Malignant Glioma Cancer Stem Cells Regulate Tumor Cell Invasiveness: Implications For Invasive Characteristics In Aggressive Brain Tumors

IntroductionMalignant gliomas are the most common and most aggressive primary brain tumors that respond poorly to current treatment strategies of cytoreductive neurosurgery, involved field radiotherapy,and adjuvant chemotherapy,which are characterized by rapid cell proliferation,a high level of invasiveness into the surrounding brain and increased vascularity.The capability of glioma cells to invade extensively within the central nervous system is a major cause of the high morbidity rate of primary malignant brain tumors.The invasion of tumor cells into brain tissue is a pathologic hallmark of WHO gradesⅡ-Ⅳgliomas and contributes significantly to the failure of current therapeutic treatments.Glioma cell invasion involves the attachment of tumor cells to extracellular matrix(ECM),degradation of ECM components,and subsequent penetration into adjacent brain structures.These processes are accomplished in part by matrix metalloproteinases(MMPs),a family of zinc-dependent proteinases.The members of the MMPs family have the ability to degrade macromolecules of the ECM and are responsible for tumor invasion and infiltration,limiting the effectiveness of the neurosurgical resection of brain tumors.Recent reports have suggested that expressions of gelatinase-A(MMP-2) and gelatinase-B(MMP-9) play an important role in neoplastic tissue invasion or metastasis,since these are critical factors in basement membrane degradation.A better understanding of glioma invasion mechanism will help in developing therapeutic strategies to decrease the invasion of gliomas.Recently,a large body of evidence has assigned a key role in tumor formation and progression to cancer stem cells(CSCs) by virtue of its ability to self-renew,and differentiate into subpopulations of cells similar to those present in the initial tumor, more importantly,the isolated CSCs are able to form tumors after in vivo implantation. Also glioma CSCs participate in cancer chemoresistance and redioresistance mechanism.To date,invasion has been defined as an important mechanism in glioma tumorigenesis,However,few reports were involved in invasiveness characteristic of glioma CSCs,therefore we isolated and identified glioma CSCs from primary cultured GBM cells-GBM1 and GBM cell lines U87,9L,examined whether glioma CSCs possess potent invasiveness potential and determine tumor progression through promoting tumor invasive processes,and detected the expressions of MMP-2 and -9 in different subpopulations and xenografts in brain tumor models.Here,we reported that malignant glioma CSCs might provide a powerful enhancement for the invasion of glioma through a gelatinases-dependent mechanism.Targeting gelatinases specifically on glioma CSCs may provide a new insight and method for patient therapy by inhibiting glioma invasiveness.MethodsCells and culture processTumor specimen was obtained from surgical resection of one GBM patient,as approved by the Institute Review Boards at the Neurosurgery Department of Cedars-Sinai Medical Center,LA,CA.Tumor tissue was washed,minced,and enzymatically dissociated.U87 cell line and 9L cell line was purchased from ATCC.Tumor cells were resuspended in DMEM/F12 medium containing 10%fetal bovine serum(FBS) Subsequently monolayer-growing tumor cells were cultured in DMEM/F12 supplemented with epidermal growth factor(EGF),and basic fibroblast growth factor(bFGF)(R&D Systems,Minneapolis,MN).Subsphere-forming and differentiation assayThe spheres were harvested,dissociated into single cells and were diluted to1-2 cells/well for the subsphere-forming assay Simultaneously,the monolayer cells left in the flasks after harvesting of the sphere were transferred into medium without growth factors but permissive for differentiation.ImmunocytostainingSpheres and differentiated glioma cells were grown in precoated chamber slides and fixed,and then were stained with the following antibodies:anti-CD133(mouse monoclonal IgG1;1:10;Milteny Biotec),anti-nestin(mouse monoclonal IgG1;1:200; Chemicon),anti-β-tubulinⅢ(mouse monoclonal IgG1;1:200;Chemicon),anti-GFAP (rabbit polyclonal;1:1000;Chemicon),anti-myelin/oligodendrocyte(mouse monoclonal; 1:1000;Chemicon).The primary antibodies were detected with Cy3 or FITC-conjugated anti-mouse or anti-rabbit IgG antibodies(1:200,Jackson Immuno Reserch).Establishment of in vivo glioma modelSix- to 8-week-old athymic nude mice and adult male F344 rats weighing between 200g and 220 g obtained from Medical Animal Center at Cedars-Sinai Medical Center,LA,CA, stereotactically inoculated with either purified glioma sphere cells or non-sphere cells (50,000/animal) in the right corpus striatum.Invasion assaysCell invasion assay kit(ECM 550,polycarbonate membrane,24-well plate with 12 cell culture inserts,8-μm pore size,Chemicon) was used.The number of invasive cells was counted as the indicator of glioma cell invasiveness.Histological and immunohistochemical analysis of glioma-bearing brain sectionsThe animals were euthanized 6 weeks after implantation.Brains harvested from tumor cells implanted animals were frozen into sections,which were stained with H&E staining.Also staining were performed using primary antibodies against MMP-2 and MMP-9(1:50,mouse monoclonal IgG,Calbiochem).Xenografts analysisMatched sphere cells or non-sphere cells were implanted intracranially with 50,000 cells per animal as described previously.The animals were euthanized on 14,21 and 28 day post-implantation,respectively.Real-time Quantitative Reverse transcription-PCRSphere cells and non-sphere cells from three glioma cell lines,and from xenografts of brain tumor models were respectively subjected to total RNA extraction with RNeasy kit (QIAGEN) and reverse transcripted into cDNA.The real-time PCR for CD 133,Nestin, MMP-2 and MMP-9 was performed.The relative mRNA level of these genes was compared between sphere cells and non-sphere calls,as well gelatinases were detected in normal brain tissues and tumor tissues derived from different time points post-implantation,respectively.Protein analysisWestern blots contained 20μg protein extracted from glioma cells,and tumor tissues or normal brain tissues derived from xenografts on a 4-20%Tris/glycine SDS-PAGE gel (Invitrogen) and were transferred to 0.45-μm PVDF(Invitrogen).Blots were incubated with antibodies directed against MMP-2 and MMP-9,and in HRP-conjugated IgG.Blots were developed and exposed to film.β-actin as control.Statistical analysisStatistical analyses were done using SPSS 13.0.Data graphed with error bars represent mean and SD.A two-sided Student's t test was used to determine the significance of any differences.P＜0.05 as statistical significance. ResultsIsolation and identification of glioma CSCsPrimary cultured sphere glioma cells formed subspheres under limiting passage conditions consistent with self renewal,expressed neural stem cell markets(CD133, nestin),and generated multiple-lineage daughter cells,and formed a neoplasm after intracranial implantation.In contrast,few cells of the non-sphere forming tumor cells in the monolayer culture stained positive for CD133 and nestin.Together these results fulfilled the criteria of two distinct tumor cell subpopulations:CSCs and non-CSCs.Glioma CSCs presented increased invasive ability combined with elevated levels of MMP-2,but not MMP-9 in vitroTo determine invasiveness capacity for different subpopulations in three glioma cell lines,respectively.The results showed that significantly more sphere cells invaded out of the membrane than non-sphere forming cells(P＜0.01),and single sphere cells formed new subspheres after they passed through ECM gel.Real-time RT-PCR was used to detect gelatinases mRNA expression.MMP-2 mRNA expression in glioma CSCs ranged from 6.56- to 9.43-fold higher than that in autologous non-CSCs(P＜0.0001),however,only modest differences of MMP-9 mRNA expression between CSCs and non-CSCs subpopulations were observed in three glioma cell lines(P＞0.05),moreover the mRNA expression levels of MMP-9 were lower than those of MMP-2 strikingly,especially in glioma CSCs subpopulation.Western blot results showed markedly increased secretion of MMP-2 from glioma CSCs versus non-CSCs subpopulation,but except a trace of expression in CSCs subpopulation derived from GBMI cell,MMP-9 expression wasn't detected in other subpopulations. Tumors derived from sphere glioma cells showed increased tumor invasiveness and proliferationRegardless of the differing origins of glioma cell lines,the brains implanted with sphere tumor cells presented highly proliferative and invasive neoplasms with extensive vascularity.However,among the animals implanted with non-sphere tumor cells,only 2 of 6 rats implanted with 9L cell line formed small neopalsms,in other brains tumors weren't formed 6 weeks after implantation.The expression of MMP-2 and MMP-9 were highly up-regulated in brain tumors derived from glioma CSCsImmunohistochemical results showed,regardless of tumor cell source,MMP-2 was expressed both in normal brain tissues and tumor tissues,and in tumor tissues, MMP-2 was present dominantly in the center of tumor,in contrast,MMP-9 expression was present in very low levels in normal brain tissues,but in the tumors it was greatly restricted to regions at the infiltrating and invading border of the tumor.Real-time RT-PCR results revealed that the mRNA expression levels of MMP-9 were very low and only a trace of expression of MMP-2 mRNA was detected in normal brain tissues;the expression of MMP-2 and MMP-9 mRNA were markedly elevated in tumor tissues at 3 time points post-implantation with the growth of malignant glioma. The mRNA expression levels of both MMP-2 and MMP-9 at 28 days were striking increased compared with that post-implantation for 14 days.However,the growth rate of MMP-9 mRNA level along with the time after implantation was higher than that of MMP-2,especially on 28 days versus on 14 days post-implantation.Western blot assay showed that MMP-2 was detected at 72 kDa in both tumor tissues and normal brain tissues,but the expression in tumor tissues was much higher than that in normal brain tissues.MMP-9 was detected at 92 kDa in tumor tissues,but not in normal brain tissues except modest expression in those derived from U87 sphere cells.In addition,western blot results showed increased secretion of MMP-2 and MMP-9 along with the time post sphere tumor cell implantation.Both MMP-2 and MMP-9 expressions 28 days post-implantation were evidently increased compared to those from 14 days,particularly for MMP-9.Conclusions1 In our study,we confirmed cultured sphere glioma cells from glioma cell lines formed new spheres with the capacity for self renewal,expressed gene associated with neural stem cells,generated daughter cells of different phenotypes from one mother cell, and formed tumors after intracranial implantation both in mice and rats.These sphere cells could be used for further study as optimal CSCs.2 Our findings above suggest the fact that glioma CSCs enhance the invasiveness potential of malignant glioma,at least in part,is because the small subpopulation shares both characteristics of glioma tumor cells and NSCs simultaneously.Our results of CSCs inducing the increased invasiveness in cell invasion assay and in xenograft evaluations extend these findings of a key role of CSCs in glioma invasive tumorigenesis.Glioma CSCs maybe contribute to tumor progression through promoting tumor invasion.3 Together these results suggest that glioma CSCs can up-regulate the expression of MMP-2 that promote their invasion and migration in vitro.Also of interest was our finding that MMP-9 might exhibit stronger invasion and infiltration capability in glioma tumorigenesis than MMP-2 in vivo.Glioma CSCs may possess increased invasiveness ability and enhance tumor invasive progression through gelatinases-related mechanism under different circumstances.