer (Thermo Fisher Scientific) and the cytoplasmic and nuclear fractions were separated using a Nuclear Extract kit (Active Motif Carlsbad, CA, USA). To detect all of the proteins, equal amounts of cytoplasmic proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Blocking was performed in 10% bovine serum albumin in 1X Tris-buffered saline. Western blotting was performed with the primary antibody of CD24 (1:100 dilution), phosphorylated Akt (1:1,000 dilution), Akt (1:1,000 dilution), phosphorylated ERK (1:1,000 dilution), ERK (1:1,000 dilution), phosphorylated NF-κB (1:1,000 dilution), and NF-κB (1:1,000 dilution) overnight at 4°C. Finally, the bands were visualized using enhanced chemiluminescence (ECL Plus, GE Healthcare Life Sciences, Pittsburgh, PA, USA). All of the western blots were checked for equal protein loading using Ponceau staining.

The intraperitoneal xenograft model

Female 5-week-old athymic nude mice (BALB/c Slc-nu/nu) were purchased from Japan SLC and maintained in accordance with the institutional guidelines of Osaka Medical College. All of the animal studies were carried out according to approved experimental protocols. To assess tumorigenicity, 80×104 purified CD24-positive and CD24-negative cells were suspended in 1 ml PBS and were then intraperitoneally injected into each of the 10 female nude mice (5-week-old). The body weight and abdominal circumference were measured weekly. On the 50th day after the initiation of treatment, all of the mice were sacrificed, and the presence of macroscopic disease and the ascites volumes and organ weight were examined. Dissemination tumors were resected, fixed in formalin and paraffin sections were made for hematoxylin and eosin staining.

Statistical analysis

The statistical analyses in this study were performed using the JMP software program (SAS Institute, Japan). Fisher's exact probably test was used to evaluate the correlations between the immunohistochemical and clinical data. The endpoints were progression-free survival (defined as the time from the first day of chemotherapy until disease progression, based on the findings of imaging studies) and overall survival (defined as the time from the first day of chemotherapy to death from any cause). The univariate analyses of the histology, progression-free survival and overall survival were determined according to the Kaplan-Meier method using the log-rank test and the Cox proportional hazards model, respectively. Differences with p-values of <0.05 were considered to indicate statistical significance.

Results

The expression of CD24 in ovarian cancer patients

The data that we investigated included the patients' age, histology, International Federation of Gynecology and Obstetrics (FIGO) stage, peritoneal cytology, the presence of lymph node metastasis, the presence of peritoneal metastasis and the outcome, as shown in Table I. The 174 epithelial ovarian cancers included serous adenocarcinomas (n=48), mucinous adenocarcinomas (n=21), endometrioid adenocarcinomas (n=32), clear cell adenocarcinomas (n=32), serous surface papillary carcinomas (SSPCs; n=26) and others (n=15). The disease stages of the 174 investigated patients were classified as follows: stage I, n=56; stage II, n=14; stage III, n=85 and stage IV, n=19. Representative examples of immunohistochemical staining with anti-CD24 antibody are shown in Fig. 1. CD24 was expressed in a significantly higher percentage of patients with serous adenocarcinoma (87.5%) and SSPCs (84.6%) than among patients with other histological subtypes (p<0.01). In the analysis of the FIGO stage, CD24 was expressed in 53.6, 78.6, 77.7 and 79.0% of the patients with stage I, II, III and IV disease, respectively. CD24 was expressed in 70.8% of the patients with positive peritoneal cytology and 68.1% of the patients with negative peritoneal cytology. The expression of CD24 was observed in 78.7% of the patients with peritoneal metastasis, in comparison to 60.0% of the patients without peritoneal metastasis. The expression of CD24 was observed in 87.9% of the patients with lymph node metastasis, in comparison to 60.7% of the patients without lymph node metastasis. Significant associations were observed between the CD24 expression rate and the FIGO stage (p<0.01), and the presence of peritoneal (p<0.01) and lymph node metastasis (p<0.01).

Figure 1. Sequential sections immunohistochemically stained with CD24 in ovarian carcinoma. A representative example of CD24-negative stage 2C serous adenocarcinoma (A), and CD24-positive stage 3C serous adenocarcinoma (B). (A and B) ×40 original magnification. The scale bars represent 100 µm.

Table I. The immunohistochemistry results.

Table I.

The immunohistochemistry results.

	CD24
Variables	Negative	Positive	P-value
Age, years			0.18
<50	18 (37.5)	30 (62.5)
≥50	34 (27.0)	92 (73.0)
Histology			<0.01
Serous adenocarcinoma	6 (12.5)	42 (87.5)
Mucinous adenocarcinoma	13 (61.9)	8 (38.1)
Clear cell adenocarcinoma	12 (37.5)	20 (62.5)
Endometrioid adenocarcinoma	13 (40.6)	19 (59.4)
SSPCs	4 (15.4)	22 (84.6)
Others	4 (26.6)	11 (73.4)
Stage			<0.001
I	26 (46.4)	30 (53.6)
II	3 (21.4)	11 (78.6)
III	19 (22.3)	66 (77.7)
IV	4 (21.0)	15 (79.0)
Peritoneal cytology			0.75
Positive	38 (29.2)	92 (70.8)
Negative	14 (31.8)	30 (68.1)
Dissemination			<0.01
Positive	20 (21.3)	74 (78.7)
Negative	32 (40.0)	48 (60.0)
Lymph node metastasis			<0.01
Positive	4 (12.1)	29 (87.9)
Negative	35 (39.3)	54 (60.7)
Recurrence			<0.01
Positive	18 (19.1)	76 (80.9)
Negative	34 (42.5)	46 (57.5)
Died or survived			0.02
Died	11 (18.6)	48 (81.4)
Survived	41 (35.7)	74 (64.3)

The prognostic impact of the CD24 expression status in patients with ovarian carcinoma

The CD24 expression status was investigated in relation to patient survival. All patients underwent a primary staging operation such as total abdominal hysterectomy, bilateral salpingo-oophorectomy, omentectomy, and/or lymphadenectomy. Furthermore, all patients received 6-course conventional TC (paclitaxel and carboplatin) therapies as the adjuvant chemotherapy. The median follow-up duration was 29.2 months (range, 2–122 months). During the follow-up period, 94 of 174 patients (54%) had recurrent disease (18 from the CD24-negative group and 76 from the CD24-positive group), and 59 of 174 patients (33.9%) died (11 from the CD24-negative group and 48 from the CD24-positive group), as shown in Table I. As shown in Fig. 2, progression-free and overall survival were stratified according to the CD24 expression status using the Kaplan-Meier method with the log-rank test. The CD24 expression status was found to be significantly associated with both the progression-free survival (p<0.01) and overall survival (p<0.05).

Figure 2. The survival curves generated using the Kaplan-Meier method in 174 ovarian cancer patients. The progression-free survival (A) and overall survival (B) of the patients were stratified according to the expression of CD24. P-values were calculated using the log-rank test.

TGF-β stimulates EMT and CD24 expression

To study the EMT mechanism, we cultured cisplatin-resistant ovarian cancer cell lines (Caov-3 cells) in medium containing TGF-β. TGF-β is a major inducer of EMT during embryonic development, the pathogenesis of fibrotic disorders and cancer progression (26–28). It was first described as an inducer of the EMT in normal mammary epithelial cells (29). We examined whether there was a correlation between the EMT mechanism induced by TGF-β and the expression of CD24 in Caov-3 cells. The data from a quantitative PCR indicated that the expression of Snail was increased with TGF-β in the cultured medium, and that the expression of E-cadherin was decreased with TGF-β in the cultured medium, as shown in Fig. 3A. The expression of CD24 was also increased with TGF-β in the cultured medium. These findings suggest that TGF-β stimulates not only the EMT phenotypes but also the expression of CD24 in Caov-3 cells.

Figure 3. The correlation between the EMT mechanism induced by TGF-β and 1% hypoxia. (A) TGF-β enhanced the expression of Snail or CD24, and reduced the expression of E-cadherin in Caov-3 cells. (B) A condition of 1% hypoxia for 12 h enhanced the expression of Snail or CD24, and reduced the expression of E-cadherin in Caov-3 cells. **p<0.01.

Hypoxia stimulates EMT and CD24 expression

A hypoxic microenvironment, which is common to cancer cells emerges as an important factor in the induction of a pathological EMT, which is a key link in cancer progression (30). We examined whether there is a correlation between the EMT mechanism induced by hypoxia and the expression of CD24 in Caov-3 cells. We cultured Caov-3 cells under the condition of 1% hypoxia for 12 h. As shown in Fig. 3B, the data from quantitative PCR indicated that the expression of Snail was increased, and that the expression of E-cadherin was decreased under 1% hypoxia in comparison to normoxia. The expression of CD24 was also increased under 1% hypoxia (Fig. 3B). Next, we performed flow cytometry to examine whether the CD24 fraction of the Caov-3 cells was changed under the condition of 1% hypoxia. The CD24 fraction under normoxia and 1% hypoxia was 52.9 and 72.9%, respectively (data not shown). These findings suggest that hypoxic conditions stimulate not only the EMT phenotypes but also the expression of CD24 in Caov-3 cells.

CD24 induces EMT in ovarian cancer cells

We examined whether cells with the overexpression of CD24 induced the EMT mechanism in Caov-3 cells. As shown in Fig. 4A, the data from a quantitative PCR indicated that the expression of Snail was increased, and that the expression of E-cadherin was decreased in the CD24-overexpressing cells in comparison to control cells. We confirmed that CD24 directly induced the EMT phenotypes in Caov-3 cells and that the CD24-positive Caov-3 cells showed the acquisition of spindle cell morphology, as shown in Fig. 4B.

Figure 4. CD24 induces the EMT mechanism in ovarian cancer cells. (A) CD24-overexpressing cells enhanced the expression of Snail, and reduced the expression of E-cadherin. **p<0.01 (B) The CD24-positive Caov-3 cells showed the acquisition of spindle cell morphology compared with the CD24-negative Caov-3 cells.

CD24-positive cells have enhanced proliferation, a highly invasive phenotype, colony formation and are associated with cisplatin resistance in ovarian cancer cells

Several groups have reported that the EMT contributes to the properties of cancer stem cells (31,32). We examined whether CD24 could increase the colony-forming activity of CD24-positive Caov-3 cells. As shown in Fig. 5, after 28 days of culture, CD24-positive Caov-3 cells formed more colonies than CD-negative Caov-3 cells.

Figure 5. CD24-positive cells show enhanced colony formation. After 28 days of culture, CD24-positive Caov-3 cells formed more colonies than the CD24-negative Caov-3 cells.

To obtain further insight into the biological relevance of CD24-positive cells, we analyzed their effect on the proliferative activity of Caov-3 cells by manual cell counting. As shown in Fig. 6A, CD24-positive Caov-3 cells showed an enhanced proliferation rate after 7 days of culture in comparison to CD24-negative Caov-3 cells. We next examined the correlation between the expression of CD24 and the invasiveness of ovarian cancer. CD24-overexpressing cells showed 7.5-fold greater invasion in comparison to the control cells (p<0.01). Moreover, CD24 siRNA-transfected cells significantly inhibited the invasive activity of ovarian cancer (Fig. 6B).

Figure 6. CD24-positive cells show enhanced proliferation and a highly invasive phenotype in Caov-3 cells. (A) After 7 days of culture, CD24-positive Caov-3 cells showed an enhanced proliferation rate in comparison to CD24-negative Caov-3 cells. (B) The correlation between the expression of CD24 and the invasiveness of ovarian cancer as determined by an invasion assay. *p<0.05; **p<0.01.

To clarify whether the CD24-positive cells were more resistant to cisplatin than CD24-negative cells, we divided the Caov-3 cells into CD24-positive and CD24-negative fractions using MACS and performed an MTS assay with various concentrations of cisplatin. As shown in Fig. 7A, the CD24-positive Caov-3 cells were more resistant to cisplatin treatment than the CD24-negative Caov-3 cells. We confirmed that the CD24-overexpressing cells were also more resistant to cisplatin treatment than the CD24 siRNA-transfected cells (Fig. 7B).

Figure 7. CD24-positive cells associated with cisplatin resistance in Caov-3 cells. (A) CD24-positive Caov-3 cells were found to be more resistant to cisplatin treatment than CD24-negative cells by an MTS assay. (B) CD24-overexpressing cells were also found to be more resistant to cisplatin treatment than the CD24 siRNA-transfected cells by the MTS assay. *p<0.05; **p<0.01.

The long-term differentiation potential of CD24-positive cells

During long-term culture, in addition to increasing their cell mass, CSCs also undergo asymmetric division to generate a cell population with heterogeneous phenotypes and low tumorigenicity (33). We divided the Caov-3 cells into CD24-positive and CD24-negative fractions using MACS, and examined whether the CD24-positive and negative fractions of Caov-3 cells changed in the fraction after 7 days of culture by flow cytometry, as shown in Fig. 8. After 7 days of culture, 66.5% of the CD24-positive fractions cells were CD24-positive and 33.5% were CD24-negative. Interestingly, after 7 days of culture, 87.4% of the CD24-negative fractions cells were CD24-negative, while only 12.6% were CD24-positive. These results suggest that CD24-positive cells show self-renewal potential and cell differentiation potential after prolonged culture.

Figure 8. The long-term differentiation potential of CD24-positive cells. The CD24-positive or negative fractions of Caov-3 cells were evaluated in the fraction after 7 days of culture by flow cytometry. After 7 days of culture, 66.5% of the CD24-positive fractions cells were CD24-positive and 33.5% were CD24-negative; 87.4% of the CD24-negative fractions cells were CD24-negative.

CD24 differentially promotes the activation of the Akt and ERK pathway via the TGF-β signaling cascade in ovarian cancer cell lines

TGF-β is major inducer of the EMT during embryonic development, the pathogenesis of fibrotic disorders and cancer progression (26–28). It was first described as an inducer of the EMT in normal mammary epithelial cells (29). As the TGF-β signaling pathway is located upstream of both the PI3k/Akt and MAPK pathways, we anticipated that there would be a connection between CD24 and the TGF-β signaling pathway. To determine whether CD24 has any effect on the TGF-β signaling pathway, the Akt phosphorylation status and the ERK phosphorylation status were assessed by western blotting. As shown in Fig. 9, both Akt and ERK were significantly more phosphorylated in Caov-3 cells that were treated with TGF-β than in Caov-3 cells that were not treated with TGF-β. The phosphorylation of both Akt and ERK was significantly enhanced in CD24-overexpressing Caov-3 cells in comparison to the scrambled transfected Caov-3 cells. Moreover, we examined whether CD24 had any effect on the NF-κB phosphorylation status. NF-κB was significantly more phosphorylated in Caov-3 cells that were treated with TGF-β, and was significantly enhanced in CD24-overexpressing Caov-3 cells in comparison to control cells (data not shown). These data suggest that CD24 somehow enhances the TGF-β signaling pathway and that the PI3K/Akt and MAPK pathways may be promoted via this enhanced TGF-β signaling.

Figure 9. Western blotting reveals that CD24 enhances the TGF-β induced phosphorylation of Akt and ERK in Caov-3 cells. Caov-3 cells were treated in according to the procedure described in Materials and methods. After treatment with or without 10 ng/ml of TGF-β for 10 min, cells lysates were subjected to SDS-page, and then western blotting was performed with an anti-phospho-Akt antibody or an anti-Akt antibody (A) and an anti-phospho-ERK antibody or an anti-ERK antibody (B). The values represent the means ± SE from at least three independent experiments. *p<0.05; **p<0.01.

CD24 enhances the intra-abdominal disseminated tumorigenesis of ovarian cancer in vivo

Peritoneal dissemination is the primary route of progression in human ovarian cancer, and the amount of ascites and the disseminated tumor burden are correlated with the prognosis of human patients (34). We therefore examined the correlation between the expression of CD24 and the intra-abdominally disseminated tumorigenesis of ovarian cancer. We divided the Caov-3 cells into CD24-positive and CD24-negative fractions using MACS, and athymic nude mice were intraperitoneally inoculated with CD24-positive Caov-3 cells and CD24-negative cells, as described in Materials and methods. The appearance of the mice is shown in Fig. 10A. After 50 days, intra-abdominal dissemination was clearly detected in the athymic nude mice that were intraperitoneally inoculated with CD24-positive cells in comparison to CD24-negative cells. After performing a histological examination, the abdominal tumors were found to be papillary adenocarcinomas, which was consistent with the characteristics of the Caov-3 cells. The mean abdominal circumference after 50 days in the athymic nude mice that were intraperitoneally inoculated with CD24-positive cells was significantly higher than that observed in the mice with CD24-negative cells (Fig. 10B). The mean ascites volume after 50 days in the athymic nude mice that were intraperitoneally inoculated with CD-24 positive cells was also significantly higher than that observed in the mice with CD24-negative cells (Fig. 10B). The disseminated tumor weight after 50 days in the athymic nude mice that were intraperitoneally inoculated with CD24-positive cells was significantly higher than that in mice that were inoculated wit CD24-negative cells (Fig. 10B).

Figure 10. CD24-negative Caov-3 cells or CD24-positive Caov-3 cells were intra-abdominally injected into athymic nude mice. The body weight, abdominal circumference, and organ dissemination were determined at 50 days after the cell injection. (A) The physical appearance of the representative mice, and magnified views of the intra-abdominal dissemination pattern of the same mouse. (B) The body weight, abdominal circumference, and the total weight of the organs with dissemination and distant metastasis are shown for each group. *p<0.05.

Discussion

We found that CD24 was expressed in 70.1% of ovarian cancers and that the expression of CD24 was an independent predictor of survival in patients with ovarian cancer. This finding is in line with previous reports (35,36). Moreover, we demonstrated that the expression of CD24 was correlated with the FIGO stage (p<0.01), and the presence of peritoneal (p<0.01) and lymph node metastasis (p<0.01). In addition, CD24 induced the EMT phenomenon, which was involved in cell invasion, the highly proliferative phenotype, resistance to chemotherapy, and the properties of cancer stem-like cells (CSCs). Moreover, we clarified, for the first time, that the CD24-induced proliferation and invasiveness in ovarian cancer were dependent on the activation of PI3K/Akt, NF-κB and ERK.

EMT, which is associated with a key step in tumor metastasis via the induction of a highly invasive phenotype, has been intensely studied (4,37). We previously showed that the expression of EMT-related proteins was correlated with the status of tumor metastasis or the prognostic value in ovarian carcinoma (38). Given that 33 of 174 patients (19.0%) had EMT-positive status, as represented by both a reduced E-cadherin expression and the presence of nuclear Snail expression [4 (12.1%) in the CD24-negative group and 29 (97.9%) in the CD24-positive group], we concluded that a strong relationship existed between the expression of CD24 and EMT status (p<0.05) (data not shown). In this study, we demonstrated that CD24-positive ovarian cancer cells consistently acquired an EMT phenotype. These data indicate that CD24-positive cells are involved in the EMT process in ovarian cancer. In the present study, we also demonstrated that CD24-positive ovarian carcinomas have greater potential for intra-abdominal tumor cell dissemination and ascites production than CD24-negative ovarian carcinomas in in vivo models.

EMT is a clinically important multifaceted pathological program that endows cancer cells with the ability of invasion, resistance to apoptosis and dissemination. EMT-inducing transcription factors such as Snail, Twist and Slug are aberrantly expressed in human cancers and are responsible for a poor prognosis in patients (5). In the present study, we demonstrated that CD24 enhanced the expression of Snail, and decreased the expression of E-cadherin in the Caov-3 cells. We herein demonstrated, for the first time, that CD24 amplified cell growth-related intracellular signaling via the PI3K/Akt and MAPK pathways by affecting the EMT signal pathways. Some groups have reported that EMT also contributes to the properties of CSCs (31,32,39), and CSCs have been found to transiently acquire stem cell-like properties as a consequence of EMT (40). The CSC hypothesis states that solid tumors are hierarchically organized and sustained by a minority of the tumor cell population with stem cell-like properties, such as self-renewal, multilineage differentiation and tumorigenicity. CD24 is a mucin-like cell surface glycoprotein that is frequently overexpressed in various human cancers and is correlated with a poor prognosis (35,41). Recent reports have revealed that CD24-positive cells function as CSCs in colon cancer, hepatocellular carcinoma, pancreatic carcinomas, and ovarian carcinomas (23,24,41,42). In the present study, we also demonstrated that CD24-positive cells show self-renewal potential and cell differentiation potential after prolonged culture, and increase the colony-forming activity, invasive activity, proliferation activity, and resistance to cisplatin in comparison to CD24-negative cells. We also demonstrated that CD24 induced the EMT phenomenon, which was involved in cell invasion, resistance to chemotherapy, and the formation of side populations of CSCs (7); however, we could not clarify what regulated the expression of CD24, and CD24-positive cells were identical with ovarian cancer stem cells in that study. In the present study, we indicated that CD24 plays a critical role in regulating the EMT phenomenon in ovarian cancer. Our study provides novel insight into the interaction between CSCs and the EMT program and a better understanding of the mechanism underlying the involvement of CD24 in ovarian cancer metastasis. Given the findings in the present study, we believe that CD24 is a key molecule of metastatic progression in the EMT phenomenon and a promising therapeutic target for advanced ovarian cancer.

Acknowledgements

This study was supported by a JSPS KAKENHI grant no. 25462621 (to Y. Terai).

Glossary

Abbreviations

Abbreviations:

References