Abstract
Epithelioid sarcoma (ES) is a relatively rare, highly malignant soft tissue sarcoma. The mainstay of treatment is resection or amputation. Currently other therapeutic options available for this disease are limited. Therefore, a novel therapeutic option needs to be developed. In the present study, we established a new human ES cell line (ESX) and analyzed the characteristics of its cancer stem-like cells/cancer-initiating cells (CSCs/CICs) based on ALDH1 activity. We demonstrated that a subpopulation of ESX cells with high ALDH1 activity (ALDHhigh cells) correlated with enhanced clonogenic ability, sphere-formation ability, and invasiveness in vitro and showed higher tumorigenicity in vivo. Next, using gene expression profiling, we identified CD109, a GPI-anchored protein upregulated in the ALDHhigh cells. CD109 mRNA was highly expressed in various sarcoma cell lines, but weakly expressed in normal adult tissues. CD109-positive cells in ESX predominantly formed spheres in culture, whereas siCD109 reduced ALDH1 expression and inhibited the cell proliferation in vitro. Subsequently, we evaluated the expression of CD109 protein in 80 clinical specimens of soft tissue sarcoma. We found a strong correlation between CD109 protein expression and the prognosis (P = 0.009). In conclusion, CD109 might be a CSC/CIC marker in epithelioid sarcoma. Moreover, CD109 is a promising prognostic biomarker and a molecular target of cancer therapy for sarcomas including ES.
Introduction
Epithelioid sarcoma (ES) is a relatively rare and highly malignant soft tissue sarcoma (STS) accounting for <1% of all STSs [1]. The mainstay of treatment is aggressive, radical local resection or amputation. Currently other therapeutic options available for ES are limited. Therefore, a novel therapeutic option needs to be developed.
Recent studies have revealed that several human cancers contain a small subpopulation of cells called cancer stem-like cells (CSCs)/cancer initiating cells (CICs), which are defined by the ability of self-renewal, multi-differentiation potential, and tumorigenesis. Therefore, CSCs/CICs are believed to be responsible for the progression and relapse of cancer [2]. In the current study, we isolated CSCs/CICs based on aldehyde dehydrogenase 1 (ALDH1) activity. Human ALDHs are a family of NAD (P)+-dependent enzymes involved in detoxifying a wide variety of aldehydes to their corresponding weak carboxylic acids [3]. They serve to detoxify both xenobiotic aldehydes (eg. cyclophosphamide) and many other intracellular aldehydes, including ethanol and vitamin A [4]. Therefore, ALDH activity is important for drug resistance and the response to oxidative stress [5]. Recently ALDH1 activity was used, either alone or in combination with cell surface markers, to identify CSCs/CICs in hematologic malignancies and carcinomas derived from the lung and prostate [6-8].
We established a new ES cell line (designated ESX) from a 73-year-old woman. Next, we investigated CICs/CSCs in ES cell lines and isolated CSCs/CICs based on ALDH activity. Finally, we demonstrate that CD109 is a potential CSC/CIC marker that may be useful as a prognostic biomarker and a molecular target of cancer therapy for sarcomas, including ES.
Materials and Methods
Ethics Statement
Mice were maintained and experimented on in accordance with the guidelines of and after approval by the Ethics Committee of Sapporo Medical University School of Medicine, Animal Experimentation Center under permit number 08-006. Any animal found unhealthy or sick was promptly euthanized. All studies were approved by the Institutional Review Board of Sapporo Medical University Hospital. Written informed consent was obtained from all patients according to the guidelines of the Declaration of Helsinki.
Primary tumor
A 73-year-old Japanese woman was admitted to our hospital with a 9-month history of swelling of the left thigh. The swelling had gradually enlarged and become painful. A well-demarcated elastic soft mass was palpable in the medial aspect of the left thigh. Magnetic resonance imaging revealed a subcutaneous tumor and lymph node metastases in the inguinal region (Figure S1A). The tumor (3×3 cm) was homogeneously isointense relative to skeletal muscle in T1-weighted images, whereas it was heterogeneously iso- and hyperintense relative to skeletal muscle in T2-weighted images. Computed tomography revealed no pulmonary metastasis. The serum CA125 level was 6.6 U/ml (normal: <40 U/ml). Open biopsy showed that the tumor was composed of sheets of large cells with vesicular chromatin, prominent nucleoli, and amphophilic cytoplasm, with peripheral palisading of epithelioid cells around necrotic areas (Figure S1B). Immunohistochemical analysis revealed that the tumor was positive for AE1/AE3 and vimentin, but negative for CD34, CA125, and S-100. (Figure S1C). Although the tumor was weakly positive for INI1 analyzed by immunohistochemistry, fluorescence in situ hybridization (FISH) analysis revealed the heterozygous deletion of INI1 in 17 of 50 tumor cells (34%) (Figure S1D). Upon these findings, the tumor was diagnosed as proximal-type epithelioid sarcoma. Wide resection of the tumor and lymph node dissection were performed, but systemic chemotherapy was not. Unfortunately, pulmonary metastases developed 12 weeks after surgery and the patient died 16 weeks after the definitive surgery.
Establishment of a new ES cell line, ESX
The resected specimen of the primary tumor was rinsed with phosphate-buffered saline, cut into small pieces with a scalpel and cultured in Iscove’s modified Dulbecco’s Eagle’s medium (IMDM; GIBCO BRL, Grand Island, NY) with 10% heat-inactivated fetal bovine serum (FBS; HyClone Laboratories, Inc., South Logan, UT). The tumors were incubated at 37°C in 5% CO2. The cell line (ESX) was maintained for more than 24 months.
Cell lines
Human osteosarcoma cell lines (NY, U2OS and HOS), human Ewing sarcoma cell lines (SKES, WES, and RDES), the human synovial sarcoma cell line FUJI, and the human ES cell line VA-ES-BJ were purchased from the Japanese Collection of Research Bioresources Cell Bank (Tokyo, Japan) and American Type Culture Collection (Manassas, VA, USA). The human synovial sarcoma cell line YaFuSS and FU-EPS-1 were gifts from Dr J. Toguchida (Kyoto University) [9] and Dr H. Iwasaki (Fukuoka University) [10]. The human osteosarcoma cell line OS2000 and KIKU, and the human malignant fibrous histiocytoma cell lines MFH2003 and MFH2004 were established in our laboratory [11-14].
ALDEFLUOR assay
The ALDEFLUOR kit (StemCell Technologies, Vancouver, Canada) was used to separate the population with high ALDH1 activity. Cells (1×106) were suspended in ALDEFLUOR assay buffer containing an ALDH1 substrate, bodipy-aminoacetaldehyde, at the concentration of 1μmol/L and incubated for 50 min at 37°C according to the manufacturer’s protocol. A specific inhibitor of ALDH1, diethylaminobenzaldehyde (DEAB), was used at 50mmol/L as a negative control.
CD109-positive cell sorting
The cells were washed once with PBS and then centrifuged at 440g at 4°C for 5 min using an LX120 (Tomy, Tokyo, Japan). The cell pellets were resuspended and incubated for 60 min at 4°C with a 100-fold dilution of a mouse anti-CD109 antibody (R&D Systems). Then samples were washed with PBS 3 times and stained and incubated for 60 min at 4°C with a 500-fold dilution of an FITC-labeled anti-mouse secondary antibody (KPL, Gaithersburg, MD). Cell sorting was performed using a FACSAria II (BD Bioscience, San Jose, CA). Collected data were analyzed using BD FACSDiva V6.1.3 (BD Bioscience). Propidium isodide (PI; Life Technologies Corp.) was used to stain live cells.
RNA preparation and PCR analysis
Total RNAs were extracted from cells using the RNeasyMini Kit (Qiagen, Hilden, Germany). cDNA was synthesized using Superscript III and an oligo(dT) primer (Life Technologies Corp.). Human Multiple Tissue cDNA Panels I and II, and the Human Fetal Multiple Tissue cDNA Panel (Clontech; Mountainview, CA) were used as normal tissue cDNAs. PCR was performed using KOD Dash (TOYOBO, Osaka, Japan) to detect CD109. The primer sequences used were 5’-TTGAATTCCCAATCCTGGAG-3’ and 5’-TTGTTGCCACTAACCACCAA-3’. The PCR mixture was denatured at 98°C for 2 min, followed by 30 cycles at 98°C for 15s, at 55°C for 2s, and at 74°C for 30s. GAPDH and beta-actin were used as internal controls. Real-time PCR was performed using the StepOne system (Life Technologies Corp.). Primers and probes were designed using the TapMan Gene expression assay (Life Technologies Corp.). Thermal cycling was performed with 40 cycles of 95°C for 1s, followed by 60°C for 20 min. Each experiment was done in triplicate and normalized to the GAPDH gene as an internal control.
siRNA
CD109 siRNA (siCD109) (5’-AAAGUUUGGACUCUGAUGACACCCA-3’) was designed using BLOCK-it RNAi (Life Technologies Corp.). Control siRNA was obtained from Life Technologies Corp. The siRNAs were transfected using Lipofectamine RNAiMAX transfection reagent (Life Technologies Corp.)
Spherical colony formation assay
Cells were plated at 1000 cells per well in six-well ultra-low attachment plates (Corning Inc., Corning, NY) and cultured in DMEM/F12 medium with 10ng/ml hEGF, 10ng/ml hbFGF, and 2% B-27 (Life Technologies Corp.) at 37°C in 5% CO2. On day 7, the number of colonies was counted under an inverted contrast microscope.
Cell proliferation assay
Cells were seeded in duplicate at a density of 2.5×104 cells/well in 24-well plates. On the following day, siRNAs were transfected. At 48 hr, 72 hr and 120 hr cells were trypsinized and counted with a Coulter Counter (Beckman Coulter, Inc. Brea, CA).
Basement membrane matrix invasion assay
Invasiveness was analyzed using a Matrigel™ Invasion Chamber (BD Biosciences). Briefly, 2.5×104 cells were placed on inserts in the wells in IMDM with 10% FBS. After 52 hours, the cells were stained using a Hemacolor staining kit (Merck Millipore, Billerica, MA) and the degree of migration was determined.
Gene expression profiling
RNA from ALDHhigh cells was labeled with Cy5 dye and RNA from ALDHlow cells were labeled with Cy3 dye. The probe mixture was hybridized for 40 hours at 65°C using a Human Whole Genome Microarray (G4112F) (Agilent Technologies, Santa Clara, CA). The array was scanned after washing with a G2565BA Microarray Scanner and fluorescent signals were acquired using Feature Extraction software (Agilent Technologies). The average expression ratio of Cy5 to Cy3 was determined per gene. A dye swap experiment was also done to label ALDHhigh and ALDHlow cells with Cy3 and Cy5, respectively. An average ratio of more than 2.0, reproducible in 2 experiments, was determined to indicate differential up-regulation in ALDHhigh cells. The accession number of ArrayExpress is E-MEXP-3826. We focused on membrane protein-related genes for cell sorting and as therapeutic targets using antibodies. Therefore, we selected membrane protein-related genes based on information obtained from GeneCards (http://www.genecards.org) (Table S2), followed by screening of mRNA expression in ALDHhigh and ALDHlow cells by RT-PCR (data not shown).
Xenografting
ALDHhigh and ALDHlow cells freshly sorted from cell lines were washed and suspended in PBS. Then the cells (1×102, 1×103 and 1×104 cells in 50μl of PBS) were mixed with an equal volume of Matrigel (BD Science) and subcutaneously injected into the bilateral sides of the lower back in female non-obese diabetic/severe combined immunodeficiency (NOD/SCID) mice (NOD.CB17-Prdkcscid/J, Charles River Laboratory, Yokohama, Japan). Tumor growth was monitored weekly for 10 weeks. Then the xenografted tumors were resected and analyzed.
Immunohistochemical staining
Formalin-fixed paraffin-embedded sections from 80 STS patients who underwent resection for stage I-III tumors and chemotherapy or radiotherapy for stage IV tumors between 2004 and 2009 in the Sapporo Medical University Hospital were used for CD109 staining as previously described [15]. The reactivity of the anti-CD109 antibody was determined by the staining pattern of the tumor cell membrane and graded as follows: 0 (no staining), 1 (partial staining of the membrane), 2 (mild to moderate circumferential staining of the membrane) and 3 (strong circumferential staining of the membrane). If the score was 2 or 3 in more than 10% of the tumor cells, it was considered to be positive. The clones used, antigen retrieval methods, and commercial sources of the antibodies used in the study are listed in Table S1.
Statistical methods
The Mann-Whitney test was used to compare in vitro data and the differences in tumor volume using IBM SPSS Statistics (IBM Corp., Armonk, NY). The Fisher exact test was used to compare the associations between the CD109 expression level and clinicopathological factors using SAS software 9.3 (SAS Institute, Cary, NC). Postoperative disease-free survival (DFS) and overall survival (OS) were estimated using Kaplan-Meier plots. Prognostic significance was evaluated by the log-rank test. Univariate and multivariate analyses for hazards ratios (HR) in OS were performed by Cox’s proportional-hazards regression with backward selection using IBM SPSS Statistics. A probability of less than 0.05 was considered statistically significant.
Results
Establishment of the ES cell line ESX
A tumor cell culture obtained from a patient with ES of the left thigh (Figure S1) was maintained for over 1 year and designated ESX (Figure 1A). The ESX cells were spindle-shaped with large nuclei and grew as adherent, tightly packed monolayers, but had no epithelial cell morphology. The morphology was maintained across all cell passages. Karyotype analysis revealed massive rearrangement of chromosomes (Figure 1B). Immunohistochemical examination revealed that the atypical cells were positive for vimentin and AE1/AE3, but negative for CD34 (Figure 1C). Subcutaneous inoculations of ESX cells into NOD/SCID mice produced growing tumors. Histologically the xenografted tumors consisted of a distinct nodular arrangement of the tumor cells, a tendency to undergo central degeneration and necrosis, and an epithelioid appearance with cytoplasmic eosinophilia. Immunostaining analysis of the xenografted tumors also revealed a staining pattern similar to that of the ESX cells (Figure 1C) and the original tumor (Figure S1C). These findings indicated that the established cell line was consistent with the profile of the original tumor. The characteristics of ESX were consistent with the profile of the highly malignant original tumor.
Identification of an ALDHhigh population in ES cell lines
We performed ALDEFLUOR assay to detect ALDHhigh populations containing CSCs/CICs in the epithelioid sarcoma cell lines. As shown in Figure 2A, all 3 ES cell lines (ESX, VA-ES-BJ, and FU-EPS-1) contained ALDHhigh populations, although the proportion of ALDHhigh cells varied. The mean proportions of ALDHhigh cells were 36.6%, 14.2 % and 13.8% in ESX, VA-ES-BJ and FU-EPS-1, respectively. The proportions of ALDHhigh cells in ES cell lines were higher than in the other sarcoma cell lines (Figure S2). The proportion of ALDHhigh cells in ESX was significantly higher than in the others (p<0.001) (Figure 2B). We then analyzed the differentiation abilities of ALDHhigh and ALDHlow ESX cells in vitro 9 days and 12 days after sorting (Figure 2C). ALDHhigh and ALDHlow cells showed a tendency to differentiate into ALDHlow and ALDHhigh, respectively. These results could indicate the flexible plasticity of CSCs/CICs of ES cells. The frequency of ALDHhigh cells converted from ALDHlow cells after in vitro culture was 16.3% on Day 12. On the other hand, when sorted ALDHhigh cells were cultured in vitro the frequency of remaining ALDHhigh cells was 36.2% on Day 12, indicating that ALDHhigh cells could maintain higher enzyme activity than ALDHlow cells.
Cancer-initiating ability of ALDHhigh cells of ES cell lines
In a previous study, we showed that CSCs/CICs could generate floating spheroid-like bodies in a serum-free medium [16]. We therefore determined whether ALDHhigh and ALDHlow cells of ESX could generate spherical colonies. As shown in Figure 3A, most ALDHlow cells died and the others formed a few small colonies. In contrast, the number of colonies derived from ALDHhigh cells was significantly higher than that from ALDHlow cells (p<0.001) (Figure 3B). Next, we examined the expression of stem/progenitor cell-related genes Sox2, Oct3/4, and Nanog [17,18]. Interestingly, in the ALDHhigh population the mRNA expression of Sox2, Oct3/4, and Nanog was lower than in ALDHlow (Figure 3C). These findings were in marked contrast to the CSCs/CICs of carcinomas, which suggested that gene expression of stem cell-related genes was not involved in the cancer-initiating ability, at least in epithelioid sarcoma [19]. In the ALDHhigh cells of VA-ES-BJ and FU-EPS-1, the mRNA expression of stem/progenitor cell-related genes was not lower than in ALDHlow cells (Figure S3).
To determine whether CSCs/CICs were abundant in the ALDHhigh population, we performed xenograft transplantation of ALDHhigh and ALDHlow cells of ESX, VA-ES-BJ, and FU-EPS-1 into NOD/SCID mice. The sorted ALDHhigh and ALDHlow cells were injected subcutaneously into mice, and tumor growth was monitored weekly for 10 weeks. In ESX, 1×103 ALDHhigh cells showed tumorigenicity in one of four mice. In contrast, 1×103 of ALDHlow cells failed to form tumors. On the other hand, 1×104 ALDHhigh and ALDHlow cells both showed tumorigenicity (Figure 3, D and E). However, the frequency of tumor formation was lower for ALDHlow cells than for ALDHhigh cells (Figure 3F). The histology of the tumors derived from ALDHhigh and ALDHlow cells of ESX showed no major differences (Figure 3G). In the case of FU-EPS-1, neither ALDHhigh nor ALDHlow cells could form tumors. In VA-ES-BJ, ALDHlow cells formed a tumor but ALDHhigh cells did not (Figure 3F). These findings suggested that ALDHhigh cells of ESX contained a higher number of tumorigenic cells compatible with CSCs/CICs than ALDHlow cells. In contrast, we considered that VA-ES-BJ and FU-EPS-1 did not contain a CSCs/CICs population in ALDHhigh cells that was similar to that of ESX. Therefore, ESX was used for the further characterization of the ALDHhigh population containing CSCs/CICs.
To examine the invasive potential of ALDHhigh and ALDHlow cells in ESX, we performed in vitro basement membrane matrix invasion assay. The microscopic features of the invading cells of ALDHhigh and ALDHlow are shown in Figure 4A. The number of invading cells for ALDHhigh was significantly higher than that for ALDHlow cells (P<0.001) (Figure 4B). Furthermore, the mRNA expression of epithelial-mesenchymal transition (EMT)-related genes Snail1 and Twist1 was upregulated in ALDHhigh cells, supporting the invasive ability of the ALDHhigh population (Figure 4C), which was compatible with the characteristics of CSCs/CICs.
Identification of the novel marker CD109 for the CSCs/CICs of ES
We screened the upregulated genes in ALDHhigh cells using a cDNA microarray to identify membrane protein-related genes for cell sorting and therapeutic targets using antibodies. The upregulated membrane protein-related genes in ALDHhigh cells of ESX are summarized in Table S2. Among them, we selected CD109, which was upregulated in ALDHhigh cells, as a representative marker (Figure 5A). The other molecules listed in Table S2 were not upregulated in ALDHhigh cells assessed by RT-PCR (data not shown). CD109, a GPI-anchored glycoprotein, was originally identified as a leukemia antigen. It has been reported that CD109 is expressed on activated T lymphocytes and platelets, endothelial cells and a subpopulation of CD34+ hematopoietic stem and progenitor cells [20,21]. CD109 has also been reported as the cell surface antigen on acute myeloid leukemia [20]. On the other hand, in the other ES cell lines there was no difference of CD109 mRNA expression between ALDHhigh and ALDHlow cells. These results supported the hypothesis that CD109 could be essential to maintain the cancer-initiating ability of CSCs/CICs in ALDHhigh cells of ESX.
We assessed the CD109 mRNA expression profiles in sarcomas and human fetal and normal adult tissues by RT-PCR. In addition to ES, CD109 mRNA expression was observed in 9 of 14 (64%) sarcoma cell lines (Figure 5B). As shown in Figure 5C, CD109 mRNA was weakly expressed in fetal skin, spleen, liver, lung, kidney, heart, pancreas and colon tissues. In adult tissues, CD109 mRNA was also weakly expressed in the lung, heart, small intestine, and testis. However, CD109 protein was expressed in alveoli of the lung, but not in heart and colon tissues (Figure 5D). These results suggested that CD109 could be an antigen highly expressed in various human sarcomas.
CD109 is associated with cancer-initiating ability and TGFβ/Smad signaling
The proportion of CD109-positive cells was 0.2% in ESX (Figure 6A), which was lower than that of ALDHhigh cells (Figure 1A). However, higher expression of ALDH1 mRNA was detected in CD109-positive cells than in CD109-negative cells (Figure 6B). Moreover, the number of spherical colonies derived from CD109-positive cells was higher than that from CD109-negative cells (Figure 6, C and D). In addition, the expression of stemness-related genes was lower in CD109-positive cells than in CD109-negative cells (Figure 6E). These features of CD109-positive cells were similar to those of ALDHhigh cells and suggested that CD109 might regulate cancer-initiating ability in ESX.
Next, we examined the effect of CD109 knockdown on cell proliferation. ESX cells were treated with siCD109, trypsinized and counted after 48 hr, 72 hr and 120 hr. Expression of both mRNA and protein of CD109 was downregulated (Figure S4, A, B and C). As shown in Figure 7A, siCD109 significantly inhibited cell proliferation (p<0.05). These results also supported the idea that CD109 plays an important role in cancer-initiating ability.
CD109 is a component of the TGF-β1 receptor 1 complex and negatively regulates TGF-β/Smad signaling [22]. Therefore, we hypothesized that CD109 could positively regulate ALDH1 activity and negatively regulate TGFβ1R1 expression in the TGF-β/Smad signaling pathway. We examined whether downregulation of ALDH1A1 and upregulation of TGFβ1R1 were induced by silencing of CD109. As shown in Figure 7B, ALDH1A1 mRNA was downregulated in siCD109-transfected ESX cells. On the other hand, TGFβ1R1 mRNA was upregulated in the siCD109-transfected cells (Figure 7C). Moreover, silencing of CD109 reduced the ALDHhigh population in ESX cells (Figure 7, D and E). These findings suggested that CD109 positively regulated ALDH1A1 activity and negatively regulated the TGF-β/Smad signaling pathway.
CD109 protein is associated with poor prognosis in soft tissue sarcoma patients
To determine the clinical relevance of CD109 expression in soft tissue sarcomas (STS), we evaluated CD109 expression by immunohistochemistry in the primary extremity lesions of 80 STSs. The clinical characteristics of these patients are summarized in Table S3. Representative staining patterns with the anti-CD109 antibody are shown in Figure 8A. Positive expression of CD109 protein was identified in 18% (15/80) of the STSs. Higher expression of CD109 was significantly associated with histologic grade, tumor stage, and distant metastases (p=0.021, 0.0012, and 0.0003 respectively). However, no other significant correlation was found between CD109 expression and other clinical parameters of the STSs. As shown in Figure 8B and 8C, positive CD109 expression, including in well-differentiated liposarcomas, was significantly associated with decreased probabilities of overall survival (OS) and disease-free survival (DFS) (P=8.3×10-5 and 4.5×10-4, respectively). Moreover, excluding well-differentiated liposarcomas, positive CD109 expression was also significantly associated with decreased probabilities of OS and DFS (P=0.006 and 0.049, respectively). The OS rates at 5 years were 46.7% for CD109-positive patients and 85.3% in those who were CD109 negative. Several variables were tested to assess whether they had an impact on survival. Univariate and multivariate analyses revealed that the histologic grade and expression of CD109 were independent risk factors for poor outcome (Table 1, Table 2). The hazard ratio (HR) of OS for the CD109-positive group was 3.85 (95% confidence interval CI, 1.40-10.55). Thus, CD109 might be a predictive biomarker for both distant metastasis and the prognosis of STSs in the clinical setting.
Table 1. Univariate analysis of the variables in overall survival.
Variable | Hazard ratio | 95% CI | P value |
---|---|---|---|
Tumor size | |||
<5cm | 1 | ||
5-10cm | 3.38 | 0.70-16.2 | 0.129 |
>10cm | 2.21 | 0.46-10.6 | 0.321 |
Tumor depth | |||
Superficial | 1 | ||
Deep | 24.0 | 0.024-24154 | 0.367 |
Histologic grade | |||
I | 1 | ||
II | 3.18 | 0.28-35.1 | 0.344 |
III | 20.4 | 2.66-157 | 0.004 |
CD109 expression | |||
Negative | 1 | ||
Positive | 5.72 | 2.13-15.3 | 0.001 |
CI: confidence interval
Table 2. Multivariate analysis of the variables in overall survival.
Variable | Hazard ratio | 95% CI | P value |
---|---|---|---|
Histologic grade | |||
I | 1 | ||
II | 2.03 | 0.177-23.39 | 0.570 |
III | 13.9 | 1.75-111.7 | 0.013 |
CD109 expression | |||
Negative | 1 | ||
Positive | 3.85 | 1.40-10.55 | 0.009 |
CI: confidence interval
Discussion
In the present study, we (i) established and characterized the new ES cell line ESX; (ii) demonstrated that the ALDHhigh population of ESX contained CSCs/CICs showing in vitro and in vivo tumorigenesis; (iii) found that high expression of CD109 in ALDHhigh cells was important to maintain tumorigenesis as the important feature of CSCs/CICs; and (iv) showed the prognostic impact of CD109 expression on patients with STSs. ES is a rare, slow-growing malignant tumor. With an infiltrative growth pattern and a propensity for extension along fascial planes, and nerves, it is characteristically associated with multiple local recurrences and late metastasis [23]. The effects of multiagent chemotherapy and radiotherapy remain unclear; therefore, novel therapeutic options need to be developed. The establishment of an experimental model is imperative to investigate the biological characteristics of ES and develop novel therapeutic options. To the best of our knowledge, only 11 human ES cell lines have been reported to date [10,24-33]. ESX might therefore facilitate further studies on the biological characteristics of this rare tumor entity.
Using ESX, we tried to identify CSCs/CICs of ES. Previously, we performed side population (SP) analysis to identify CSCs/CICs of sarcomas [16]. However, SP cells were hardly detected in ESX (data not shown). Therefore, we used the ALDEFLUOR assay. In this study, we demonstrated that an ALDHhigh population existed in ESX, and that the ALDHhigh cells possessed repopulating capacity and high tumor-forming ability in vitro and in vivo. ALDH is a cytosolic isoenzyme involved in the detoxification of intracellular aldehydes by oxidation and conversion of retinol to retinoic acid, and it confers resistance to chemotherapeutic agents such as cyclophosphamide [34]. Therefore, it makes sense that ALDHhigh cells contain CSCs/CICs. Visus et al. were the first to describe the use of ALDH1A1 as a putative CSC/CIC marker in head and neck squamous cell carcinomas [35]. In the field of sarcoma, Awad et al. tested ALDH activity in Ewing sarcoma, defining CSCs/CICs as those cells that showed the highest ALDH activity [36]. However, there are no other reports regarding CSCs/CICs of sarcoma. This is the first report identifying and characterizing of CSCs/CICs of ES. The proportion of ALDHhigh cells in ES cell lines was clearly higher than in the other sarcoma cell lines (Figure S2). These results were compatible with the chemotherapy-resistant characteristics of ES.
Expression of the stem cell-related genes Sox2, Oct3/4, and Nanog is also used to characterize CSCs/CICs [37]. These are essential for the maintenance of pluripotent embryonic stem cells and germ cells, as well as CSCs/CICs [38]. Twist and Snail, which could promote EMT, are also considered to be key factors in the maintenance of CSCs/CICs. Twist and Snail induce not only increased potential for invasiveness and metastases, but also increased ability to form spheres and generate tumors in xenografts [39]. In this study, the ALDHhigh population expressed Twist 1 and Snail 1 at higher levels than ALDHlow cells. ALDHhigh cells of ESX had higher invasive ability, which was considered to be compatible with previous reports [7,8]. On the other hand, expression of Sox2, Oct3/4 and Nanog was lower in ALDHhigh cells. The reason for this discrepancy in the expression status between ALDHhigh and ALDHlow cells in the case of ESX remains unknown. Velcheti et al. reported that high expression of Sox2 was correlated with good prognosis in patients with non-small cell lung carcinomas [40]. These results suggested that Sox2 was not necessarily related to tumorigenicity and malignant potential. Alternatively, we speculate that CSCs/CICs of ES might have more differentiated characteristics with downregulation of Sox2 than those of carcinomas.
Using a cDNA microarray, we found that CD109 was upregulated in ALDHhigh cells. CD109, a GPI-anchored glycoprotein, was originally identified as a leukemia antigen. It has been reported that CD109 is expressed on activated T lymphocytes and platelets, endothelial cells and a subpopulation of CD34+ hematopoietic stem and progenitor cells [20,21,41]. CD109 is also expressed in keratinocytes and contributes to the inhibition of extracellular matrix production in scleroderma [42]. In addition, it is also expressed in malignancies of the lung, esophagus, cervix, urinary tract and breast and plays a role in the tumor growth of oral cancer [43]. The present study also demonstrated that CD109 was highly expressed in sarcoma but not in normal tissues. The expression status suggested that CD109 might be a candidate therapeutic target not only for sarcoma but also for epithelial cancer.
CD109 is a TGF-β co-receptor, a component of the TGF-β1 receptor 1 (TGFβ1R1) complex. It accelerates TGF-β receptor degradation and negatively regulates TGF-β/Smad signaling [44,45]. In some human cancers, CD109 actually impairs TGF-β/Smad signaling [46]. TGF-β can play both tumor-suppressive and tumor-promoting roles. Especially in the early phase of cancer initiation, TGF-β acts as an anti-oncogenic factor [47]. In this study, ALDHhigh cells of ESX showed higher tumorigenicity and higher expression of CD109. In addition, silencing of CD109 upregulated TGFβ1R1 mRNA in ESX cells. These findings suggested that CD109 expressed in ALDHhigh cells of ESX promoted cancer-initiating ability as the result of the CD109-mediated inhibition of TGF-β.
CD109-positive cells of ESX highly expressed ALDH1 and showed higher tumorigenicity than CD109-negative cells. Furthermore, the expression status of stemness-related genes of CD109-positive cells was similar to that of ALDHhigh cells of ESX. Therefore, we considered that CD109 could be a representative molecule of ALDHhigh cells. In addition, knockdown of CD109 decreased ALDH1 activity in ESX cells. These results suggested that CD109 might regulate ALDH1 activity and confer the characteristics of CSCs/CICs in ESX cells. However, the proportion of CD109-positive cells was lower than that of ALDHhigh cells of ESX. Therefore we hypothesized that other factors might also regulate ALDH1 activity.
The evaluation of the clinical specimens of STSs revealed a strong correlation between CD109 expression and DFS and OS, suggesting that CD109 could be a promising prognostic biomarker in STSs. Although it is reported that CD109 is preferentially expressed in the early stage of tumorigenesis in oral tumor and urothelial carcinomas [43,46], we demonstrated that high expression of CD109 was significantly associated with advanced stage in STSs. These results suggest that the pathophysiological function of CD109 protein in sarcomas is different from that in carcinomas. Therefore, we speculate that CD109 expression in STSs is deeply involved in invasion and metastasis.
In conclusion, we established the novel ES cell line ESX. Next, we investigated CICs/CSCs in ES cell lines and isolated CSCs/CICs based on ALDH activity. Finally, we demonstrated that CD109 is a potential CSC/CIC marker, prognostic factor and molecular target for STSs, including ES.
Supporting Information
Funding Statement
Supported by grants from JSPS KAKENHI (21249025, 20390403, 22689041), Management Expenses Grants from the Government to the National Cancer Center (23-A-10, 23-A-44), NOASTEC (H24-S-5) and Suhara Memorial Foundation (H24-12). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
References
- 1. Weiss SW, Goldblum JR (2008) Enzinger and Weiss’s soft tissue tumors. Mosby, Inc. [Google Scholar]
- 2. Jordan CT, Guzman ML, Noble M (2006) Cancer stem cells. N Engl J Med 355: 1253-1261. doi: 10.1056/NEJMra061808. PubMed: 16990388. [DOI] [PubMed] [Google Scholar]
- 3. Sládek NE (2003) Human aldehyde dehydrogenases: potential pathological, pharmacological, and toxicological impact. J Biochem Mol Toxicol 17: 7-23. doi: 10.1002/jbt.10057. PubMed: 12616643. [DOI] [PubMed] [Google Scholar]
- 4. Vasiliou V, Pappa A, Petersen DR (2000) Role of aldehyde dehydrogenases in endogenous and xenobiotic metabolism. Chem Biol Interact 129: 1-19. doi: 10.1016/S0009-2797(00)00211-8. PubMed: 11154732. [DOI] [PubMed] [Google Scholar]
- 5. van den Hoogen C, van der Horst G, Cheung H, Buijs JT, Lippitt JM et al. (2010) High aldehyde dehydrogenase activity identifies tumor-initiating and metastasis-initiating cells in human prostate cancer. Cancer Res 70: 5163-5173. doi: 10.1158/0008-5472.CAN-09-3806. PubMed: 20516116. [DOI] [PubMed] [Google Scholar]
- 6. Cheung AM, Wan TS, Leung JC, Chan LY, Huang H et al. (2007) Aldehyde dehydrogenase activity in leukemic blasts defines a subgroup of acute myeloid leukemia with adverse prognosis and superior NOD/SCID engrafting potential. Leukemia 21: 1423-1430. doi: 10.1038/sj.leu.2404721. PubMed: 17476279. [DOI] [PubMed] [Google Scholar]
- 7. Jiang F, Qiu Q, Khanna A, Todd NW, Deepak J et al. (2009) Aldehyde dehydrogenase 1 is a tumor stem cell-associated marker in lung cancer. Mol Cancer Res 7: 330-338. doi: 10.1158/1541-7786.MCR-08-0393. PubMed: 19276181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Li T, Su Y, Mei Y, Leng Q, Leng B et al. (2010) ALDH1A1 is a marker for malignant prostate stem cells and predictor of prostate cancer patients' outcome. Lab Invest 90: 234-244. doi: 10.1038/labinvest.2009.127. PubMed: 20010854. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Ishibe T, Nakayama T, Okamoto T, Aoyama T, Nishijo K et al. (2005) Disruption of fibroblast growth factor signal pathway inhibits the growth of synovial sarcomas: potential application of signal inhibitors to molecular target therapy. Clin Cancer Res 11: 2702-2712. doi: 10.1158/1078-0432.CCR-04-2057. PubMed: 15814652. [DOI] [PubMed] [Google Scholar]
- 10. Nishio J, Iwasaki H, Nabeshima K, Ishiguro M, Naumann S et al. (2005) Establishment of a new human epithelioid sarcoma cell line, FU-EPS-1: molecular cytogenetic characterization by use of spectral karyotyping and comparative genomic hybridization. Int J Oncol 27: 361-369. PubMed: 16010416. [PubMed] [Google Scholar]
- 11. Wada T, Uede T, Ishii S, Matsuyama K, Yamawaki S et al. (1988) Monoclonal antibodies that detect different antigenic determinants of the same human osteosarcoma-associated antigen. Cancer Res 48: 2273-2279. PubMed: 2450650. [PubMed] [Google Scholar]
- 12. Nabeta Y, Kawaguchi S, Sahara H, Ikeda H, Hirohashi Y et al. (2003) Recognition by cellular and humoral autologous immunity in a human osteosarcoma cell line. J Orthop Sci 8: 554-559. doi: 10.1007/s00776-003-0663-5. PubMed: 12898310. [DOI] [PubMed] [Google Scholar]
- 13. Tsukahara T, Kawaguchi S, Ida K, Kimura S, Tamura Y et al. (2006) HLA-restricted specific tumor cytolysis by autologous T-lymphocytes infiltrating metastatic bone malignant fibrous histiocytoma of lymph node. J Orthop Res 24: 94-101. doi: 10.1002/jor.20019. PubMed: 16419974. [DOI] [PubMed] [Google Scholar]
- 14. Kimura S, Kozakai Y, Kawaguchi S, Tsukahara T, Ida K et al. (2008) Clonal T-cell response against autologous pleomorphic malignant fibrous histiocytoma antigen presented by retrieved HLA-A*0206. J Orthop Res 26: 271-278. doi: 10.1002/jor.20497. PubMed: 17853492. [DOI] [PubMed] [Google Scholar]
- 15. Tsukahara T, Kawaguchi S, Torigoe T, Kimura S, Murase M et al. (2008) Prognostic impact and immunogenicity of a novel osteosarcoma antigen, papillomavirus binding factor, in patients with osteosarcoma. Cancer Sci 99: 368-375. doi: 10.1111/j.1349-7006.2008.00695.x. PubMed: 18271936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Murase M, Kano M, Tsukahara T, Takahashi A, Torigoe T et al. (2009) Side population cells have the characteristics of cancer stem-like cells/cancer-initiating cells in bone sarcomas. Br J Cancer 101: 1425-1432. doi: 10.1038/sj.bjc.6605330. PubMed: 19826427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Nichols J, Zevnik B, Anastassiadis K, Niwa H, Klewe-Nebenius D et al. (1998) Formation of pluripotent stem cells in the mammalian embryo depends on the POU transcription factor Oct4. Cell 95: 379-391. doi: 10.1016/S0092-8674(00)81769-9. PubMed: 9814708. [DOI] [PubMed] [Google Scholar]
- 18. Chambers I, Colby D, Robertson M, Nichols J, Lee S et al. (2003) Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Cell 113: 643-655. doi: 10.1016/S0092-8674(03)00392-1. PubMed: 12787505. [DOI] [PubMed] [Google Scholar]
- 19. Yasuda K, Torigoe T, Morita R, Kuroda T, Takahashi A et al. (2013) Ovarian cancer stem cells are enriched in side population and aldehyde dehydrogenase bright overlapping population. PLOS ONE (in press). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sutherland DR, Yeo E, Ryan A, Mills GB, Bailey D et al. (1991) Identification of a cell-surface antigen associated with activated T lymphoblasts and activated platelets. Blood 77: 84-93. PubMed: 1984805. [PubMed] [Google Scholar]
- 21. Murray LJ, Bruno E, Uchida N, Hoffman R, Nayar R et al. (1999) CD109 is expressed on a subpopulation of CD34+ cells enriched in hematopoietic stem and progenitor cells. Exp Hematol 27: 1282-1294. doi: 10.1016/S0301-472X(99)00071-5. PubMed: 10428505. [DOI] [PubMed] [Google Scholar]
- 22. Finnson KW, Tam BY, Liu K, Marcoux A, Lepage P et al. (2006) Identification of CD109 as part of the TGF-beta receptor system in human keratinocytes. FASEB J 20: 1525-1527. doi: 10.1096/fj.05-5229fje. PubMed: 16754747. [DOI] [PubMed] [Google Scholar]
- 23. Hollmann TJ, Hornick JL (2011) INI1-deficient tumors: diagnostic features and molecular genetics. Am J Surg Pathol 35: e47-e63. doi: 10.1097/PAS.0b013e31822b325b. PubMed: 21934399. [DOI] [PubMed] [Google Scholar]
- 24. Reeves BR, Fisher C, Smith S, Courtenay VD, Robertson D (1987) Ultrastructural, immunocytochemical, and cytogenetic characterization of a human epithelioid sarcoma cell line (RM-HS1). J Natl Cancer Inst 78: 7-18. PubMed: 2432306. [DOI] [PubMed] [Google Scholar]
- 25. Kelland LR, Bingle L (1988) Radiosensitivity and characterisation of a newly established cell line from an epithelioid sarcoma. Br J Cancer 58: 322-325. doi: 10.1038/bjc.1988.211. PubMed: 3179184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Stenman G, Kindblom LG, Willems J, Angervall L (1990) A cell culture, chromosomal and quantitative DNA analysis of a metastatic epithelioid sarcoma. Deletion 1p, a possible primary chromosomal abnormality in epithelioid sarcoma. Cancer 65: 2006-2013. doi: 10.1002/1097-0142(19900501)65:9. PubMed: 2196989. [DOI] [PubMed] [Google Scholar]
- 27. Gerharz CD, Moll R, Ramp U, Mellin W, Gabbert HE (1990) Multidirectional differentiation in a newly established human epithelioid sarcoma cell line (GRU-1) with co-expression of vimentin, cytokeratins and neurofilament proteins. Int J Cancer 45: 143-152. doi: 10.1002/ijc.2910450126. PubMed: 1688830. [DOI] [PubMed] [Google Scholar]
- 28. Roché H, Julia AM, Jozan S, Muller C, Dastugue N et al. (1993) Characterization and chemosensitivity of a human epithelioid sarcoma cell line (SARCCR 2). Int J Cancer 54: 663-668. doi: 10.1002/ijc.2910540423. PubMed: 8099901. [DOI] [PubMed] [Google Scholar]
- 29. Sonobe H, Furihata M, Iwata J, Oka T, Ohtsuki Y et al. (1993) Morphological characterization of a new human epithelioid sarcoma cell line, ES020488, in vitro and in vivo. Virchows Arch B Cell Pathol Incl Mol Pathol 63: 219-225. doi: 10.1007/BF02899265. PubMed: 7685133. [DOI] [PubMed] [Google Scholar]
- 30. Cordoba JC, Parham DM, Meyer WH, Douglass EC (1994) A new cytogenetic finding in an epithelioid sarcoma, t(8;22)(q22;q11). Cancer Genet Cytogenet 72: 151-154. [DOI] [PubMed] [Google Scholar]
- 31. Helson C, Melamed M, Braverman S, Traganos F, Preti R et al. (1995) Va-es-bj - an epithelioid sarcoma cell-line. Int J Oncol 7: 51-56. PubMed: 21552805. [DOI] [PubMed] [Google Scholar]
- 32. Iwasaki H, Ohjimi Y, Ishiguro M, Isayama T, Kaneko Y et al. (1996) Epithelioid sarcoma with an 18q aberration. Cancer Genet Cytogenet 91: 46-52. doi: 10.1016/S0165-4608(95)00315-0. PubMed: 8908166. [DOI] [PubMed] [Google Scholar]
- 33. Kusakabe H, Sakatani S, Yonebayashi K, Kiyokane K (1997) Establishment and characterization of an epithelioid sarcoma cell line with an autocrine response to interleukin-6. Arch Dermatol Res 289: 224-233. doi: 10.1007/s004030050184. PubMed: 9143739. [DOI] [PubMed] [Google Scholar]
- 34. Russo JE, Hilton J (1988) Characterization of cytosolic aldehyde dehydrogenase from cyclophosphamide resistant L1210 cells. Cancer Res 48: 2963-2968. PubMed: 3365687. [PubMed] [Google Scholar]
- 35. Visus C, Ito D, Amoscato A, Maciejewska-Franczak M, Abdelsalem A et al. (2007) Identification of human aldehyde dehydrogenase 1 family member A1 as a novel CD8+ T-cell-defined tumor antigen in squamous cell carcinoma of the head and neck. Cancer Res 67: 10538-10545. doi: 10.1158/0008-5472.CAN-07-1346. PubMed: 17974998. [DOI] [PubMed] [Google Scholar]
- 36. Awad O, Yustein JT, Shah P, Gul N, Katuri V et al. (2010) High ALDH activity identifies chemotherapy-resistant Ewing's sarcoma stem cells that retain sensitivity to EWS-FLI1 inhibition. PLOS ONE 5: e13943. doi: 10.1371/journal.pone.0013943. PubMed: 21085683. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Nakatsugawa M, Takahashi A, Hirohashi Y, Torigoe T, Inoda S et al. (2011) SOX2 is overexpressed in stem-like cells of human lung adenocarcinoma and augments the tumorigenicity. Lab Invest 91: 1796-1804. doi: 10.1038/labinvest.2011.140. PubMed: 21931300. [DOI] [PubMed] [Google Scholar]
- 38. Jaenisch R, Young R (2008) Stem cells, the molecular circuitry of pluripotency and nuclear reprogramming. Cell 132: 567-582. doi: 10.1016/j.cell.2008.01.015. PubMed: 18295576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Schatton T, Frank MH (2008) Cancer stem cells and human malignant melanoma. Pigment. Cell - Melanoma Res 21: 39-55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Velcheti V, Schalper K, Yao X, Cheng H, Kocoglu M et al. (2013) High SOX2 levels predict better outcome in non-small cell lung carcinomas. PLOS ONE 8: e61427. doi: 10.1371/journal.pone.0061427. PubMed: 23620753. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Hwang SM, Kim MJ, Chang HE, Hong YJ, Kim TS, et al. (2013) Human platelet antigen genotyping and expression of CD109 (human platelet antigen 15) mRNA in various human cell types. Biomed Res Int 2013: 946403 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Man XY, Finnson KW, Baron M, Philip A (2012) CD109, a TGF-beta co-receptor, attenuates extracellular matrix production in scleroderma skin fibroblasts. Arthritis Res Ther 14: R144. doi: 10.1186/ar3877. PubMed: 22694813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Hagiwara S, Murakumo Y, Sato T, Shigetomi T, Mitsudo K et al. (2008) Up-regulation of CD109 expression is associated with carcinogenesis of the squamous epithelium of the oral cavity. Cancer Sci 99: 1916-1923. PubMed: 19016750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Hagiwara S, Murakumo Y, Mii S, Shigetomi T, Yamamoto N et al. (2010) Processing of CD109 by furin and its role in the regulation of TGF-beta signaling. Oncogene 29: 2181-2191. doi: 10.1038/onc.2009.506. PubMed: 20101215. [DOI] [PubMed] [Google Scholar]
- 45. Bizet AA, Liu K, Tran-Khanh N, Saksena A, Vorstenbosch J et al. (2011) The TGF-beta co-receptor, CD109, promotes internalization and degradation of TGF-beta receptors. Biochim Biophys Acta 1813: 742-753. doi: 10.1016/j.bbamcr.2011.01.028. PubMed: 21295082. [DOI] [PubMed] [Google Scholar]
- 46. Hagikura M, Murakumo Y, Hasegawa M, Jijiwa M, Hagiwara S et al. (2010) Correlation of pathological grade and tumor stage of urothelial carcinomas with CD109 expression. Pathol Int 60: 735-743. doi: 10.1111/j.1440-1827.2010.02592.x. PubMed: 20946523. [DOI] [PubMed] [Google Scholar]
- 47. Ikushima H, Miyazono K (2012) TGF-beta signal transduction spreading to a wider field: a broad variety of mechanisms for context-dependent effects of TGF-beta. Cell Tissue Res 347: 37-49. doi: 10.1007/s00441-011-1179-5. PubMed: 21618142. [DOI] [PubMed] [Google Scholar]
- 48. Miura Y, Keira Y, Ogino J, Nakanishi K, Noguchi H et al. (2012) Detection of specific genetic abnormalities by fluorescence in situ hybridization in soft tissue tumors. Pathol Int 62: 16-27. doi: 10.1111/j.1440-1827.2011.02739.x. PubMed: 22192800. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.