| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print September 4, 2008
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Pathology,* Keio University School of Medicine, Tokyo; the Institute for Biological Sciences, Science University of Tokyo, Noda City; and the National Center for Child Health and Development, Tokyo, Japan
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
The origin of the Schwannian stromal cell is controversial, and the relationship between Schwannian stromal cells and neuroblastic cells has not yet been clarified. One hypothesis is that both the Schwannian stromal cells and the neuroblastic cells arise from pluripotent neoplastic cells originating from the neural crest. Mora and colleagues5 showed that Schwann cells have the same genetic features as neuroblastic cells. An alternative theory, proposed by Ambros and colleagues6,7 and Katsetos and colleagues8 stated that the Schwann cells in neuroblastic tumors are likely to be reactive in nature and may have been recruited from nonneoplastic tissues surrounding the tumor cells. Ambros and colleagues9 also reported that the Schwann cells induced NB cell neural differentiation and apoptosis in vitro. As yet, there is little research on the interaction between NB cells and Schwann cells. We have examined whether NB cells can induce the differentiation of bone marrow stromal cells or mesenchymal stem cells into Schwann cells.
Tumor stroma is indispensable for tumor development. Cell growth, apoptosis, motility, and the invasiveness of the cancer cells are all critically influenced by stroma-tumor interactions. Human bone marrow stromal cells (hBMSCs) are multipotential and can differentiate into tissues such as endothelium, muscle, myofibroblast, and neural cells.10 Recently, some reports showed that bone marrow cells constitute part of tumor vessels and fibrous stroma.11-13 Animal models have been developed for in vivo research on tumor stroma,13,14 however, these models, using rodent stroma, may not be appropriate for examining the biological function of human tumor stroma. Here, we examined whether hBMSCs are capable of forming tumor stroma by inoculating hBMSCs into xenografted tumors in NOD/SCID mice.
The Schwannian stroma in NB tumors is a very specific stroma compared with other tumor stroma concerning differentiation and prognosis of NB tumors. We examined whether nonneoplastic cells (hBMSCs or human fibroblasts) could differentiate into Schwann cells in NB tumors. We conducted in vivo experiments in which hBMSCs or human fibroblasts were inoculated into xenotransplanted NBs in NOD/SCID mice and the tumor stroma analyzed using immunohistochemistry. The ability of hBMSCs or human fibroblasts to differentiate into Schwann cells through co-culture with NB cells was also examined.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Bone marrow was aspirated from the costa or sternum of patients with lung neoplasms or pectus excavatum, and suspended in Iscove’s modified Dulbecco’s medium. Mononucleated cells, including stromal cells, were isolated by centrifugation for 30 minutes at 400 x g (room temperature) using Lymphosepar (IBL, Fujioka, Japan), washed twice in Iscove’s modified Dulbecco’s medium, and seeded in a 10-cm dish containing Iscove’s modified Dulbecco’s medium with 10–6 mol/L hydrocortisone, 10–4 mol/L 2-mercaptoethanol (Sigma-Aldrich, Sheinheim, Germany), 100 IU/ml penicillin, 100 µg/ml streptomycin, 15% heat-inactivated fetal bovine serum, and 12.5% heat-inactivated horse serum (Moregate BioTech, Bulimba, Australia). The hBMSCs adhered 1 day after bone marrow isolation and were expanded for 3 to 4 days. After the nonadherent cells were decanted, the adherent stromal layer was trypsinized and transferred to two 10-cm plates. Flow cytometry showed that the hBMSCs were CD13+, CD29+, CD44+, CD106+, CD166+, CD14–, CD31–, CD34–, CD43–, CD45–, CD71–, CD105–, and S100beta–. All human material was used with the approval (approval numbers 13-1 and 12-1) of the Ethics Committee of Keio University School of Medicine and with the signed, informed consent of patients.
Cell Lines and Culture Conditions
Human mesenchymal stem cells (hMSCs; Osiris Therapeutics Inc., Columbia, MD) were cultured using the human mesenchymal stem cell bullet kit (BioWhittaker, Inc., Walkersville, MD). HMSCs have been reported to differentiate into osteogenic, chondrogenic and adipogenic lineages. Flow cytometry showed that the hMSCs were CD29+, CD44+, CD105+, CD106+, CD166+, CD14–, CD31–, CD34–, CD43–, CD45–, and S100beta–.15 The NB cell lines SK-N-DZ and SK-N-AS (CRL-2149 and CRL-2137; American Type Culture Collection, Rockville, MD) were cultured in Dulbecco’s modified Eagle’s medium with 100 IU/ml penicillin, 100 µg/ml streptomycin, and 10% heat-inactivated fetal bovine serum (JRH Biosciences, Lenexa, KS).SK-N-DZ cells are derived from a poorly differentiated embryonal NB and exhibit MYCN amplification (30 times). Retinoic acid induces differentiation of the SK-N-DZ cells.16 SK-N-AS cells are derived from a poorly differentiated embryonal NB and do not exhibit MYCN amplification. Retinoic acid partially inhibits proliferation, and SK-N-AS cells fail to differentiate after retinoic acid treatment.17 The Ewing’s sarcoma cell line RD-ES (HTB-116, American Type Culture Collection) was cultured in RPMI 1640 medium with 100 IU/ml penicillin, 100 µg/ml streptomycin, and 10% heat-inactivated fetal bovine serum. The human fibroblast cell lines MRC-5 (CCL-171, American Type Culture Collection), KMS-6 (JCRB0432; Health Science Research Resources Bank, Osaka, Japan), and HE-1 (IFO050297, Health Science Research Resources Bank) were cultured in minimum essential medium Eagle with 100 IU/ml penicillin, 100 µg/ml streptomycin, and 10% heat-inactivated fetal bovine serum. All cells, including the primary hBMSCs, were grown at 37°C in the presence of 95% humidity and 5% CO2.
Retroviral Vector and Its Producer Cells
The enhanced green fluorescent protein (EGFP) retroviral vector and the amphotropic producer cells, PG13/MSGFP, were provided by Dr. Robert G. Hawley of the American Red Cross, Washington, DC. The retroviral vector was constructed to contain only the 5' and 3' long terminal repeat promoter (LTR) and EGFP gene, and not the neomycin gene. The PG13/MSGFP producer cells were cultured in Dulbecco’s modified Eagle’s medium containing 4.5 g/L glucose plus 10% fetal bovine serum. The supernatant of the PG13/MSGFP cells was used for the retroviral transductions of the hBMSCs and hMSCs.
Transduction of Bone Marrow Stromal Cells
HBMSCs or hMSCs were seeded in a 10-cm dish. When the cells reached 80% confluence, the supernatant was discarded and 5 ml of a viral solution containing 2.5 ml of the supernatant of amphotropic producer cells (PG13/MSGFP), an equal volume of media for the primary hBMSCs or hMSCs, and 4 µg/ml of polybrene (Sigma, St. Louis, MO), were added. The cells were then incubated for 2 hours. Four rounds of transductions were performed, and 5 ml of fresh media was added after the last round of transduction. After 24 hours, the supernatant of the stromal cells was changed with fresh media. The cells were transduced with EGFP-retroviral vector and allowed to expand for1 week; the presence of EGFP-positive cells was detected using flow cytometry.
Transplantation of NB Cells and hBMSCs or hMSCs in NOD/SCID Mice
NOD/SCID mice were maintained under specific, pathogen-free conditions and used at 6 to 10 weeks of age. All experimental procedures and protocols were approved by the Animal Ethics Committee of the Keio University School of Medicine. NB cells, SK-N-DZ or SK-N-AS, were suspended (1 x 107 cells) in 100 µl of Dulbecco’s modified Eagle’s medium and transplanted subcutaneously into the NOD/SCID mice. Two weeks after transplantation, hBMSCs or hMSCs containing the EGFP-retroviral vector (3 x 106 cells) in 50 µl of Iscove’s modified Dulbecco’s medium were inoculated into the subcutaneous NB tumors of the NOD/SCID mice. NB tumors inoculated with media only served as negative controls. The NB tumors inoculated with or without hBMSCs or hMSCs were excised 9 days after inoculation and processed for histological examination. The process was repeated using Ewing’s sarcoma cells RD-ES (1 x 107 cells) transplanted subcutaneously into NOD/SCID mice and the subcutaneous Ewing’s sarcoma inoculated with hBMSCs.
Examination of EGFP-Positive Cells
The subcutaneous NB tumors from the NOD/SCID mice were cut in half and fixed in 4% paraformaldehyde for 24 hours. One half of each tumor was embedded in paraffin for histological examination and immunohistochemistry. The other half was frozen in Tissue-Tec OCT (Sakura, Tokyo, Japan) in isopentane with dry ice. Cryosections (6 to 8 µm) were fixed in 4% paraformaldehyde for 5 minutes. EGFP-fluorescence was observed using a confocal laser microscope (LSM510; Carl Zeiss, Tokyo, Japan) with a 488-nm laser and a 505- to 530-nm bandpass filter.
Immunohistochemistry and Immunocytochemistry
The paraffin-embedded xenografted NB tumor sections were stained with a mouse monoclonal GFP antibody (1:2000; MBL, Nagoya, Japan), using an indirect immunohistochemical technique to identify the EGFP-positive cells. The Schwann cells in the NBs were stained with an anti-S100 polyclonal antibody (1:1000; DakoCytomation, Glostrup, Denmark), a S100beta monoclonal antibody (1:100; Santa Cruz Biotechnology, Santa Cruz, CA), and a human NGFRp75 monoclonal antibody (1:200, Santa Cruz Biotechnology). The stromal components of the NBs were stained with an 
-smooth muscle actin monoclonal antibody (1:200, DakoCytomation). Immunohistochemistry for Egr-2 was performed on cryosections of NB tumors from patients with an Egr-2 rabbit polyclonal antibody (1:400, Santa Cruz Biotechnology). Antigen retrieval was performed with 10 mmol/L citrate buffer (pH 6.0) for S100beta and -smooth muscle actin, heated in a boiling steamer for 10 minutes, and cooled down to room temperature for 20 minutes. For S100, GFP, and human NGFRp75, antigen retrieval was performed with 0.1% trypsin (Difco Laboratories, Detroit, MI) at room temperature for 30 minutes. All sections were blocked with 2.5% normal horse serum for 30 minutes at room temperature and incubated with the primary antibodies in a humidity chamber overnight at 4°C, and visualized with diaminobenzidine (ImmPress Reagen kit; Vector Laboratories, Burlingame, CA). The sections were immunohistochemically dual-stained for EGFP-positive and S100-positive cells using alkaline phosphatase with Fast blue BB (for the mouse anti-GFP antibody), and horseradish peroxidase with diaminobenzidine (for the anti-S100 antibody). The sections were also immunohistochemically dual-stained for EGFP-positive and -smooth muscle actin-positive cells using alkaline phosphatase with Fast blue BB (for the mouse anti-GFP antibody), and horseradish peroxidase with diaminobenzidine (for the -smooth muscle actin antibody).
Immunofluorescence Observations
Cryosections of xenografted NB tumor were processed for dual-immunofluorescence staining of EGFP-positive/S100-positive and EGFP-positive/Egr2-positive cells. EGFP-positive cells were stained with a mouse monoclonal GFP antibody (1:2000) and a secondary goat anti-mouse Cy3-conjugated antibody (1:100; Jackson ImmunoResearch Laboratories, West Grove, PA). S100-positive cells were stained with a S100 polyclonal antibody (1:1000) and a secondary goat anti-rabbit Cy5-conjugated antibody (1:100, Jackson ImmunoResearch Laboratories). Egr-2-positive cells were stained with an Egr-2 rabbit polyclonal antibody (1:200) and a secondary goat anti-rabbit Cy5-conjugated antibody (1:100). All sections were blocked with 5% normal goat serum before incubation with primary antibodies. Digital images were obtained using a Zeiss confocal laser microscope (LSM510) with a 543-, 633-nm laser and a 560-, 650-nm bandpass filter.
Co-Culture of hBMSCs or hMSCs and NB Cells
The hBMSCs, hMSCs, or human fibroblasts were trypsinized and placed on six-well plates for transwells (pore size, 0.4 µm; Corning Costar, Cambridge, MA) at a density of 2 x 104 stromal cells/well. The transwells were inserted and 1 x 105 NB cells, either SK-N-DZ or SK-N-AS in a suspension of Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, were added to each transwell. The hBMSCs were also co-cultured in contact with NB cells, in the same wells of the six-well plates, at a density of 1 x 104 stromal cells, and 1 x 104 NB cells/well. Every 3 days the cultures were refreshed with an equal volume of Dulbecco’s modified Eagle’s medium for the NB cells and media for the stromal cells or fibroblasts. A proportion of the wells were treated with 1 µg/ml of transforming growth factor-β (TGF-β), 5 µg/ml glial-derived neurotrophic factor (GDNF), 25 µg/ml sensory and mortar neuron-derived factor (SMDF), 20 µg/ml 3,3,5-triiodo-L-thyonine sodium salt (T3), 1 mg/ml L-throxine sodium salt pentahydrate T4, 0.5 mmol/L forskolin, and 10 µg/ml of nerve growth factor (NGF) to examine the induction of Schwann cell differentiation by the hBMSCs or fibroblasts.18-22 Before co-culture, hBMSCs were labeled using a fluorescent cell linker kit (PKH26-GL, PKH67-GL; Sigma-Aldrich Co., Vienna, Austria) following the instructions of the manufacturer.
Quantification of S100beta Induction
After1 week of culture, the hBMSCs labeled with the green fluorescent cell linker kit (PKH67-GL) were stained with aS100beta monoclonal antibody and a secondary goat anti-mouse Cy3-conjugated antibody, and mounted with mounting medium with 4,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories). Digital images were obtained using a confocal laser microscope (LSM510; Carl Zeiss) with a 405-, 488-, 543-nm laser and a 420-, 505-, 560-nm bandpass filter. S100beta-positive hBMSCs, after co-culture with NB cells, were quantified by the ArrayScan VT1 system (Cellomics, Pittsburgh, PA). The hBMSCs labeled with the red fluorescent cell linker kit (PKJ26-GL) were stained with aS100beta monoclonal antibody and a secondary rabbit anti-mouse fluorescein isothiocyanate (FITC) antibody (DakoCytomation). The nuclei were stained with Hoechst 33342. Images were captured by the Cellomics ArrayScan VT1 from a total of 50 fields (1 field = 660 µm x 660 µm) per well using three separate fluorescent filters at x10 objective magnification. Channel 1 (blue, 365 nm x 50 nm bandwidth filter) excites Hoechst 33342, channel 2 (green, 475 nm x 40 nm bandwidth filter) excites objects stained with FITC-labeled S100beta, and channel 3 (red, 575 nm x 25 nm bandwidth filter) excites PKH26. Morphometry was analyzed using the Target Activation BioApplication assay protocol (Cellomics). The Target Activation Bioapplication reports the fluorescent intensities of each object in the field, based on user-defined fluorescent intensity thresholds set in the blue, green, and red images. This application also reports the number of cells that were defined as an object/cell with both green and red fluorescent intensity values, inside a range defined by the user. The percentage of S100beta-FITC-positive hBMSCs from the total of hBMSCs labeled with red fluorescent linker PKH26-GL, is reported.
Flow Cytometry Analysis of NB Cells after Co-Culture with hBMSCs or hMSCs
The SK-N-DZ cells and SK-N-AS cells were stained with a ssDNA rabbit polyclonal antibody (1:200, DakoCytomation) and a secondary swine anti-rabbit FITC-conjugated antibody (1:100, DakoCytomation), after co-culture with hBMSCs or hMSCs, to identify apoptosis of the NB cells. The NB cells were also stained with a MIB-1 mouse monoclonal antibody (1:100, DakoCytomation) and a secondary rabbit anti-mouse FITC-conjugated antibody (1:100, DakoCytomation) after co-culture to analyze cell proliferation.
Immunoblotting
HBMSCs or fibroblasts in six-well plates were washed with phosphate-buffered saline, dissolved in RIPA buffer containing 1 mmol/L ethylenediaminetetraacetic acid, 0.1% sodium dodecyl sulfate, 1% IGEPAL, 150 mmol/L NaCl, 10 mmol/L Tris, and 10 µg/ml aprotinin (Sigma), and centrifuged (15,000 rpm, 15 minutes, 4°C). The supernatant of the cell lysates was transferred to a new tube. The protein concentrations were determined using a standard bovine serum albumin curve (DC Protein Assay Kit II; Bio-Rad, Hercules, CA). Equal amounts of protein were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis using NuPAGE 4 to 12% Bis-Tris gel (Invitrogen, Carlsbad, CA) and blotted onto a polyvinylidene difluoride membrane. The membrane was blocked with 5% skim milk (Becton Dickinson, Franklin Lakes, NJ) in Tris-buffered saline containing Tween 20 (Sigma-Aldrich) (TBS-T). Egr-2 or HSC70 was detected using enhanced chemiluminescence (ECL; Amersham, Buckinghamshire, UK) and primary antibodies against human Egr-2 (rabbit polyclonal IgG, diluted 1:1000; Santa Cruz Biotechnology) or human HSC70 (mouse monoclonal IgG, diluted 1:1000; Santa Cruz Biotechnology) and horseradish peroxidase-conjugated secondary antibody (goat anti-rabbit IgG-HRP, diluted 1:1000; Santa Cruz Biotechnology).
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The primary hBMSCs were spindle-shaped and grew constantly (Figure 1A1)
. A proportion of the hBMSCs differentiated into endothelial cells, as shown by the uptake of Dil-Ac-LDL on day 10 (Figure 1A2)
. On day 30, numerous lipid droplets in the cytoplasm of the hBMSCs and red-colored droplets were observed when the cells were examined using oil red O stain (Figure 1A, 3 and 4)
. The primary hBMSCs were harvested from human bone marrow and transduced with an EGFP-retroviral vector. More than 70% of the primary hBMSCs were EGFP-positive on day 9 after transduction (Figure 1B)
. Sixty to seventy percent of the hMSC cell line were EGFP-positive after the retroviral transduction (Supplemental Figure S1, see http://ajp.amjpathol.org). The flow cytometry results of S100beta– in hBMSCs and hMSCs did not change after transduction.
|
The NB cells, SK-N-DZ or SK-N-AS, were transplanted subcutaneously into NOD/SCID mice. After 2 weeks, tumors with diameters of 5 mm were observed in the subcutaneous tissues. Primary hBMSCs transduced with the EGFP-retroviral vector were inoculated into the NB tumors. Figure 2, a and b
, shows NB tumors (SK-N-DZ and SK-N-AS, respectively) with stromal components exhibiting an irregular net-like formation. The distribution of EGFP-positive cells was diffuse among the stromal components of the NB tumors (Figure 2, c and d
; SK-N-DZ and SK-N-AS, respectively). Green-fluorescent cells in the frozen NB tumor sections (SK-N-DZ and SK-N-AS) were observed using a confocal laser microscope (Figure 2, e and f
, respectively). These results show that the hBMSCs inoculated into the NB tumors survived and constituted a portion of the tumor stroma. Reconstitution of the tumor stroma using hMSCs was also seen (Supplemental Figure S2, see http://ajp.amjpathol.org).
|
Schwann cells in the xenografted NB tumors were detected using a S100 polyclonal antibody, which reacted with both human and murine S100 proteins, similar to the method used for the pathological diagnosis of NB tumors. S100-positive cells were not observed in the NB tumors (SK-N-DZ) that were not inoculated with hBMSCs (Figure 3h)
. We examined the expression of S100beta, Schwann cell-specific molecules, and human NGFRp75 using immunostaining. S100beta- and p75-positive cells were detected in the xenografted NB tumors inoculated with hBMSCs (Figure 3, b and c)
, but were not detected in the NB tumors without the inoculation of hBMSCs (Supplemental Figure S3A, i and j; see http://ajp.amjpathol.org). Tumor stroma with S100-, S100beta-, and p75-positive cells was also observed in SK-N-AS tumors injected with hBMSCs (Supplemental Figure S3A, b–d; see http://ajp.amjpathol.org). These results show that Schwannian stroma derived from murine cells was not induced in the human NB tumors. On the other hand, S100/S100beta-positive cells were observed in the irregular net-like stroma of NB tumors (SK-N-DZ) inoculated with hBMSCs (Figure 3, a and b)
. These S100/S100beta-positive cells were spindle-shaped and found among the stromal components (Figure 3, a and b)
. We examined whether these S100-positive cells originated from the inoculated EGFP-hBMSCs using immunohistochemical or immunofluorescent double-staining for S100 and EGFP. We found the S100-positive cells were similarly almost always positive for EGFP (Figure 3, d–g)
. The induction of S100-positive stromal cells in SK-N-AS cell-derived NB tumors inoculated with hBMSCs was also observed (Supplemental Figure S3A, e–h, see http://ajp.amjpathol.org). These results suggest that the S100-positive Schwann cells were derived from the inoculated hBMSCs.
|
. The induction of S100/S100beta/p75-positive stromal cells in NB tumors inoculated with hMSCs was also seen (data not shown). These results suggest that hBMSCs or hMSCs constitute human stromal components in vivo and differentiate into S100-positive Schwannian stroma only when in NBs similar to clinical NB tumors.
We also examined the expression of Egr-2, a transcription factor expressed during the early development of Schwann cells, in NB tumors.23
In differentiating NB tumors from patients, a small number of Egr2-positive cells were observed in the stroma of differentiating NB tumors (Figure 4, a and b)
. In xenografted NB tumors, Egr-2-positive cells were detected in the stroma, and most of these Egr-2-positive cells were also EGFP-positive (Figure 4, c–f
; and Supplemental Figure S3B, see http://ajp.amjpathol.org). These results support the proposal that inoculated hBMSCs differentiate into Schwann cells.
|
Serial observations of NB tumors (SK-N-DZ) inoculated with hBMSCs showed that EGFP-positive hBMSCs formed blood vessel-like structures that contained red blood cells in their lumen (Figure 5a)
. We examined whether the EGFP-positive cells expressed certain endothelial markers (CD31, CD34, or CD105), or pericytic markers (-smooth muscle actin), by immunohistochemically double-staining the cells with anti-GFP and anti-CD31, CD34, CD105, or -smooth muscle actin. The EGFP-positive cells that formed vessel-like structures expressed only -smooth muscle actin, and not CD31, CD34, or CD105 (Figure 5b)
. Thus, the EGFP-hBMSCs seem to function as pericytes, forming the lining of vessel walls in NB tumors. We observed a quantitative alteration in the stromal elements induced by the inoculation of hBMSCs into NB tumors. The stromal network of NB tumors inoculated with hBMSCs, which were immunostained using an -smooth muscle actin antibody, was more widespread than that of NB tumors not inoculated with hBMSCs (Figure 5, c and d)
. Anti--smooth muscle actin reacts with both murine and human -smooth muscle actin, so actin-positive cells (Figure 5c)
were present in the murine stromal elements of the xenografted human NB tumors not inoculated with hBMSCs. However, many more -smooth muscle actin-positive cells were seen in NB tumors inoculated with hBMSCs, and the area of the -smooth muscle actin-positive cells was relatively more widespread because the -smooth muscle actin-positive cells were large and more frequent (Figure 5d)
. The -smooth muscle actin-positive cells were almost always immunohistochemically positive for GFP (Supplemental Figure S4, see http://ajp.amjpathol.org). These results indicate that the hBMSCs inoculated into NB tumors survived as a stromal component and were incorporated into blood vessel structures as pericytes and formed myofibroblasts.
|
We examined whether NB cells could induce the differentiation of hBMSCs or hMSCs into Schwann cells in vitro. We cultured hBMSCs or hMSCs with human NB cells (SK-N-DZ or SK-N-AS). We cultured stromal cells and NB cells with cell-to-cell contact, and stromal cells and NB cells using the transwell plates with micropores (0.4 µm), without cell-to-cell contact. The expression of S100beta, a Schwann cell-specific marker, was observed in cultures under both conditions. On the other hand, S100beta-positive cells were not observed in hBMSCs cultured without NB cells (Figure 6Aa)
. Quantitative analysis using the ArrayScan VT1 system revealed significantly more induction of S100beta FITC fluorescence in hBMSCs when co-cultured with NB cells, both when there was or was not cell-to-cell contact with NB cells, compared with that in hBMSCs not co-cultured with NB cells. These results show that NB cells can induce the differentiation of stromal cells to Schwann cells in vitro. We found that a cytokine cocktail containing growth factors such as TGF-β, NGF, GDNF, SMDF, T3, T4, and forskolin,18-22
did not affect the expression of S100beta in the presence or absence of NB cells (Figure 6)
.
|
We examined whether Egr-2, a transcription factor expressed during the early development of Schwann cells, was induced during the differentiation of hBMSCs or hMSCs to Schwann cells in vitro. The expression of Egr-2 was significantly elevated in hBMSCs or hMSCs co-cultured with NB cells, compared with hBMSCs or hMSCs not co-cultured with NB cells (Figure 7)
. Egr-2 expression was induced in stromal cells in co-cultures with SK-N-DZ or SK-N-AS cells. The expression of Egr-2 in the absence of NB cells was not affected by the addition of a cytokine cocktail containing the growth factors, TGF-β, NGF, GDNF, SMDF, T3, T4, and forskolin (Figure 7A)
. Furthermore, hMSCs were co-cultured with NB cells in the presence of TGF-β (10 µg/ml) because TGF-β was suspected to be involved in the tumor-stromal interactions that cause hMSCs to differentiate into Schwann cells.24
However, the addition of TGF-β did not influence the expression of Egr-2 in the co-cultures (Figure 7B)
. The result of hBMSCs co-cultured with SK-N-AS or SK-N-DZ in the presence of TGF-β (or cytokine cocktail) was similar to that of hMSCs.
|
To test whether the differentiation of Schwann cells was specific to certain cell types, we examined whether human fibroblasts, MRC-5 (originating from human fetal lung tissue) and KMS-6 (originating from a whole human embryo), differentiated into Schwann cells in the presence of NB cells. MRC-5 or KMS-6 cells were co-cultured with NB cells (SK-N-DZ) using transwell plates and the expression of Egr-2 in the MRC-5 or KMS-6 cells was analyzed using immunoblotting. There was a slight increase in the expression of Egr-2 in the MRC-5 cells co-cultured with the SK-N-DZ cells, whereas there was no significant change in Egr-2 expression in the KMS-6 and HE-1 cells (Figure 8)
. These results suggest that human fibroblasts are a phenotypically heterogeneous cell population.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Recently, anti-tumor therapy using mesenchymal stem cells has been tested in malignant tumors such as glioma, colon cancer, and Kaposi’s sarcoma, based on the hypothesis that mesenchymal stem cells could constitute the tumor stroma.25-38 These studies showed that tumor cell proliferation or tumor growth was impaired by the mesenchymal stem cells carrying therapeutic genes or molecules.
We have developed an animal model for human tumor-stroma research using human stromal cells derived from bone marrow. We found that the primary stromal cells or mesenchymal stem cells from human bone marrow may comprise the tumor stroma of xenografted human tumor, such as NB or Ewing’s sarcoma, in murine subcutaneous tissues. This experimental model may be useful to study in vivo interactions between neoplastic cells and stromal cells and for the advancement of tumor-stroma biology through histology, immunohistochemistry, or in situ hybridization techniques. Tumor stroma composed of hBMSCs or hMSCs may be functional in human xenografted tumors, because the inoculated hBMSCs or hMSCs developed into NB-specific Schwannian stroma and formed the lining of microvessels.
The size or weight of the xenografted NB tumors inoculated with hBMSCs and those without hBMSCs did not differ (data not shown). In the co-culture experiments, both apoptosis and proliferation of NB cells were augmented by their association with hBMSCs. Using flow cytometry, we observed an increased number of ssDNA-positive NB cells and an increased number of MIB-1 (Ki-67)-positive NB cells after co-culture with hBMSCs (Supplemental Figure S5, see http://ajp.amjpathol.org). These results show that the inoculation of hBMSCs into NB tumors may not have influenced tumor growth despite the accompanying Schwannian stroma formations.
The presence and volume of the Schwannian stroma is closely related with the prognosis of patients with NB. Schwannian stroma-rich/stroma-dominant NB tumors are associated with a more favorable outcome.2 The development of Schwannian stroma is thought to be associated with the differentiation of NBs.2 Ambros and colleagues9 reported that neural differentiation was enhanced and the apoptotic rate was increased in NB cells co-cultivated with normal Schwann cells in vitro. Liu and colleagues39 also reported that Schwann cells influenced proliferation, differentiation, and apoptosis of NB cells and angiogenesis in NB, using a murine sciatic nerve-engrafted NB xenograft model. Using cytogenetic procedures, Mora and colleagues5 reported that Schwannian cells are derived from neoplastic NB cells. However, Ambros and colleagues7 stated that the Schwann cells in NB tumors originated from nonneoplastic cells because the neoplastic NB cells showed aberrations in the number of chromosomes and the Schwann cells contained a normal number of chromosomes, when examined using fluorescence in situhybridization. Liu and colleagues,39 using the sciatic nerve-engrafted NB xenograft model, also concluded that the Schwann cells infiltrating into xenografted NB tumors were derived from the murine peripheral nerve tissues around the NB tumors, but not from the inoculated human NB cells. We only observed Schwann cells in NB tumors containing hBMSCs or hMSCs, which is compatible with the theory that the Schwannian stroma originates from nonneoplastic cells. In addition, we did not observe Schwann cells in xenografted human NB tumors that were not inoculated with hBMSCs or hMSCs. This may indicate that the factors produced by human NB cells cannot cause murine stromal cells to develop into stromal elements, such as microvessels. These observations do not necessarily mean that Schwannian stroma in patients with NBs originate from bone marrow cells. Cells such as fibroblasts that are recruited from neighboring tissues may constitute the tumor stroma because we observed that human fibroblasts (MRC-5 cells) differentiated into Schwann cells when co-cultured with SK-N-DZ cells.
Both of the two NB cell lines, SK-N-DZ and SK-N-AS, could induce Schwannian cell differentiation of hBMSCs in vivo and in vitro, but the differentiation of induction ability was observed. Two pathologists evaluated independently the volume of S100/S100beta-positive stroma in SK-N-DZ and SK-N-AS tumors by microscope. The evaluation showed that S100/S100beta-positive stroma in SK-N-DZ tumors with human stromal cells was more than that in SK-N-AS tumors (Supplemental Figure S3, a and b; see http://ajp.amjpathol.org). Furthermore, S100beta induction in vitro in hBMSCs by co-culture with SK-N-DZ cells was more effective than by SK-N-AS cells (Figure 6B)
. SK-N-DZ cells exhibit MYCN amplification of 30 times and SK-N-AS did not exhibit MYCN amplification. Retinoic acid has been reported to induce cellular differentiation in SK-N-DZ cells.16
By contrast, retinoic acid is also reported to partially inhibit the proliferation of SK-N-AS cells and not to induce cellular differentiation.17
The mechanism of differentiation and reactivity to induction differs between SK-N-DZ cells and SK-N-AS cells.
The hBMSCs and hMSCs differentiated into Schwann cells when they were co-cultured with NB cells using transwell plates. Humoral factors have been suggested to induce BMSCs and MSCs to differentiate into Schwann cells. Several humoral factors, including forskolin, NGF, SMDF, and T3, are reported to be related to Schwann cell differentiation.22 However, we did not observe the differentiation of BMSCs or MSCs into Schwann cells after the addition of these humoral factors. The expression of the TGF-β receptor in Schwann cells and their progenitors has been documented, and the TGF-β signaling pathway is reported to be involved in Schwann cell development, myelination, proliferation, and apoptosis.18,24 However, the addition of TGF-β did not significantly influence the differentiation of hBMSCs or hMSCs into Schwann cells in this study (the data for hBMSCs with TGF-β was not shown). It is possible that an unknown humoral factor(s) may be produced by the NB cells.
We used S100 as a marker of Schwannian stroma in the histological examination of the NBs, similar to the protocol used for the clinicopathological evaluation of NBs. However, S100 protein is always expressed in hBMSCs and hMSCs in vitro because of the existence of adipocytes among the hBMSCs and hMSCs. Thus, we used the Schwann cell-specific marker S100beta and NGFRp75 for increased precision. Egr-2 was also used to evaluate the early stage differentiation of Schwann cells. Egr-2 expression was monitored using immunoblotting during the Schwann cell differentiation of the hBMSCs and hMSCs co-cultured with NB cells. Several molecules including Egr-1, Egr-2, GAP-43, O4, MAG, and CNPase are expressed during Schwann cell differentiation.23,40-42 Egr-2 is a transcription factor expressed in Schwann cell precursors during the early stages of differentiation, and the increased expression of Egr-2 is seen after nerve impairment.23 Egr-2 is thought to be a marker for earl-stage cellular differentiation into Schwann cells.
We also examined whether other types of human cells, such as fibroblasts from human embryo, fetal lung or skin tissue, were also able to differentiate into Schwann cells in NB tumors. MRC-5 cells (originating from human fetal lung tissue) were induced to differentiate into Schwann cells by co-culturing with NB cells SK-N-DZ, as seen with the hBMSCs. However, the KMS-6 cells (originating from a human embryo) and the HE-1 cells (originating from human skin) did not differentiate into Schwann cells. Thus, fibroblasts may represent a heterogeneous cell population. Differences in gene expression between the MRC-5, KMS-6, and HE-1 cells may explain the molecular mechanism responsible for the induction of Schwann cell differentiation. Egr-2 induction was only seen in the co-cultures of fibroblasts (MRC-5) and SK-N-DZ cells but not in the co-cultures of fibroblasts and SK-N-AS cells. This may be attributable to the difference in differentiation between the SK-N-DZ cells and the SK-N-AS cells. In this model, the hBMSCs contributed to the formation of microvessel structures such as pericytes or myofibroblasts. This phenomenon may be compatible with evidence showing that Flk-1+ or Flt-1+ cells from bone marrow are involved in tumor angiogenesis because some hBMSCs and hMSCs are known to express Flk-1and Flt-1.11,12 The use of Schwann cells as a therapy for NB may be reasonable, but Schwann cells are hard to obtain. A useful source of Schwann cells may be hBMSCs and hMSCs because these cells have both proliferative capacity and multipotency.
We have presented an animal model for the reconstruction of human tumor stroma by injecting hBMSCs or mesenchymal stem cells into xenografted NB tumors. We have shown that stromal cells derived from bone marrow differentiate into Schwann cells in NB tumors in vivo and in co-culture with NB cells in vitro. These observations suggest that the origin of Schwannian stroma in NB tumors may be nonneoplastic cells rather than NB cells.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (07-17), Grants-in Aid for Pediatric Research (9C-5, 12C-1) from the Ministry of Education in Japan (05-045-0203), and Keio Gijuku Academic Development Funds and a special grant-in-aid for innovative and collaborative research projects at Keio University.
Supplemental material for this article can be found on http://ajp.amjpathol.org.
Accepted for publication July 8, 2008.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
