Introduction
Ovarian cancer is the prime cause of death from gynecological malignancies and the fifth leading cause of cancer-related deaths among women. Over 85% of ovarian malignancies are categorized as epithelial ovarian cancer (EOC). Relapse and drug resistance are major barriers in the management of patients with EOC [1]. The reasons for cancer recurrence and lack of response to chemotherapy are unknown. Recent studies have indicated that cancer stem cells or tumor-initiating cells (CSCs/TICs) are not only the potential origin of the tumor, but also the source of EOC relapse and chemo-resistance [2,3].
Based on extensive clinical-pathologic analysis and molecular genetic studies, a dualistic model of ovarian tumorigenesis has been proposed in which ovarian cancers can be divided into two subgroups: Type I tumors, which are low-grade and histologically heterogeneous, and Type II tumors, which are high-grade and mostly serous [4,5]. Type II tumors, of which the most common and lethal is high grade serous ovarian carcinoma (HGSC), represent 75% of all ovarian carcinomas and are responsible for 90% of ovarian cancer deaths [4,5]. Recently, a stem-like subtype of ovarian cancer was reported. This subtype had worse disease-free and overall survival in high grade, malignant ovarian tumors, and was associated with poorer disease-free survival in high-grade serous cancer [6].
The isolation of putative, tumorigenic CSCs has been achieved by using various ovarian cancer cell lines, patients' ascites, and primary, human ovarian tumor specimens following in vitro or in vivo passage [7,8,9]. However, patients' ascites do not recapitulate all aspects of primary tumors and cell lines may not reflect unmanipulated, primary cells. To date, there have been no reports on isolation of an ovarian cancer stem cell (OCSC) population directly from the main HGSC bulk.
In this study, we isolated cancer cells from six HGSC specimens and defined the characteristics of these cell clones and identified the potential OCSCs. This is the first report on the identification of OCSCs from human HGSC, and may provide new therapeutic approaches to treat this highly aggressive tumor.
Materials and Methods
Tissue collection, isolation, and culture of tumor stem cells
Six samples of tumor tissue specimens (Table 1) were obtained at the time of primary surgery from consenting patients with ovarian carcinoma, in accordance with the Scientific and Ethical Committee of the Shanghai First Maternity and Infant Hospital affiliated with Tongji University. We have obtained written informed consent from all study participants. All of the procedures were done in accordance with the Declaration of Helsinki and relevant policies in China. One aliquot of each sample was reviewed by two experienced pathologists to verify the histologic assessment, and a second aliquot was used for isolation, purification, and culture of tumor stem cells. Within 30 min of surgery, tumor samples were mechanically dissociated and enzymatically digested in a 1:1 solution of Type III collagenase/hyaluronidase (0.1%; Sigma, USA) for 30 min at 37°C and incubated at 37°C for 2 h in a shaking bath. Sterile gauzes (pore diameter sizes: 200 mesh) were used to remove clumps and erythrolysis was performed in hypotonic solution (0.2% NaCl), followed by 1.2% NaCl to stop lysis. Cells were washed with D-Hanks solution and passed through a 40-mm filter. Single stem cells were placed at a density of 103 cells/mL in stem cell medium, consisting of serum-free medium DMEM/F12 (Invitrogen, USA) supplemented with 10 ng/mL basic fibroblast growth factor (bFGF; PeproTech, USA), 20 ng/mL epidermal growth factor (EGF; PeproTech, USA), 5 μg/mL insulin (Sigma, USA), and 0.4% bovine serum albumin (BSA; Sigma, USA). Cells were cultured in 6-well ultra-low attachment plates (Corning,USA) as described by Ponti et al. [10].
Table 1
Case description and tumor sphere formation.Characteristics of primary high-grade serous ovarian carcinomas (HGSC) and stem-like properties in vitro and in vivo. Cells from freshly dissociated HGSC tissues (n = 6) were examined for sphere and tumor formation., The ability of samples to generate cancer spheres in vitro was evaluated by prolonged culture in growth factor-containing serum-free medium. NOD/SCID mice were injected with sphere-forming cells to assess the potential to form tumors in vivo. Cells from all six primary tumors generated xenografts and could be passaged in mice
Sphere formation and differentiation
Under stem cell conditions, spheroid colonies were apparent after one week of culture. To assess the self-renewal ability of the cells, sphere cells were mechanically disaggregated and single cells were plated in 96-well ultra-low attachment plates in the stem cell medium described above. The medium was changed twice a week to renew growth factors. Single, dissociated primary sphere cells give rise to secondary spheres that, in turn, formed tertiary spheres. The sphere forming ability of the cells was photographed using a phase contrast microscope (Axiovert 100, Zeiss, Germany). Cellular aggregates with a diameter larger than 50 μm were classified as ‘spheres'.
To evaluate potency of differentiation of tumor sphere cells, cells in the spheres were picked out with a pipette and grown on collagen-coated dishes with DMEM containing 10% fetal bovine serum (FBS) without growth factors. Cell morphology was inspected every day using phase-contrast microscopy. After 14 d of culture in differentiation conditions, the cells were labeled using monoclonal antibodies against cytokeratin-7 (CK7; 1:500; Santa Cruz Biotechnology, CA) or cancer antigen-125 (CA125; 1:500; Santa Cruz Biotechnology) for 2 h at 37°C, then washed in blocking solution and incubated with fluorescein isothiocyanate (FITC)-conjugated rat anti-mouse IgG (1:300; Invitrogen, USA) for 1 h at room temperature. Cells were then washed in blocking solution and incubated with 4, 6-diamidino-2-phenylindole (DAPI; 1:1,000; Sigma) for 1 min, followed by three washes with phosphate buffered saline (PBS). Stained cells were visualized with a fluorescent microscope (Olympus, Japan).
Analysis of stem cell gene expression by reverse transcription (RT)-PCR
Total RNA was isolated from cells and tumor tissue using TRIzol reagent (Invitrogen, USA). The quality of the total RNA was evaluated by optical density (A260/280 ratio of >1.9). RT-PCR was performed using an RNA PCR kit (Invitrogen, USA) according to the manufacturer's protocol. β-Actin was amplified as an internal control. RT-PCR products were analyzed on 3% agarose gels stained with ethidium bromide and photographed under UV illumination. Primer sequences, lengths of predicted amplified gene fragments, and annealing conditions are listed in Table 2.
Immunofluorescence analysis
Sphere cells were washed twice with PBS and fixed in 4% paraformaldehyde in PBS for 20 min. After washing with PBS, the cells were blocked with 5% BSA and 0.05% Triton X-100. Spheroids were stained with primary antibodies against Oct-4 (1:2000, Santa Cruz Biotechnology), Nanog (1:2000; Chemicon Millipore,USA), and Sox2 (1:1000; R&D Systems, Minneapolis, MN) overnight at 4°C. Cells were then washed in blocking solution and incubated with phycoerythrin (PE)-conjugated goat anti-rabbit IgG (1:250; Invitrogen, USA) for 1 h at room temperature. After washing, the nuclei were counterstained with DAPI for 1 min. Fluorescence microscopy was performed with a Leica TCS-SP5 laser-scanning confocal microscope (Leica, Germany).
Flow cytometric analysis
Both single cells from tumor spheres and adherent cells were blocked with 10% goat serum, and then the cells were incubated with mouse monoclonal antibody anti-CD117 (1:500; Invitrogen, USA), mouse monoclonal antibody anti-CD24 antibody (1:400; R&D Systems), mouse monoclonal antibody anti-ABCG2 antibody (1:500; Abcam), mouse monoclonal antibody anti-CD44 (1:200; Neomarker,USA), mouse monoclonal antibody anti-ALDH1 (1:500; BD Biosciences),USA or isotype control mAb for 45 min at 4°C. After resuspending in 1% BSA/PBS, cells were incubated with FITC-labeled goat anti-rabbit or mouse IgG (Molecular Probes,USA) (1:300; Chemicon Millipore) for 1 h at 4°C. For CD133 detection, we incubated cells with anti-CD133/2 (phycoerythrin (PE)-conjugated; Miltenyi Biotec) at 4°C for 20 minutes. After washing, labeled cells were analyzed using a FACScan flow cytometer (BD Biosciences). Data were analyzed using Cell Quest software (BD Biosciences).
Histology and immunohistochemical staining
Tumor tissues were immediately immersed in 10% neutralized formalin for 24 h, then embedded in paraffin and processed by standard histological methods. Serial, 5-µ-thick coronal sections were cut from paraffin blocks, attached to slides with poly-L-lysine, and sections in each series were stained with hematoxylin and eosin (HE) for histologic examination. HE-stained histologic slides from each patient were reviewed and corresponding paraffin embedded tissue blocks were selected by one observer of the Department of Pathology. From each paraffin block selected, 4-micron serial sections were cut. Immunohistochemical studies were carried out with the GTVision™ III Detection System (Including DAB)/Mo&RKit (Gene Tech Company Limited, Shanghai, China) according to the manufacturer's protocol. Each specimen was then incubated with the following primary antibodies (Gene Tech Company Limited): CK7 (1:100), CA125 (1:200), p53 (1:150), Wilms' tumor gene (WT1, 1:100), cytokeratin 20(CK-20, 1:200) and Villin (1:200). (all were purchased from Gene Tech Company Limited, Shanghai, China).Appropriate positive and negative controls were included with each run of immunostains. The staining for p53 was nuclear, whereas it was membranous and cytoplasmic for CK7 and CA125. For WT1, staining was predominantly nuclear with lower amounts in the cytoplasm. The diagnosis of HGSC was established using morphology and immunohistochemistry according to the two-tier grading system recently proposed by the MD Anderson Cancer Center [11]. Sections were examined microscopically using digital photometry.
Chemotherapy resistance assays
Rates of resistance to drugs were assessed using the MTT assay (Sigma, USA). Single cells were obtained from cancer sphere dissociation and plated in 96-well ultra-low attachment plates (Corning) at 3 × 103 cells per well in stem cell medium. In parallel, adherent cells from spheroid colonies were seeded at 3 × 103 cells per well in 96-well black-walled plates (Corning). After 48 h, both undifferentiated and adherent cells were treated for 72 h with 0.1-100 μmol/L cisplatin (Sigma, USA) or from 0.1 nmol/L to 10 μmol/L paclitaxel (Sigma; n = 5 per drug dose). At each time point (24 and 48 h), 20 μL of MTT solution (5 mg/mL) was added to each well and the plate was incubated for 4 h at 37°C. The medium was then replaced with 150-μL dimethylsulfoxide (DMSO, Sigma, USA). The absorbance values were determined at 570 nm with an ELISA reader (Biotek, USA). Blank control wells were used for adjusting background absorbance to zero. All assays were performed three times and repeated at least twice.
Xenograft tumor formation
All animal experiments were approved by and performed according to the guidelines set by the Animal Research Ethics Board at Shanghai First Maternity and Infant Hospital, Tongji University School of Medicine. Female non-obese diabetic (NOD)/severe combined immunodeficient (SCID) mice (4-5 weeks old) were obtained from (Shanghai SLAC Laboratory Animal Co., Ltd (Shanghai, China). Adherent cells and sphere cells were enzymatically dissociated to obtain single-cell suspensions, diluted in PBS, mixed with Matrigel, and injected subcutaneously into NOD/SCID mice (n = 6 per group) at the following serial dilutions: 103, 104, 105, and 106 cells. Mice were observed for changes in behavior, appearance, or weight, and inspected for tumor appearance by observation and palpation. Tumor growth was measured weekly using a caliper. Mice were killed by cervical dislocation and the tumors were collected. Tumor sections were stained with HE and examined by immunohistochemical analysis to study tumor histology, as described above. All specimens were evaluated independently by two pathologists to compare mouse xenografts with patient tumors.
Statistical analysis
A probability level of p < 0.05 was used throughout to determine statistical significance. Treatment groups were compared with the control group using one-way analysis of variance (ANOVA) and Dunnett's Multiple Comparison post-tests.
Results
A subpopulation of cells with self-renewing and sphere formation capacity in HGSC
We cultured cancer stem-like cells from six different HGSC primary tumors using a proliferative medium containing EGF, bFGF, and insulin to maintain cells in an undifferentiated state. After 1 week of culture, non-adherent spherical clusters of cells were observable (Fig. 1A, top left), and these continued to expand for 2-3 weeks in serum-free media. To investigate cell self-renewal, cultured spheres were dissociated into single cells and allowed to grow in the same stem cell-selective medium. Similar progeny spheres emerged after 5 d (Fig. 1A, top center). At day 10, spheroids began to take shape (Fig. 1A, top right). By day 14, spheroids had completely formed and became well-rounded structures composed of numerous, compacted cells (Fig. 1A, bottom left). The dissociated firstgeneration spheres were able to generate second-generation and third-generation spheres. The sphere cultures could be maintained for more than 30 passages (n = 3 independent experiments). These results demonstrated that tumor sphere cells had self-renewing characteristics.
Fig. 1
Human HGSC cells formed anchorage-independent, self-renewing spheres. (A). Phase-contrast images of cells in serum-free culture conditions for 7 d (top left). Phase-contrast images of a single sphere-derived cell cultured in a 96well ultra-low attachment plate under serum-free conditions. The propagation of a single cell was recorded at day 5 (top center), day 10 (top right), day 14 (bottom left and right). Scale bar, 200 µm. When spheres were transferred to differentiation conditions, cells migrated out of the core of the sphere and extended processes were noted (bottom center and right) (B). Immunocytochemistry showed that adherent cells expressed sensitive and specific markers for ovarian cancer (CA125 and CK7). Nuclei were counterstained with DAPI. (40×).
When spheres were cultivated under differentiation conditions, we observed that floating, undifferentiated cells had attached to the bottom of the culture plates after one day of culture. These cells had gradually migrated from tumor spheres and differentiated into large and adherent spheres resembling the parental HGSC cells (Fig. 1A, bottom center, and right). These adherent cells expressed high levels of the CA125 and CK7, sensitive and specific markers for ovarian cancer (Fig. 1B). Taken together, these findings provide evidence that a subpopulation of spheroids from HGSC self-renew under stem cell-selective conditions and, under differentiation conditions, assume an epithelial tumor phenotype.
Sphere cells preferentially expressed stem cell genes
Using RT-PCR, we examined the expression of specific markers known to be associated with stem and/or progenitor cells, including Nestin, Oct-4, Nanog, Sox-2, Bmi-1 and Nestin, all of which are genes that are essential for the maintenance of an undifferentiated state. We found that sphere cells expressed higher mRNA levels of Oct-4, Nanog, Sox-2, Bmi-1 and Nestin, suggesting that these cells had a stemness phenotype (Fig. 2). To better characterize the ovarian CSC population, we further analyzed the expression of several stem/progenitor cell surface markers. As shown in Fig. 2, expression of CD133, CD44, CD24, ALDH1, CD117, ABCG2, and CD24 markers was consistently and significantly higher in sphere cells than in the corresponding adherent cells or tumor cells.
Fig. 2
RT-PCR analyses of expression of stem cell genes in sphere-forming cells (SFCs) under stem cell-selective conditions, compared with parental bulk tumor population cells (OC) and SFCs under differentiation conditions (Different). Lanes 1-3 correspond to tumor T1 gene expression under the three conditions. Lanes 4-6 similarly denote tumor T2 sphere-forming cell gene expression under the same conditions. β-actin mRNA expression was used as an internal control.
Sphere cells express markers of pluripotent embryonic stem (ES) cells
Current evidence suggests that the expression of Oct-4, Nanog, and Sox2 is essential for the maintenance of pluripotency in ES cells and germ cells, and that these genes are down-regulated in all adherent somatic cell types. Expression of stemness genes is also used to characterize CSCs [7,9]. To examine the subcellular localization of Oct4, Nanog, and Sox2 in spheroid body-forming cells, immunofluorescent staining of these markers was performed.Oct4, Nanog, and Sox2 proteins were positively stained within the perinuclear region and cytoplasm of spheroid body-forming cells (Fig. 3). These results are consistent with the robust levels of Oct4, Nanog, and Sox2 mRNAs expressed in these cells (Fig. 2). Taken together, the results indicate that sphere cells possess an undifferentiated, ES cell phenotype.
Fig. 3
Expression of pluripotency markers in sphere-forming cells. Spheres are positive for Oct-4, Nanog, and Sox2. Confocal images were double-labeled with primary antibodies indirectly labeled with PE and the nuclear stain DAPI. Representative images are shown. Scale bars represent 50 µm.
Expression of putative stem cell markers in sphere cells and adherent cells
Different cell surface antigens have been identified in several cancers as candidate CSC biomarkers. CD44, CD24, CD117, CD133, ABCG2, and ALDH1, or a combination of these markers, have all been used to define CSCs in multiple human epithelial cancers including ovarian cancer. Therefore, to examine whether spheroid cells were enriched for CSCs, we analyzed the expression of multiple potential cancer stem cell markers in sphere cells by flow cytometry. FACS analysis demonstrated the presence of very high levels of cells that were positive for CSC markers in the spheroid cell population, whereas CSC markers were significantly downregulated in adherent cells (Fig. 4, Table 3). These results suggest that tumor spheres could enrich for cells expressing cell markers associated with certain stem/progenitor cell properties, which could be potential candidate cell surface markers for ovarian cancer stem-like cells.
Table 3
Expression of putative stem cell markers in sphere cells and adherent cells. mean ± SD; n = three independent experiments)
Fig. 4
Immunophenotypic analysis of HGSC sphere-forming cells and adherent cells. Sphereforming cells and adherent cells were stained with CD44, CD24, CD117, CD133, ABCG2, and ALDH1 primary antibodies and subjected to flow cytometry. Red lines correspond to sphere-forming cells. Green lines are isotype controls. One representative staining of three independent experiments is shown.
Sphere-forming cells displayed chemoresistance in vitro
CSCs in several solid tumors, including HGSC, are known to be more chemoresistant than the non-CSC population. To examine chemoresistance in sphere-forming cells, we determined the expression of multifunctional efflux transporters from the ABC gene family responsible for drug resistance of CSCs. Expression of ABCG2 was high in CSCs (Fig. 2, 3), suggesting that these cells may be chemoresistant. To examine whether the ovarian cancer stem-like cells displayed resistance to conventional chemotherapies, sphere cells and adherent cells were exposed to cisplatin or paclitaxel, basic chemotherapeutic agents for ovarian cancer treatment. Cell survival assays were conducted to determine the concentration of each drug that inhibited cell growth by 50% (IC50). The results showed greater resistance of sphere cells cultured under stem cell (undifferentiated) conditions compared to adherent cells (Fig. 5; p < 0.05). IC50 values for cisplatin were greater for sphere cells under stem cell conditions (18.84 μmol/L ) than those of adherent cells (7.35 μmol/L). Similarly results were obtained for paclitaxel, IC50 (IC50 = 9.73 and 2.56 μmol/L for sphere cells and adherent cells, respectively. These results indicate that cancer stem-like cells may contribute to drug resistance commonly found in HGSC.
Fig. 5
Sphere-forming cells showed increased resistance to conventional chemotherapeutic agents cisplatin and paclitaxel. Sphere-forming cells and adherent cells were treated with cisplatin (20 μmol/L; left) or paclitaxel (2 μmol/L; right) for 4 h. Cell survival was determined by MTT assay. Sphere-derived cells demonstrate significant resistance to drug treatment. *p < 0.05 compared with adherent cells. Data are presented as the mean ± S.D. of three independent experiments.
Sphere forming cells exhibited higher tumorigenicity in vivo
Tumor sphere cells obtained from six patients were injected into female NOD/SCID mice. All tumors generated xenografts and could be passaged in mice (Table 1). Tumorigenicity experiments with one of the tumors (T2) showed that as few as 1 × 104 sphere cells were able to form a tumor when subcutaneously injected into nude mice, while 2 × 106 adherent cells were needed to produce a similar tumor (Table 4). Moreover, spheroid body-forming cells generated subcutaneous tumors with larger volumes and within a shorter time frame compared to adherent cells (Fig. 6A, B and C). HE staining of tumors grown in mice after injection of ovarian cancer-initiating cells revealed the presence of a complex papillary pattern and slit-like spaces, and malignant cells with marked nuclear atypia and abundant mitotic activity (Fig. 6D). In addition, immunohistochemistry analysis showed that the xenografts expressed the HGSC differentiation markers CA125, CK7, p53, and WT1, but were negative for CK20 and Villin (Fig. 6D). HE staining and immunohistochemistry analysis revealed that the histological features and immunoprofile of xenograft tumors derived from spheroid body-forming cells and adherent cells were similar to the parental tumor. These data indicate that a stem cell population exists in HGSC following in vivo transplantation.
Table 4
Tumor formation of sphere cells and adherent cells. Tumor forming capacity of sphere-derived and adherent cells. Increasing numbers of cells (103, 104, 105, and 106) from tumor T2 were subcutaneously injected into nude mice. Tumor generation was evaluated 100 d after implantation
Fig. 6
(A). Tumor xenografts derived from 1 ×106 adherent cells (A) or sphere cells (B) injected in NOD/SCID mice. Sphere cell-derived tumors were substantially larger. (C) Xenografted tumor growth curve (D) Xenografts generated from both adherent and sphere cells resemble the original patient tumor (100×). Data are representative of three independent experiments.
Discussion
Currently, therapy for ovarian cancer is largely dependent on FIGO stage and grade rather than type; however, this is likely to change in the future with the development of new chemotherapeutic agents and targeted therapies for EOC [12]. Therefore, central pathology review becomes mandatory in ovarian carcinoma clinical trials when treatment relies on morphological subtype or any other pathological parameter. At present, based on morphologic, immunohistochemical, and molecular genetic studies, a dualistic model of ovarian tumorigenesis has been put forward in which ovarian cancers can be divided into two subgroups: type I and type II. Type II tumors presumably evolve rapidly and disseminate early in their clinical course. These tumors are highly aggressive and present in an advanced stage (stages II-IV) in more than 75% of cases [4,5]. Type II disease therefore represents most of the current public health burden for EOC. HGSC is the prototype of Type II tumors, and are chromosomally unstable and harbor TP53 mutations in more than 95% of cases [4,5]. A putative precursor lesion for this type of cancer remains unidentified. It is well known that tumors are composed of heterogeneous cell types. There is increasing evidence for the presence of a rare population of undifferentiated cells termed CSCs or TICs in malignant tumors. This subpopulation may play a pivotal role in tumor initiation, development, chemoresistance, and recurrence [2,3]. What has yet to be established is whether each HGSC cancer cell possesses the potential to initiate and sustain tumor growth, or whether the tumor is hierarchically organized so that only a subset of cells—CSCs—possess this potential.
The stem cell model of carcinogenesis suggests that cancers originate in tissue stem or progenitor cells, so many of the characteristics of normal adult stem cells have also been attributable to CSCs [13]. Reynolds et al. [14] first developed an in vitro technique termed the neurosphere assay to quantify activity of neural stem cells Similar assays have also been used in normal breast tissue, with the formation of mammospheres demonstrating the presence of a stem cell population This anchorage-independent, serum-free culture system yielded clonogenic stem-like cells that possessed many attributes common to normal stem cells [15]. Recent studies have used sphere cell culture techniques to isolate, enrich, maintain, or expand potential CSC subpopulations from various types of cancer [7,9,10]. Sphere formation is one of the most commonly used assays to identify and isolate CSCs [15]. In EOC, spherical colonies are found in patient ascites and solid ovarian tumors [7,9]. Using stem cell conditions and sphere formation assays, we collected non-adherent cells from six different patients to form spheroids. The ability of cells to generate spheres after serial passages demonstrates their self-renewal potential. Under differentiation conditions, floating cells could adhere and differentiate. Adherent ovarian cancer cells were phenotypically very similar to the major cancer cell population present in the original ovarian tumor. This demonstrates the existence of a precise hierarchy for the formation of HGSC, based on the generation of large numbers of cell progeny from a relatively small number of self-renewing undifferentiated cells.
CSCs can be identified and isolated by flow cytometry according to cell surface markers [16,17]. Different cell surface antigens have been identified in several cancers as candidate CSC biomarkers, although none of these are CSC-specific [17,18,19,20]. Here we examined the expression of surface markers that are highly expressed in sphere-forming cells. Potentially, the identification of specific CSC-associated markers from expression profiles of HGSCs could be correlated with patient survival, leading to improved strategies for therapy
Using RT-PCR analysis, we found that sphere-forming cells highly expressed stemness related genes compared with adherent cells or parental bulk tumor cells. Protein expression for some genes was confirmed by immunofluorescence analysis. Oct-4 is a POU family homeoprotein initially expressed in the inner cell mass of embryos and is essential for the maintenance of pluripotency. Nanog is a diver gent homeoprotein that is capable of maintaining self-renewal in ES cells. Over-expression of Nanog is associated with an increased self-renewal capacity in ES cells [21]. The transcription factor SOX2 is essential for maintaining the pluripotent phenotype in ES cells and is a partner of Oct-4 in regulating several ES cell-specific genes [22,23]. Together, our data suggest that spheres contain cells that possess pluripotent and self-renewal capacity. Thus, Oct-4, Nanog, and SOX2 may have important roles in the initiation and/or progression of HGSC and, consequently, may serve as important molecular diagnostics or therapeutic targets for the development of novel treatment strategies in HGSC patients.
Chemoresistance is one of the greatest clinical in the successful treatment of ovarian cancer. Approximately 80% of ovarian cancer patients that initially respond to standard chemotherapy later develop a recurrent, therapy-resistant, lethal disease [1]. The survival of a very small percentage of chemotherapy-resistant CSCs could facilitate the development of recurrent progressive disease [24]. We demonstrated that sphere cells showed higher resistance to the chemotherapeutic agents, cisplatin and paclitaxel. Although the mechanisms of this drug resistance remain to be elucidated, several studies have revealed the possible involvement of ATP-binding cassette (ABC) drug transporters [8]. In this study, spherical cells highly expressed ABCG2, a member of the ABC transporter transmembrane proteins, which may be involved in the efflux capacity of CSCs. Bmi-1, a member of the Polycomb group (PcG) family, participates in the self-renewal and maintenance of CSCs [25]. Silencing of Bmi-1 has been shown to sensitize ovarian cancer cells to cisplatin and induce apoptosis [26]. We demonstrated high expression levels of Bmi-1 in sphere-forming cells, suggesting that this gene may contribute to chemoresistance in these cells. In addition to the in vitro functional characterization of sphere cells, the in vivo transplantation into NOD/SCID mice was necessary to verify tumorigenicity [16,17].
Our results indicate that compared to adherent cells, sphere cells exhibited enhanced tumorigenic capacity, in that tumors were formed with the inoculation of fewer cells. When an equal number of cells were inoculated, the volume of tumors derived from spheroid body-forming cells was much larger and required a shorter period of time than that of the adherent cells. Notably, phenotypic analysis of tumors derived from sphere cells showed that they closely resembled that of the original tumor (Fig. 6; Tables 1, 3). These data demonstrate the presence of a hierarchy of cells within HGSC, in which only a fraction of the cells have the ability to generate a new tumor. Thus, tumorigenic HGSC cells from most tumors appear to exhibit the properties of CSCs.
In summary, we have demonstrated that HGSC contains a subpopulation of stem cell-like cells that form spheroids when cultured in serum-free media and possess self-renewal capacity, strong tumor-initiating ability, and higher resistance to chemotherapy. These data may lead to a considerable increase in our understanding of the biology of HGSC. Importantly, these results raise the possibility that HGSC is a stem cell malignancy, which could lead to the development of better diagnostic and therapeutic methodologies.
Acknowledgements
This work was supported in part by the National Natural Science Foundation of China (No. 81300466, Luo XZ), Shanghai Natural Science Foundation (No. 13ZR1404500, Luo XZ) and the health bureau of Shanghai (No. 2007Y48, He QZ).
Conflict of Interest
The authors have declared no conflict of interest.
Get Permission
PubMed ID
