SEARCH

SEARCH BY CITATION

Keywords:

  • Bone marrow–derived mesenchymal stem cell;
  • Malignant transformation;
  • Chromosomal instability;
  • Fibrosarcoma

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosures
  9. References

Despite recent emerging evidence suggesting that cancer stem cells subsist in a variety of tumors, it is not yet fully elucidated whether postnatal stem cells are directly involved in tumorigenesis. We used murine bone marrow–derived mesenchymal stem cells (BMMSCs) as a model to test a hypothesis that tumorigenesis may originate from spontaneous mutation of stem cells. In this study, we demonstrated that murine BMMSCs, after numerous passages, obtained unlimited population doublings and proceeded to a malignant transformation state, resulting in fibrosarcoma formation in vivo. Transformed BMMSCs colonized to multiple organs when delivered systemically through the tail vein. Fibrosarcoma cells formed by transformed BMMSCs contained cancer progenitors, which were capable of generating colony clusters in vitro and fibrosarcoma in vivo by the second administration. The mechanism by which BMMSCs transformed to malignant cells was associated with accumulated chromosomal abnormalities, gradual elevation in telomerase activity, and increased c-myc expression. Moreover, BMMSCs and their transformed counterpart, fibrosarcoma-forming cells, demonstrated different sensitivity to anti-cancer drugs. BMMSCs/fibrosarcoma transformation system may provide an ideal system to elucidate the mechanism of how stem cells become cancer cells and to screen anti-sarcoma drugs.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosures
  9. References

Recently, the possibility that tumors originate from cancer stem cells has been proposed based on the fact that only a small percentage of cancer cells form tumors and that tumor cells exhibit stem cell properties such as hierarchical self-renewal, unlimited proliferation, differentiation capabilities, and use of stem cell–associated signaling pathways to maintain “stemness” [1]. Although cancer stem cells have been reported in leukemia [24], breast cancer [5], and brain tumor [6, 7], the exact origin of cancer stem cells remains unclear. In this study, we tried to assess whether bone marrow–derived mesenchymal stem cells (BMMSCs) are indeed involved, at least in part, in the generation of cancer stem cells, because long-living stem cells may have higher probability of accumulating mutations, resulting in tumorigenesis. To date, however, it is not clear whether post-natal stem cells can be transformed to malignant cells under the standard culture conditions without any genetic manipulation, exposure to viruses, and irradiation. Given the recent report that bone marrow–derived cells contribute to gastric cancer formation [8] and that human mesenchymal stem cells (MSCs) may be transformed to cancer cells [9], it is important to further explore the role, if any, of bone marrow–derived stem cells in promoting tumor formation. It has been reported that mouse embryonic fibroblasts were able to become established lines with a constant or rising growth rate after continuous in vitro culturing [10], implying the possibility of spontaneous immortalization of murine BMMSCs under similar culture conditions.

In the meantime, multipotent adult progenitor cells (MAPCs) were discovered as an infrequent population of BMMSCs maintained in the adult bone marrow compartment [11]. They are considered an extremely important stem cell resource for regenerative medicine because they were reported to possess pluripotent stem cell characteristics, similar to embryonic stem cells (ESCs); they differentiate into endoderm-, mesoderm-, and ectoderm-derived cells and proliferate extensively without entering into senescence. However, it is important to further verify characteristics of the cells with extensive proliferation in adult bone marrow.

In this study, we found that continuous passages led BMMSCs to spontaneous immortalization. After additional passages, the immortalized BMMSCs became transformed into malignant cells, capable of forming fibrosarcomas in vivo, at least partially, due to accumulated chromosomal abnormalities, amplified c-Myc expression, and elevated telomerase activity. This consecutive conversion of BMMSCs to malignant cells provides an excellent model to study the mechanisms associated with the tumorigenic potential of postnatal stem cells and explore therapeutic strategies for malignant tumors.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosures
  9. References

Mouse BMMSC Culture

Preparation and expansion of the murine BMMSCs was done based on a published method [12]. Briefly, bone marrow cells from adult C57Bl/6 mice were seeded into culture dishes, incubated for 4 hours, and washed twice with α-minimum essential medium (α-MEM) (Gibco, Grand Island, NY, http://www.invitrogen.com). The culture medium consisted of α-MEM, 20% fetal bovine serum (FBS; Equitech-Bio, Kerrville, TX, http://www.equitech-bio.com), 2 mM L-glutamine (Biofluids, Rockville, MD, http://www.biofluids.com), a combination of 100 U/ml penicillin and 100 μg/ml streptomycin (Biofluids), and 55 μM 2-mercaptoethanol (Gibco). Primary cultures (passage 0 [P0]) were passaged to disperse the colony-forming cells and seeded on freshly prepared culture dishes (P1). Initially, BMMSCs were passed at 1:2–5 dilution when they reached confluence. After BMMSCs were immortalized (at approximately P15), cells were passed at 1:10–20. To obtain single colony–derived BMMSC clones, bone marrow cells were seeded at 1–3 × 106 cells in 15-cm culture dishes and cultured as described above. Colonies were isolated at 14 to 21 days after cultivation. The proliferation rate of BMMSCs was assessed by bromodeoxyuridine (BrdU) incorporation for 4 hours using a BrdU staining Kit (Zymed Laboratories, Inc., San Francisco, http://www.zymed.com). In vitro osteogenic, adipogenic, and chondrogenic assays were done as previously reported [13, 14].

Human MSC Culture

Human bone marrow aspirates from healthy adult volunteers were purchased from AllCells, LLC (Berkeley, CA, http://www.allcells.com). To identify putative BMMSCs, single-cell suspension of 1 × 106 of bone marrow mononuclear cells was seeded into 15-cm culture dishes and nonadherent cells were removed after 4 hours of incubation at 37°C. The adherent cells were cultured with α-MEM supplemented with 15% FBS, 100 μM L-ascorbic acid 2-phosphate (Wako Pure Chemical Industries Ltd., Osaka, Japan, http://www.wako-chem.co.jp), 2 mM L-glutamine, and a combination of 100 U/ml penicillin and 100 μg/ml streptomycin. BMMSCs were plated at 1:4 dilution when the cells were approaching confluence. Isolation of dental pulp stem cells and periodontal ligament stem cells was described as previously [15, 16]. They were cultured in the same medium as that used for human BMMSCs and plated at 1:4 dilution when the cells were approaching confluence.

Transplantation and Injection of BMMSCs

Approximately 2–4 × 106 of murine BMMSCs were transplanted into 6- to 8-week-old immunocompromised (bg-nu/nu-xid) mice using hydroxyapatite/tricalcium phosphate (HA/TCP) ceramic powder (Zimmer, Inc., Warsaw, IN, http://www.zimmer.com) and harvested 8 weeks after transplantation [17]. To assess tumorigenesis and migration potential, 1 × 106 of BMMSCs were injected into bg-nu/nu-xid mice subcutaneously and intravenously through tail vein, respectively. In some experiments, tumor cells were dispersed by digesting tumors with 3 mg/ml collagenase type I (Worthington Biochem, Freehold, NJ, http://www.worthington-biochem.com) and 4 mg/ml dispase (Boehringer Mannheim, GmbH, Mannheim, Germany, http://www.boehringer.com). Then tumor cells were injected subcutaneously into second-recipient bg-nu/nu-xid mice. These procedures were performed in accordance with an approved small-animal protocol (National Institute of Dental and Craniofacial Research no. 04–317).

Histological Analysis

Transplants were harvested and demineralized with 10% EDTA before embedding. The paraffin-embedded sections were stained with hematoxylin and eosin, or Mallory trichrome, or incubated with the antibodies for vimentin (Zymed Laboratories, Inc.) and cytokeratin (DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com).

Fluorescence-Activated Cell Sorting Analysis

Cells (1 × 106) were incubated with 1 μg of phycoerythrin (PE)-conjugated antibodies for 45 minutes at 4°C. PE-conjugated isotype-matched immunoglobulin G (IgG) was used as control. Antibodies against CD45, TER119, CD13, Sca-1, Thy-1, CD34, c-kit, CD19, CD3, and CD18 were from BD Biosciences (San Diego, http://www.bdbiosciences.com), and SSEA-1 and Flk-1 were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, http://www.scbt.com).

Spectral Karyotyping and Fluorescent In Situ Hybridization

Metaphase chromosomes for spectral karyotyping (SKY) hybridization were prepared from BMMSCs at different passages. Cells in culture were incubated for 1–2 hours in 0.02 mg/ml Colcemid (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). The cells were incubated in hypotonic solution (0.075 M KCl) and fixed in methanol/acetic acid (3:1). SKY was then performed as described previously, using a combination of five different fluorochromes [18, 19]. Images were acquired with Sky acquisition software (Applied Spectral Imaging, Ltd, Migdal Haemek, Israel, http://www.spectral-imaging.com) using a spectral cube and a CCD (charge-coupled device) camera (Hamamatsu, Bridgewater, NJ, http://jp.hamamatsu.com) connected to a DMRXA microscope (Leica Imaging Systems, Cambridge, U.K., http://www.leica.com) with a custom SKY-3 optical filter (Chroma Technology Corp., Rockingham, VT, http://www.chroma.com). A minimum of 10 metaphases were imaged and karyotyped using SkyView version 1.6.2 software. To confirm the origin of the double-minute chromosomes detected by SKY, dual-color fluorescent in situ hybridization (FISH) was performed with a chromosome-painting probe for chromosome MMU15 and the c-myc gene. DNA was isolated from the BAC clone for c-myc (D15Mit17) and labeled with biotin-16-dUTP (Roche BioMolecular, Indianapolis, https://www.roche-applied-science.com) using a standard nick-translation protocol. Chromosome-painting probe for MMU15 (directly labeled with CyTm5) used for FISH hybridizations was prepared using the same protocols as for labeling of SKY chromosome-painting probes (http://www.riedlab.nci.nih.gov/protocols.asp). FISH probe for c-myc was labeled by Nick translation with biotin-16-dUTP (Roche BioMolecular) and visualized using CW4000 FISH software (Leica Imaging Systems). The condition for hybridization was as described in http://www.riedlab.nci.nih.gov/protocols.asp.

Western Blot Analysis

Cells were lysed in mammalian protein extraction reagent (Pierce Chemical Co., Rockford, IL, http://www.piercenet.com). Ten micrograms of protein per each lane was separated by SDS-PAGE, transferred to a nitrocellulose membrane, and probed with c-Myc (Santa Cruz Biotechnology, Inc.) and β-actin (Sigma, St. Louis, http://www.sigmaaldrich.com). Bound antibodies were revealed with goat anti-mouse IgG conjugated to horseradish peroxidase (HRP) (Santa Cruz Biotechnology, Inc.), and blots were developed using Super Signal chemiluminescence HRP substrate (Pierce Chemical Co.).

Quantitative Telomerase Activity Assay

The telomerase activity was evaluated using the two-step quantitative-PCR-based telomeric repeat amplification protocol assay as reported previously with slight modification [20]. Briefly, 40 μl of first-step reaction mixture, which contained 2.0 μg of cell lysate diluted into 26 μl of 0.1 mg/ml bovine serum albumin, 1X TRAP standard reaction buffer, 50 μM each deoxynucleotide triphosphate, 0.1 μg of TS oligonucleotide substrate (TS primer; AATCCGTCGAGCAGAGTT), and 0.2 μg of T4 gene protein (Amersham Pharmacia Biotech, Piscataway, NJ, http://www.amersham.com) was incubated at 33°C for 5 hours and 95°C for 10 minutes. Sybr Green real-time Q-PCR assays (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com) were carried out using 25 μl of volume containing 1.0 μl of the product from the first step and 300 nM TS and ACX primers (ACX primer; GCGCGGCTTACCCTTACCCTTACCCTAACC). Standard curve was produced for the assay using serially diluted 293 cell extracts, and all samples were run in triplicate.

Telomere Length Assay

Genomic DNA was purified from BMMSCs using a High Pure PCR Template Preparation Kit (Roche Applied Science, Mannheim, Germany, http://www.roche-applied-science.com). Telomere length was assessed by Telo TAGGG Telomere Length Assay kit (Roche Applied Science) according to the manufacturer's protocol.

Anti-Cancer Drug Treatment

BMMSCs at P1 and transformed BMMSCs at P63 seeded in 96-well plates were treated with etoposide (0–1,000 μg/ml; Alexis Biochemicals, San Diego, http://www.axxora.com, http://sigmaaldrich.com) and doxorubicin (0–30 μg/ml; Sigma) for 48 hours at 37°C. Cell viability was measured by using the Cell Counting Kit-8 from Dojindo Molecular Technologies, Inc. (Gaithersburg, MD, http://www.dojindo.com) according to the manufacturer's protocol. The inhibition effects of drugs (percentage of control) were calculated as (absorbance of non-treated cells × absorbance of treated cells)/absorbance of non-treated cells. Median effective concentration (EC50) is the drug concentration at which the cell viability becomes 50% of non-treated cells. The experiments were repeated three times. The representative plates were scanned after the reaction.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosures
  9. References

BMMSCs Were Spontaneously Transformed into Malignant Cells by Continuous Passage In Vitro

BMMSCs from adult C57Bl/6 mice were passaged continuously for more than 1 year and obtained unlimited population doublings (Fig. 1A). After additional passages, BMMSCs gradually acquired increased proliferating capacity as shown by BrdU incorporation (Fig. 1B). Initially, BMMSCs at P0 formed single-colony clusters (Fig. 1C) and showed multiple in vitro differentiation capabilities to become osteogenic (Fig. 1D), adipogenic (Fig. 1E), and chondrogenic cells (Fig. 1F) under various inductive conditions [2124]. Upon in vivo transplantation into immunocompromised mice [17], BMMSCs at P1 (Fig. 1G) generated bone and associated marrow components, indicating an in vivo osteogenic differentiation capacity (Fig. 1K). BMMSCs at early stages (P2–P5) demonstrated a slower proliferation rate and an enlarged, flattened cellular morphology, indicative of cellular senescence. However, they gradually gained rapid proliferating ability after overcoming the crisis phase with continuous regular passages [10]. At P13, BMMSCs still maintained a large cell morphology (Fig. 1H) but failed to generate any mineralized tissues in vivo (Fig. 1L), suggesting that BMMSCs have lost their osteogenic differentiation potential after they overcome the crisis phase and acquired proliferating capacity. With further continued passages (P29), the cells became smaller (Fig. 1I) and generated tumors upon in vivo transplantation (Fig. 1M). Moreover, as the passage number increased to P54, the cells showed a significantly smaller morphology (Fig. 1J) and formed tumors with a shorter latent period after transplantation compared with P29 (Fig. 1N), indicating a higher degree of malignancy.

Transformed BMMSCs Generated Fibrosarcomas and Colonized in Multiple Organs In Vivo

The tumors generated by transformed BMMSCs were composed of spindle-shaped and round atypical cells and extensively invaded the surrounding muscles and subcutaneous tissues (Fig. 2A). The blue cytoplasmic staining of tumor cells by Mallory trichrome staining suggested fibroblastic differentiation (Fig. 2B). In addition, the positive staining for vimentin (Fig. 2C) and the negative staining for cytokeratin (Fig. 2D) by immunohistochemical analysis revealed their mesenchymal origin. Taken together, these tumors generated by transformed BMMSCs were diagnosed as fibrosarcomas. BMMSCs have been shown to integrate into multiple organs after systemic administration known as “homing” [11, 25]. Therefore, we explored whether the transformed BMMSCs might similarly migrate to multiple organs when injected into immunocompromised mice through the tail vein. At 6–8 weeks after injection, transformed BMMSCs at P65 settled in several organs, including lungs (Fig. 2E), kidneys (Fig. 2F), mediastinum (Fig. 2G), vertebrae (Fig. 2H), and pericardium (data not shown) with in situ tumorigenesis. Furthermore, we investigated whether fibro-sarcoma cells isolated from tumors could regenerate tumors in vivo. Fibrosarcoma cells showed a clonogenic capability at the frequency of 2.4% from total dispersed fibrosarcoma cells (Fig. 2I). The progeny of 20 individual colonies was inoculated into secondary recipient immunocompromised mice subcutaneously. All of these 20 colonies developed fibrosarcomas within 2 weeks (Fig. 2J). This evidence implies that a small population of fibrosarcoma cells, which forms colonies in vitro, maintains the properties of forming tumors upon serial transplantation as previously reported for cancer stem cells in other tumors [36].

Transformed BMMSCs at P57 sustained the expression of CD13 and Sca-1 (Fig. 2K) but lost the expression of Thy-1, SSEA-1, and Flk-1 expressed in BMMSCs at P1 (data not shown). Hematopoietic cell markers, CD45, TER119, CD18, CD19, CD3, c-kit, and CD34, were negative in BMMSCs at both P57 and P1. These findings indicated that the transformation process changed the expression pattern of surface markers in BMMSCs, but some markers expressed in BMMSCs at P1 were still maintained.

To further clarify the features of BMMSC at the single-cell level, single colony–derived BMMSCs were isolated and their immortalization and transformation potentials were assessed. We found that five out of 100 single colony–derived BMMSCs became immortalized. All of the immortalized clones maintained in vitro osteogenic differentiation potential and four out of five clones maintained in vitro adipogenic differentiation potential at early passages (Table 1). Four out of five of these clones were eventually transformed to generate fibrosarcoma after a subcutaneous inoculation (Table 1). One of them, clone C3, showed significant bone differentiation within the newly formed fibrosarcoma (Table 1). However, when transplanted with HA/TCP, two more clones (clone D2 and H8) also generated bone after they acquired tumor formation capacity (Table 1). These findings suggest that not only multicolony-derived BMMSCs, but also single colony–derived BMMSCs, become transformed and that some cells in transformed clones partially maintain their differentiation capability, probably depending on the environment and the cell stages.

Accumulated Chromosomal Abnormalities Were Associated with the Transformation of BMMSCs

It has been postulated that genome instability is critical for tumor formation [26, 27]. We therefore examined chromosomal alterations in the transformed BMMSCs using the SKY analysis [18, 19] to find out the potential mechanism contributing to BMMSCs transformation. Fifty percent of BMMSCs showed normal chromosomes (40, XY) at P1 but displayed predominantly non-clonal numerical aberrations on remaining cells (Table 2), which may imply increased chromosome instability in murine BMMSCs. BMMSCs at P27 showed the gain of murine chromosome 2 (MMU2) in 40% of the cells (Fig. 3A; Table 2) and more than 10 double-minute chromosomes per cell involving MMU15 in 40% of the cells (Fig. 3A, yellow arrow), indicating the existence of clonal numerical and structural abnormalities. Chromosomal aneuploidy was also detected with non-clonal gains and losses of numerous chromosomes (Table 2). At P55, BMMSCs showed a greater number of double-minute chromosomes involving MMU15 in 90% of the analyzed cells (Fig. 3B, yellow arrow) and the clonal deletion of MMU14 (Fig. 3B; Table 2) with other numerical aberrations (Table 2). Chromosomal imbalances were more frequent at P55 compared with P27 (Table 2). These results suggested that accumulation of chromosomal abnormalities was a dynamic process during continuous passages, leading to the immortalization and transformation of BMMSCs. We further performed FISH analysis and identified that the double-minute chromosomes were due to an amplification of c-myc, an oncogene located on MMU15 (Fig. 3C). Western blot analysis confirmed the upregulation of c-Myc protein expression in BMMSCs at P14, P29, and P54 when compared with the expression at P1 (Fig. 3D). Next, we examined telomerase activity because c-myc has been reported to activate telomerase activity [28, 29]. The quantitative assay indicated that BMMSCs at P1 had little telomerase activity compared with the HEK293 cells used as a positive control, but after continuous passages, BMMSCs gradually gained increased telomerase activity (Fig. 3E). Non-transformed BMMSCs and transformed BMMSCs showed similar telomere lengths (Fig. 3F), suggesting that telomere length was maintained by telomerase activity during the immortalization and transformation process. These findings suggested that amplification of c-myc was, at least in part, involved in the transformation of BMMSCs through upregulated telomerase activity.

Transformation System of BMMSCs Can Be Used as a Model for Screening of Anti-Cancer Drugs

Here, we showed that transformed BMMSCs had quite different characteristics from their parental BMMSCs. We selected two anti-cancer drugs, etoposide and doxorubicin, to compare how transformed BMMSCs and their parental BMMSCs might respond to treatment. After treatment with etoposide for 48 hours, the drug concentrations to suppress the cell viability to 50% of 21.20 μg/ml non-treated cells as a control (EC50) are 185.54 ± and 1.07 ± 0.07 μg/ml (mean ± SD) for BMMSCs at P1 and P63, respectively (Fig. 4A, 4B). The inhibitory curves of etoposide to BMMSCs at P1 and P63 were significantly separated (Fig. 4B), indicating that the therapeutic effect of etoposide may be achieved without causing a severe toxic effect on normal BMMSCs at a wide range of concentrations. In contrast, the inhibitory curves of doxorubicin to BMMSCs at P1 and P63 were very close. The mean EC50 values of doxorubicin to BMMSCs at P1 and P63 were 0.49 ± 0.18 μg/ml and 0.16 ± 0.05 μg/ml, respectively (Fig. 4C, 4D), suggesting that the toxicity was relatively significant because there was no separation between “toxicity curve” (the inhibitory effect on BMMSCs at P1) and “effective curve” (the inhibitory effect on BMMSCs at P63). These findings may correlate with the fact that doxo-rubicin has adverse side effects on patients, such as the damage to cardiac muscles and the suppression of bone marrow function. Taken together, this system suggests a unique model for in vitro screening to select drugs and determine appropriate dosages to target transformed stem cells without severe side effects on normal stem cells.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosures
  9. References

This study is the first report to show that murine BMMSCs can generate fibrosarcomas in vivo after spontaneous transformation into malignant cells. We clearly demonstrated the process from stem cells to malignant cells. Despite the assumption that the origin of cancer stem cells are related to postnatal stem cells, there has been no direct evidence to demonstrate that postnatal stem cells are involved in tumorigenesis. Here, we provide direct evidence to show that murine BMMSCs could evolve into malignant tumors by spontaneous conversion of BMMSCs to transformed cells. The mechanism, by which BMMSCs are transformed into malignant cells, is correlated with accumulated chromosomal abnormalities, including structural and numerical aberrations, with increased passage numbers. Our data, including proliferation assay, histological analysis, chromosomal analysis, and quantitative telomerase activity assay, suggest that alteration of genome stability is a dynamic process, which may play a critical role to determine the fate of BMMSCs. More importantly, we showed that the double-minute chromosomes were present in the transformed BMMSCs, perhaps associated with elevated expression of c-Myc. Although many studies on human and murine cells have demonstrated a strong association between overexpression of c-Myc and tumorigenesis, this study is the first report to indicate that double-minute chromosomes were developed during the transformation process of postnatal stem cells. In addition, we found that the tumors formed by transformed BMMSCs were pathologically identified as fibrosarcomas. Transformed BMMSCs, which are a small population (5 of 100 single cell–derived clones) of adult BMMSCs, demonstrated unlimited population doublings, partially maintained differentiation capabilities in vivo, colonized in multiple organs after systemic injection, and obtained telomerase activity. These phenotypes are similar to MAPCs although transformed BMMSCs partially shared the expression profile of cell surface markers with MAPCs [11].

It is important to point out that the immortalization and transformation processes seen in murine BMMSCs were not observed in human BMMSCs under our culture conditions. To evaluate the potential of human BMMSCs to become spontaneously immortalized, we used similar strategies to find that in vitro cell culture with continuous passages led human BMMSCs to senescence without any sign of immortalization (data not shown). Human BMMSCs at P17, which had almost reached senescence, showed a very slow growth rate and demonstrated normal chromosomes (46, XX) in 14 of 15 analyzed cells, with only one revealing two non-clonal structural aberrations, deletion of chromosome 9 and 10 (data not shown). Therefore, human BMMSCs at the very latest stage of their life span are still capable of maintaining fidelity of chromosomal segregation. We tested not only human BMMSCs but other human MSCs, including dental pulp stem cells [15] and periodontal ligament stem cells [16], and did not observe spontaneous immortalization (data not shown). Genome stability in human stem cells may be maintained with complicated mechanisms that keep human cells more stable than murine cells as previously reported [30]. In the present study, murine BMMSCs at P1 showed a high degree of chromosome instability, implying that murine stem cells are susceptible to chromosomal aberration under in vitro cultivation which may associate with the clonal chromosomal abnormality such as increased gene amplification of c-myc. Very recently, however, the spontaneous transformation of human adult stem cell was reported [9]. Although the mechanisms such as increased c-myc expression and telomerase activity may contribute to the spontaneous immortalization of human MSCs [9] and murine BMMSCs, we had difficulty immortalizing human MSCs using the same continuous passage strategy as we did on murine BMMSCs. Even telomerase-transfected human BMMSCs can't be spontaneously immortalized under our experimental conditions [13]. The discrepancy between the reported finding and our observation in human BMMSCs may be due to the different experimental conditions or the origin of donors. If human BMMSCs can be spontaneously transformed to malignant cells, the clinical applications of BMMSCs for tissue engineering should be conducted very carefully and new therapeutic approaches will have to be developed to target transformed BMMSCs. Thus, understanding the mechanisms of how BMMSCs are transformed to malignant cells is critical for providing new insight into therapeutic strategies when human stem cells are applied for clinical therapies.

Spontaneous immortalization/malignant transformation in our system may mimic the process of tumorigenesis in human body. Also, the comparison of BMMSCs with transformed BMMSCs derived from their parental normal BMMSCs makes this system unique to compare the response to the drugs at different stages. We attempted to use this BMMSCs/fibrosar-coma transformation system to screen anti-cancer drugs for identifying a therapeutic index. The response of malignant cells and normal cells to drug treatment represents the effectiveness and the toxicity of the tested drug, respectively. After in vitro treatment, the EC50 of etoposide to normal BMMSCs at P1 was more than 170 folds higher than EC50 to transformed BMMSCs at P63, suggesting that etoposide has a wide therapeutic index. In contrast to etoposide, doxorubicin showed a very close EC50 between normal BMMSCs and transformed BMMSCs. These data suggest doxorubicin is more toxic than etoposide due to less specific effects on malignant cells. Indeed, this finding corresponds to the fact that doxorubicin causes worse side effects, such as the damage to cardiac muscles and the suppression of bone marrow function. It is suggested that, taken together, this system has the potential to provide a useful model for in vitro drug screening. This may help us to examine newly developed drugs and select more appropriate drugs and dosages while avoiding severe side effects.

In summary, we demonstrated that murine BMMSCs become spontaneously transformed with accumulated chromosomal abnormalities, including double-minute chromosomes, and form fibrosarcoma in vivo. Our BMMSCs/fibrosarcoma transformation system may be useful for drug screening as well as for further investigation of the mechanisms by which BMMSCs become transformed to establish stem cell–based therapies.

Table Table 1.. Characteristics of single colony–derived bone marrow–derived mesenchymal stem cell (BMMSC) clones
Thumbnail image of
Table Table 2.. The numerical chromosomal aberrations observed in 10 analyzed cells in P27 and P55
Thumbnail image of
thumbnail image

Figure Figure 1.. Murine BMMSCs were transformed into malignant cells. (A): BMMSCs were cultured under standard conditions and maintained for 1 year with more than 100 passages. They were considered immortalized cells. (B): The proliferation rate was assessed by BrdU incorporation for 4 hours. The number of BrdU-positive cells was expressed as a percentage of the total number of counted cells at the indicated passages. BMMSCs at later passages showed a higher proliferating rate in comparison with BMMSCs at earlier passages with significant statistical difference (*p < .05, n = 10). (C–F): BMMSCs at P1 maintain multiple differentiation capacity. Single colonies were formed after BMMSCs were plated at low density (1–2 × 106 cells/T-25 flask) and cultured with regular culture medium for 10 days. Single colonies derived from BMMSCs were stained with 0.1% Tulidine blue (arrows, [C]). BMMSCs were cultured with 100 μM L-ascorbate-2-phosphate, 10 nM dexamethasone, and 2 mM β-glycerophosphate for 4 weeks. Alizarin red staining showed mineralized nodule formation (arrows, [D]; original magnification: ×100). Cultured BMMSCs formed Oil Red O-positive lipid clusters after 3 weeks of induction in the presence of 0.5 μM isobutylmethylxanthine, 0.5 μM hydrocortisone, and 60 μM indomethacin (arrows; [E], original magnification: ×400). When cultured with TGF-β (2 ng/ml), BMMSCs differentiated into chondrocytes with positive immunostaining of anti–type II collagen antibody ([F]; original magnification: ×400). (G–N): Immortalization and transformation of BMMSCs. BMMSCs at P1 (G) generated bones (B) and associated hematopoietic elements (BM) upon transplantation into immunocompromised mice subcutaneously using HA/TCP (HA) (K). BMMSCs at P13 showed enlarged shape (H) and failed to generate any mineralized tissue upon in vivo transplantation. Only fibrous tissue (F) was presented around HA/TCP (HA) (L). BMMSCs at P29 became smaller (I) and showed tumor formation when transplanted into immunocompromised mice (T, [M]). BMMSCs at P54 showed a significantly smaller shape (J) and generated tumors with higher cell density and shorter latent periods compared with BMMSC at P29 when transplanted into immunocompromised mice (T, [N]). (G–J): Toluidine blue (0.1%) staining. Original magnification: ×200. (K–N): Hematoxylin and eosin staining. Original magnification: ×200. Abbreviations: BMMSC, bone marrow–derived mesenchymal stem cell; BrdU, bromode-oxyuridine; HA/TCP, hydroxyapatite/tricalcium phosphate; P, passage; TGF, transforming growth factor.

Download figure to PowerPoint

thumbnail image

Figure Figure 2.. Transformed BMMSCs generated fibrosarcomas and colonized in multiple organs. (A–D): Histology of tumors generated by BMMSCs at P67. Tumors were composed of spindle and round atypical cells and invasive to surrounding muscles (arrow). Mitosis of nuclei was also observed (yellow arrow) (A). Mallory trichrome staining showed fibroblastic differentiation of tumor cells (blue staining). Red color staining (arrows) indicated the muscles (B). Immunohistochemical staining using anti-vimentin antibody indicated that the tumor cells originated from mesenchymal cells (arrow heads) (C). Negative staining using anti-cytokeratin antibody (star) excluded the possibility of epithelial cell origin of the tumor cells. Skin epidermal tissues were positive for cytokeratin (arrows) (D). Original magnification: ×200 (A–D). (E–H): Hematoxylin and eosin staining of transformed BMMSCs at P65 showed tumor formation (T) in alveolar spaces of lungs (L, [E]), kidney (F), mediastinum (G), and vertebrae (H) after i.v. injection into mice. Gromelurus (G), renal tubules (Tu), trachea (Tr), tracheal cartilages (TC), and bone (B). Original magnification: ×200 (E–H). (I): Single cells from dispersed fibrosarcomas formed colonies in vitro when 1 × 106 cells were inoculated. (J): Single cells from dispersed fibrosarcomas re-formed tumors in the second recipients when 1 × 106 cells were inoculated. (K): FACS analysis indicated that transformed BMMSCs were positive for CD13 and Sca-1 but negative for Thy-1, SSEA-1, Flk-1, and hematopoietic markers CD45, TER119, CD18, CD19, CD3, c-kit, and CD34. Abbreviations: BMMSC, bone marrow–derived mesenchymal stem cell; FACS, fluorescence-activated cell sorting; P, passage.

Download figure to PowerPoint

thumbnail image

Figure Figure 3.. Chromosomal abnormalities in transformed BMMSCs. (A): SKY analysis of BMMSCs at P27 showed a gain of chromosome MMU2 and double-minute chromosomes involving MMU15 (yellow arrow). (B): SKY analysis of BMMSCs at P55 indicated a significant increase in the number of double minutes involving MMU15 (yellow arrow). Deletion in MMU14 and a loss of MMU8 were noted. (C): FISH analysis revealed an amplification of c-myc (yellow arrows) located on MMU15 (white arrows). (D): Western blot analysis confirmed the upregulated expression of c-Myc in BMMSCs at P14, P29, and P54 when compared with BMMSCs at P1. (E): BMMSCs gradually gained significant telomerase activity after continuous passages (*p < .05, n = 3). HEK293 cells were used as a positive control (c). (F): Telomere length of BMMSCs was not significantly changed during the continuous passages from P1 to P90. Abbreviations: BMMSC, bone marrow–derived mesenchymal stem cell; FISH, fluorescent in situ hybridization; P, passage; SKY, spectral karyotyping.

Download figure to PowerPoint

thumbnail image

Figure Figure 4.. Normal and transformed bone marrow–derived mesenchymal stem cells (BMMSCs) responded differently to anti-cancer drugs. Cell viability of BMMSCs at passage 1 (P1) and P63 was evaluated after anti-cancer drug treatment. (A, B): Treatment with etoposide (0–1,000 μg/ml) for 48 hours. BMMSCs at P1 and P63 seeded in 96-well plates were treated with the indicated concentration of etoposide. The representative plate demonstrated gradual changing of color with increased concentrations of etoposide. The change of color was more drastic in P63 than in P1 (A). The inhibitory curves to BMMSCs at P1 (black solid line) and P63 (blue dotted line) were significantly separated, indicative of a wide therapeutic index (B). (C, D): Treatment with doxorubicin (0–30 μg/ml) for 48 hours. The color of plate was changed similarly at P1 and P63 (C). There was no separation between “toxicity curve” (the inhibitory effect on BMMSCs at P1) and “effective curve” (the inhibitory effect on BMMSCs at P63) (D).

Download figure to PowerPoint

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosures
  9. References

We thank Nicole McNeil (Genetics Branch, Center for Cancer Research, National Cancer Institute) for preparation of the c-myc probe for FISH analysis, Linda Barenboim Stapleton (Genetics Branch) for preparation of our mouse SKY probes, and Sivio Gutkind (Oral and Pharyngeal Cancer Branch/National Institute of Dental and Craniofacial Research) for critical reading and review of this manuscript. This work was supported by the intramural program of National Institute of Dental and Craniofacial Research, National Institutes of Health, Department of Health and Human Services.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosures
  9. References