Invasion of human glioma biopsy specimens in cultures of rodent brain slices: a quantitative analysis

Full access

Object. The reliable assessment of the invasiveness of gliomas in vitro has proved elusive, because most invasion assays inadequately model in vivo invasion in its complexity. Recently, organotypical brain cultures were successfully used in short-term invasion studies on glioma cell lines. In this paper the authors report that the invasiveness of human glioma biopsy specimens directly implanted into rodent brain slices by using the intraslice implantation system (ISIS) can be quantified with precision. The model was first validated by the demonstration that, in long-term studies, established glioma cells survive in the ISIS and follow pathways of invasion similar to those in vivo.

Methods. Brain slices (400 µm thick) from newborn mice were maintained on millicell membranes for 15 days. Cells from two human and one rodent glioblastoma multiforme (GBM) cell lines injected into the ISIS were detected by immunohistochemistry or after transfection with green fluorescent protein—containing vectors. Preferential migration along blood vessels was identified using confocal and fluorescent microscopy. Freshly isolated (≤ 24 hours after removal) 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate—prelabeled human glioma biopsy specimens were successfully implanted in 19 (83%) of 23 cases, including 12 GBMs and seven lower grade gliomas (LGGs). Morphometric quantification of distance and density of tumor cell invasion showed that the GBMs were two to four times more invasive than the LGGs. Heterogeneity of invasion was also observed among GBMs and LGGs. Directly implanted glioma fragments were more invasive than spheroids derived from the same biopsy specimen.

Conclusions. The ISIS combines a high success rate, technical simplicity, and detailed quantitative measurements and may, therefore, be used to study the invasiveness of biopsy specimens of gliomas of different grades.

Due to their high rate of proliferation and profuse invasiveness, diffuse gliomas are invariably fatal.49 Estimates of glioma proliferation are obtained by histopathological grading34 and a variety of auxiliary methods.11,12 In contrast, brain invasiveness, an intrinsic property of diffuse gliomas, is not routinely evaluated. Diffuse gliomas are intraparenchymally metastatic tumors that invade the brain from the very onset of their development.4 As a consequence, they are not amenable to complete surgical resection and defy other current therapeutic strategies.49

Basic questions concerning the invasiveness of gliomas, such as whether their invasiveness increases with histological grade9,16,20,22,25 or whether all gliomas of the same grade and type invade the brain to a similar degree,16,40 still remain unanswered. One reason for these uncertainties is purely technical.32 Over the years, several attempts have been made to design reliable methods to evaluate the invasiveness of gliomas that have been removed at biopsy.9,16,20,40 Investigations of monolayer migration9,20,28 and filter-based transmigration,43 as well as studies in which glioma spheroids are confronted with various invasion substrates,13,15,16,40 have highlighted isolated aspects of glioma cell motility and invasiveness.32,47 None of these methods combines technical simplicity, reliability, and precision of the estimate. Moreover, some results have brought into question the very relevance of these studies to circumstances in vivo.32

In an attempt to design an ex vivo system to be used reliably to analyze the invasiveness of individual human gliomas removed at biopsy, we have turned to cultures of organotypical brain slices as recipients for implanted glioma cells. Such brain slices have been shown to survive in culture for several weeks at least, preserving tissue architecture and composition (see Discussion and references listed in the study by Gähwiler, et al.21). Recently, organotypical slices of rodent36,41 and human brain31 were successfully used to measure and modulate the invasiveness of established glioma cell lines in experiments lasting from 2 to 7 days. Results of preliminary study have indicated that such short experimental protocols do not permit the researcher sensitively to detect differences in the invasiveness of gliomas retrieved at biopsy. To validate the use of the ISIS in longer term studies we first implanted various GBM cell lines of different known in vivo invasion potentials and followed their evolution for up to 29 days. In a second step, the method was adapted to analyze tumor tissue obtained at biopsy from 27 consecutive patients. The results indicate that the ISIS is an ex vivo system that permits the precise characterization of the invasiveness of individual gliomas.

Materials and Methods

Brain Slice Cultures

Cultures of brain slices obtained from 7-day-old C57/bl6 mice were prepared using a modification of a previously published method.50 The mouse brains were removed and placed in 0.5 g/L d-glucose phosphate-buffered saline, the two cerebral hemispheres were separated, and 400-µm-thick coronal slices were cut using a tissue chopper. Cut slices were transferred onto millicell membranes and kept in six-well plates above 1 ml of brain-slice medium (MEM containing 1 g/L d-glucose, 10% heat-inactivated FCS, 0.1 g/L transferrin, 16 µg/L putrescine, 40 µg/L N-selenium, 30 µg/L triiodothyronine, 5 mg/L insulin, and 60 µg/L progesterone). The brain slices were incubated at 37°C in a humidified atmosphere containing 5% CO2. The medium was changed three times per week.

Established GBM Cell Lines

The two human GBM cell lines, GL156 and 8-MG-BA,46 and the rodent C6 glioma cell line3 were grown in glioma-cell medium (50% MEM and 50% Dulbecco modified Eagle medium), complemented with 10% FCS, 1 mM pyruvate, 33 mM glucose, 2 mM glutamine, and antibiotic agents) in a standard cell incubator. On confluence, the cells were trypsinized, centrifuged at 300 G for 10 minutes, and resuspended in 30 to 50 µl of brain-slice medium to obtain 5 × 104 cells/µl for injection. To obtain spontaneously fluorescent cells, the tumor cells were infected with a GFP-recombinant (FOCHA [retrovirus developed at the Laboratoire Transfert Génétique et Oncologie Moléculaire]—GFP) Friend murine leukemia virus—based retroviral vector.10 A single round of exposure to virus supernatant resulted in an infection covering 98% of bulk cells, as evidenced by the expression of GFP, which proved stable over time (2 years). The GFP-expressing cells exhibited a stable and unmodified phenotype, compared with nonmanipulated cells.

Human Biopsy Specimens and Cultures

Gliomas obtained during open surgery or stereotactic biopsy were retrieved in the operating room. One fragment was submitted for histological examination, a second one was snap-frozen and stored at −80°C, and the third one was transferred within 20 minutes after retrieval to glioma-cell medium supplemented with antibiotic medications and maintained under a laminar flow hood. After removal of grossly necrotic and hemorrhaged areas, each fragment was cut into pieces (0.1–0.4 mm in diameter) by using two 26-gauge needles with the aid of an operating microscope. Care was taken not to mince or tease the tissue while cutting. The tumor content of the fragment to be grafted was confirmed by a frozen section or by subsequent analysis of the paraffin-embedded specimen. Diagnoses were established according to the World Health Organization 1993 classification.34 To generate spheroids, tumor fragments and dissociated cells in glioma-cell medium were incubated in culture dishes, to which a base coat of 1% agar had been applied, for 12 to 24 days, until spheroids had formed.16,17,40 The medium was changed every 5 days. Spheroids (0.1–0.4 mm in diameter) were selected for implantation with the aid of an inverted microscope equipped with a calibrated eye-piece graticule.

Fluorescent Dye Staining

For the purpose of staining, DiI was dissolved in 100% alcohol, sonicated for 10 minutes, and added to the medium to a final concentration of 30 µg/ml.40 Glioma fragments or spheroids were incubated in the dye-containing medium for 20 to 24 hours before implantation. Slices were fixed and mounted in a manner described later in this paper. In each case, tumor cells in several slices were also detected by immunolabeling for vimentin, although DiI-prelabeled specimens were preferred for measurements of invasion.

Cell Implantation

Injection of glioma cell lines and implantation of tumor tissue collected at biopsy into the cortex of the brain slices were performed with the aid of an operating microscope. Tumor cells were injected (0.1 µl of cell suspension over 1.5 minutes) using a 0.5-µl Hamilton syringe with a 230-µm-wide outer-diameter needle affixed to the arm of a micromanipulator. Tumor fragments or spheroids were deposited onto the slice surface and gently pushed with the tip of a needle until the tumor tissue was well enveloped by the parenchyma of the slice. Slices containing injected cells were analyzed at 8 and 29 days, and tumor fragments or spheroids at 15 days after implantation.

Immunohistochemical Analysis

Slices fixed in 4% paraformaldehyde for 4 hours at 4°C were removed from the filter membrane, washed in phosphate-buffered saline, and incubated overnight at 4°C with monoclonal antibodies against HLA-ABC (concentration 1:200), and vimentin (concentration 1:200), or polyclonal antibodies against GFAP (concentration 1:400) and laminin (concentration 1:200). To detect bound antibodies, secondary anti—rabbit or anti—mouse immunoglobulin conjugated to fluorescein isothiocyanate (concentration 1:200) or to cyanidine-Cy3 (concentration 1:1000) was applied for 3 hours at room temperature. Double immunolabelings for tumor cells (HLA-ABC or vimentin) and blood vessel basal lamina (laminin) were also performed. The slices were mounted and covered.

Fluorescence Microscopy and Morphometry

Fluorescence microscopy and confocal microscopy were both performed. The images obtained were digitized and then processed using image analysis software. All areas containing invasive glioma cells were scanned by referring to nonoverlapping microscopic fields (× 200; 0.4 mm2), starting at the margin of the implant and moving away from it. Using optical sectioning, consecutive images throughout the slice thickness were acquired in each field to register all invasive cells that were present. For results, this procedure corresponded well to confocal sectioning of the slice thickness and image analysis of three-dimensional reconstituted images, but was considerably more time effective. In the case of GL15 cells, two types of invasion were separately quantified: invasion in the peritumoral zone, where tumor cells intermingled with host elements but were still in contact with the implant, and long-distance invasion of single detached cells. Coordinates of each field were recorded to map the invasion zone and invasion distance (mm), and the cell density of invasion (cells/mm2) was estimated using the point-measurement mode of the image analyzer. All invasive cells present were evaluated, yielding mean ± standard deviations of 2300 ± 1160 cells/GBM specimen (range 376–4766 cells/GBM specimen) and 310 ± 244 cells/LGG specimen (range 66–852 cells/LGG specimen).

Statistical Analysis

Statistical analysis was performed using two commercially available software packages. In each case, the maximum distance and cell density of invasion were assessed as the 90th percentile calculated from the distribution of all recorded invasive cells. Estimates of maximum invasion were preferred to median or mean values because distributions were markedly asymmetrical and because the most aggressive tumor cell subpopulations could be expected to determine the invasive behavior of the tumor in vivo.16 Two groups were compared using the Mann—Whitney, Wald—Wolfowitz, or Wilcoxon tests, and three or more groups were compared using the Kruskal—Wallis test followed by the Bonferroni—Dunn posttest. Results provided in the text are presented as the means ± standard deviations. The level of significance was set at 0.05.

Sources of Supplies and Equipment

The McIllwain tissue chopper was acquired from Mickle Laboratory (Guildford, UK) and the Millicell-CM membranes from Millipore (Saint Quentin, France). The MEM was provided by Gibco (Cergy, France), and the d-glucose, heat-inactivated FCS, transferrin, putrescine, N-selenium, triiodothyronine, insulin, and progesterone were all purchased from Sigma Chemical Co. (Saint Quentin, France). Life Technologies (Grand Island, NY) provided the Dulbecco modified Eagle medium, and Molecular Probes (Eugene, OR) the DiI. The micromanipulator used for cell implantation was purchased from Narashige, Inc. (Tokyo, Japan). The monoclonal antibodies against HLA-ABC and vimentin, and the polyclonal antibody against GFAP as well as the Vectashield coverslips were acquired from Dako Corp. (Glostrup, Denmark). The polyclonal antibody against laminin was purchased from Sigma. The fluorescein isothiocyanate was obtained from Vector Laboratories (Burlingame, CA) and the cyanidine-Cy3 from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). Carl Zeiss, Inc. (Oberkochen, Germany) manufactured the Axioplan-2 microscope, the Axiovert 135M microscope, and the KS 400 (version 3.0) image analysis software. Images were digitized using a CoolSNAP camera purchased from Roper Scientific (Tucson, AZ). Inspector software was purchased from Matrox Imaging (Rungis, France). Statistical analysis was performed with the aid of both StatView F 4-11 (Abacus Concepts, Inc., Berkeley, CA) and InStat (GraphPad Software, Inc., San Diego, CA) software.

Results

Confocal microscopy demonstrated that, after 2 to 4 weeks in culture, the brain slices retained good preservation of their cytoarchitecture and vascular network. Glioma cells did not migrate on the slice surface, but rather within the slice thickness.

Established Glioma Cell Lines

Two tumor cell lines, rodent C62 and human 8-MG-BA, were observed to develop in the brain slice as spherical non-invasive tumor masses. This behavior was similar to that exhibited by these cell lines in vivo; following injection into the rat brain, C6 and human 8-MG-BA have both displayed few signs of invasion (unpublished data). There were practically no invading cells, no matter what time point was studied (Fig. 1A and B). In sharp contrast, human GL15 cells, which form very invasive tumors when grafted into the rat brain,26 were similarly invasive in the brain slice (Fig. 1C and D). Numerous cells diffusely infiltrated the peritumoral zone. The area of this invasion zone progressively increased from Day 8 (0.074 ± 0.016 mm2; seven brain slices) to Day 29 (0.14 ± 0.04 mm2; nine brain slices) (p < 0.002; Fig. 1C and D, respectively). In addition, cells that had migrated a long distance were present at all time points studied. The maximum migration distances of these cells also increased from Day 8 (0.51 ± 0.106 mm) to Day 29 (0.92 ± 0.18 mm; p < 0.002). Cells migrating long distances were commonly seen to be oriented in parallel to blood vessels, within the slice thickness (Fig. 2A) in tight contact with their basal lamina, as demonstrated by laminin immunostaining (Fig. 2B). The reproducibility of the system proved satisfactory because the coefficients of variation of both parameters of invasion—that is, area of this invasion zone and maximum migration distances—were between 19 and 29%.

Fig. 1.
Fig. 1.

Photomicrographs demonstrating a differential type of expansion following implantation of two distinct tumor cell lines into mouse brain slices. The C6 rodent glioma cells, identified by vimentin immunofluorescence, form a rounded tumor mass 8 days after implantation (A) that expands as a solid mass up to 29 days (B), without a significant number of cells leaving it to invade surrounding neural parenchyma. In sharp contrast, GL15 human glioma cells, identified by spontaneous green fluorescence due to their preimplantation transfection with a GFP retrovirus vector, expand essentially because of a massive migration of cells out of the tumor mass, into surrounding parenchyma, between 8 (C) and 29 (D) days. The cell density of the tumor mass even seems to decrease at the expense of the infiltrating area. Bar = 400 µm (A and B) and 200 µm (C and D).

Fig. 2.
Fig. 2.

Interaction between migrating GL15 tumor cells and blood vessels. A: Results of confocal analysis of the thickness of the slice at the level of the tumor mass showing GFP-fluorescing GL15 cells embedded in the slice, the upper and lower limits of which are visualized through their content of GFAP-immunofluorescent astrocytes. B: The GFP-fluorescing GL15 cells (green) 29 days after implantation. Tumor cells migrating away from the mass (located at the top of the picture) are observed along blood vessels identified by laminin immunofluorescence (red). The most distant cells are at least 2 mm away from the tumor mass (framed area at bottom of the picture, which is analyzed and enlarged in C). C: Confocal analysis of the framed area of the brain slice in panel B showing bipolar tumor cells in tight contact with the laminin-immunostained basal lamina of the blood vessel. Bar = 50 µm (A), 200 µm (B), and 100 µm (C).

Glioma Biopsy Specimens

The ISIS allowed us to establish ex vivo tumor growth in 19 (83%) of 23 cases of diffuse gliomas, including 12 of 15 GBMs and seven of eight LGGs (Grade II or III oligodendrogliomas and oligoastrocytomas). In addition, three nonneuroectodermal tumors (two squamous-cell lung carcinomas that had metastasized to the brain and one meningioma) and one specimen of reactive gliosis (vasculitis-associated perinecrotic gliosis) were grown. In each case, the number of slices with successfully established implants varied between two and 15 (mean 10 ± 4 slices), providing proportions of success ranging from 33 to 100% (mean 76 ± 18%).

All diffuse gliomas (19 of 19 lesions) invaded the brain slice, whereas invasion was not associated with any nonneuroectodermal tumors or nonneoplastic gliosis. Gliomas invaded as single cells that spread into the otherwise unperturbed architecture of the slice (Fig. 3). Intraspecimen variability, as measured by the coefficients of variation of the two invasion estimates, was between 28 and 73%.

Fig. 3.
Fig. 3.

Photomicrographs demonstrating representative tumors 15 days after implantation of fresh tumor fragments from an LGG (A) and a GBM (B) obtained at biopsy. Note the striking differences between the two types of tumor in both the density of cells that have migrated out of the tumor mass and the distance of invasion (LGG [C] and GBM [D]). Bar = 800 µm (A and B) and 100 µm (C and D).

Quantitative intergroup analysis between diffuse gliomas revealed that GBMs were more invasive than LGGs. On average, the maximum distance and cell density of invasion were, respectively, two and four times greater for GBMs than for LGGs (Fig. 4A). Plotting out invasion cell densities against invasion distances for each group showed that the mean invasion cell density was consistently higher for GBMs than for LGGs throughout the invasion area (p < 0.0001; Fig. 4B).

Fig. 4.
Fig. 4.

Graphs showing quantitative comparison between invasive capacities of GBMs (12 specimens) and LGGs (seven specimens). A: The maximum distance of invasion from the tumor margin and the maximum cell density in the invasion area, as assessed in both cases by the 90th percentiles of distribution of all migrating cells (P90) are plotted for each tumor, and the median values were calculated (horizontal bars) separately for GBMs and LGGs. Despite a relative intragroup heterogeneity, statistical comparisons (Mann—Whitney test) show highly significant intergroup differences. B: Mean cell densities for the two groups of tumors plotted against the distance of invasion from the tumor margin. Data represent the means ± standard errors of the mean (Wilcoxon test).

Quantitative intragroup analysis additionally demonstrated heterogeneity among GBMs and LGGs. Significant differences in maximum invasion distances (p < 0.0001) were detected within both GBMs and LGGs (Fig. 4A). Furthermore, tumors exhibiting similar invasion distances could differ significantly in invasion cell density (data not shown).

To determine whether highly invasive cells could preferentially be located at the tumor—parenchyma interface, rather than in more centrally located areas, the invasiveness of specimens selectively sampled from the periphery and from deeper portions of the tumor (as confirmed by histological examination) was compared in six specimens (three GBMs and three LGGs). Maximum invasion distances were significantly higher for peripheral tumor cells in two of three GBMs—2.2 mm compared with 1.1 mm (p < 0.0001) and 1.1 mm compared with 0.7 mm (p < 0.02)—and in two of three LGGs—1 mm compared with 0.6 mm (p < 0.02) and 0.5 mm compared with 0.3 mm (p < 0.03)—whereas in the two other specimens peripheral and centrally located cells displayed similar values. These differences were mainly due to the presence at the periphery of a fraction of highly invasive cells representing 20 to 30% of all invasive cells (Fig. 5).

Fig. 5.
Fig. 5.

Graph depicting the results of a comparative analysis between invasion distances recorded for two tumor fragments, one taken from the center of a GBM and the other from the periphery of the same lesion. For each fragment, the invasion distances have been plotted as progressively increasing percentiles of the population of migrating cells, against the relevant axis of distances from the tumor margin. Note the overlap of distribution up to the 70th percentile (P70) and the left shift for higher values, indicating that the overall wider spread of tumor cells from the peripheral fragment relates to the fact that the 30% of the most invasive cells have migrated significantly farther away than their counterparts from the central one (Wald—Wolfowitz test).

Spheroids were successfully established and implanted from six (50%) of 12 GBMs and from two (33%) of six LGGs. The maximum invasion distance of spheroids was lower by 14 to 54% in all eight cases (p < 0.02), and the maximum invasion cell density was lower by 17 to 60% in seven of eight specimens (p < 0.03; Wilcoxon test) in comparison with paired directly grafted specimens. Coefficients of variation of the maximum invasion distance and maximum invasion cell density in the spheroid experiments (44 ± 11% and 37 ± 18%, respectively) were strikingly similar to the coefficients of variation calculated for the corresponding directly implanted specimens (43 ± 8% and 37 ± 16%, respectively).

Discussion

We demonstrated the efficiency of an ISIS by using mouse brain slices to analyze the invasive behavior of human glioma cells obtained from biopsy specimens. The major advantages of this approach consist of the following: 1) the high rate of successfully established specimens of gliomas of different histological grades (> 80%); 2) the use of a specific organotypical invasion substrate; 3) the precise quantification of invasion that may allow the characterization of individual tumors; and 4) the relative technical simplicity of the procedure.

Importance of an Organ-Specific Environment for Glioma Invasion

In vitro techniques designed to analyze glioma invasion are under intensive investigation.47 They should meet a number of requirements,32,47 such as technical feasibility, preservation of the cell diversity of the tumor, reproduction of in vivo phenomena, and quantification. These requirements, however, have turned out to be difficult to reconcile in a single design. The choice of an organ-specific invasion substrate may be the most critical prerequisite for the validity of a glioma invasion assay.1,14 Simplified in vitro systems are likely to be innately incapable of simulating interactions between tumor cells and the host environment.32,44 At best, monolayer migration and transmigration assays only reproduce isolated aspects of glioma invasion.32 For instance, meningioma cells that do not, as a rule, invade the brain,16 migrate faster than gliomas on plastic surfaces18 and can be very invasive when tested in Boyden chambers.48 Both monolayer and filter-based assays have been criticized for the use of migration or transmigration substrates of limited distribution in the human brain.31,47 Another obvious limitation of these assays stands out in the face of results proposing that “cross talk” between tumor and normal cells may be instrumental to glioma cell invasion.25

Confrontation assays performed using rat fetal brain aggregates, a more organ-specific substrate, have proved superior to most other in vitro techniques.32 Nevertheless, confrontation assays are dominated by cell destruction and replacement and proteolytic degradation of the substrate away from the tumor, rather than by single cell invasion.7,16,33,40 As a consequence, invasion in confrontation assays is assessed semiquantitatively7,16 or by linear regression models of the substrate destruction.33,40,45 These methods have not permitted formal statistical comparisons between tumors of similar40 or different16 grades. In contrast, as also demonstrated by a study on clonal glioma cells,31 examining single invasive cells with the aid of the ISIS permits a detailed quantitative analysis.

Importance of Avoiding Preculturing Steps

The inevitable glioma cell incubation, expansion, and passaging steps9,28,49 have both limited the number of successfully grown specimens (generally < 50%)19 and opened up the possibility of undesired clonal selection.39 Even low passage numbers (< five passages) were shown to affect differentiation42 and the cell clone repertoire,27 and to reduce up to 20-fold the invasiveness of glioma cells.43 Glioma spheroid generation is also not exempt from cell selection. Although glioma cells in spheroids were shown to preserve certain characteristics of the original tumor, such as DNA ploidy,5 clonal deletion has recently been demonstrated more often,30 and that is likely due to the intensive cell turnover, as well as cell shedding during the spheroid formation phase.5 In our hands, spheroids were less invasive than directly grafted fragments. Apart from clonal cell selection, there are other possible explanations for this unexpected result. The microenvironment of glioma-derived spheroids is characterized by cell packing and extensive cell-to-cell contacts affected by formation of gap junctions.38 In an in vitro assay performed to analyze cell migration out of spheroids, it was demonstrated that the migratory capacities of the tumor cells were inversely correlated with the number of gap junctions present in the spheroid. These results suggest that cell packing in glioma spheroids, an artifact of the forced cell cooperation in this model, may hamper detachment from the neighboring cells, a critical early step in invasion.25 In addition, glioma cells in spheroids have been shown to express integrin arrays different from those in the original tumor and to switch to massive production of extracellular matrix molecules.43 Although the effects of such a mesenchymal transdifferentiation on invasion have not been studied, integrins are involved in glioma invasion. We initially resorted to glioma spheroids with the aim to standardize our invasion assay more stringently. The presence of intraspecimen variation in spheroid experiments, however, was not lower than that observed in ISIS experiments, suggesting that variability of invasiveness was mostly due to genuine intratumoral heterogeneity.40 Direct implantation of glioma fragments after careful verification of their tumor-cell content has been successfully used in in vivo invasion studies in which human astrocytomas were grafted into the rat brain.29,51 Our results indicate that, with an appropriate invasion substrate, selected fragments can be used in an ex vivo invasion assay.

Invasiveness of Gliomas in Relationship to Tumor Grade

When confronted with an organ-specific in vitro environment, glioma cells of different origins exhibited different migratory drives. In the ISIS, the invasiveness of glioma cell lines was in good agreement with results obtained in vivo.2,26 Quantified analysis of human glioma biopsy specimens documented intergrade, intragrade, and intraspecimen differences of invasiveness. Grade IV GBMs were conspicuously more invasive than LGGs, in accordance with data from another confrontation assay study.16 Within the LGG group, however, Grade III tumors were not more invasive than Grade II tumors. That the migration capacities of glioma cells could be directly proportional to tumor grade is intuitively acceptable; however, arguments in support of this thesis are not so strong.8,9,20,49 By joining evidence provided by clinics and in vitro studies, Giese and coworkers23–25 and McDonough, et al.,38 have pointed out that glioma invasiveness does not necessarily share a linear relationship with their histological malignancy grade. We are currently examining larger groups of GBMs and LGGs to look for histological features that may be associated with this heterogeneity. Because invasiveness can vary greatly from one tumor to another,40 it has long been anticipated that invasiveness may be correlated with the clinical course of patients.5 Exploration of this hypothesis, however, has been hampered by the insufficient precision of invasion assays.40,52 Using the ISIS, the precise quantification of distance and cell density of invasion and the combination thereof allowed us to analyze individual glioma invasiveness in greater detail than that allowed by any other in vitro method.16,40 Thus, the ISIS may help to establish whether invasion per se is a determinant of the prognosis of patients with gliomas.

The ISIS as a Tool to Isolate and Study Invasive Glioma Cells

It has been suggested that invasive capacities may be particularly present in some DIGCs.49 Because DIGCs are one major source of recurrences following treatment,49 specific characterization of these cells is most relevant to strategies of glioma therapy.22,38 Malignant gliomas evolve through selection of cell clones with growing genetic instability.35 One potential correlate of this concept of clonal evolution is that DIGCs correspond to clones with a specific genetic background that have acquired the capacity to move actively, detach themselves from neighboring cells in the tumor mass, interact with the extracellular matrix and cells of the host, and proteolytically modify their environment (see Discussion in other articles22,25,37). Several results obtained using the ISIS indicate that this method readily allows the researcher to differentiate DIGCs from cells in the tumor mass. Following implantation, some cells were more apt to invade, whereas others remained in the implant. Invasive cells were clearly capable of interacting with the basal lamina of blood vessels, a known specific substrate of glioma dissemination in vivo. Topographically, such highly motile cells were present at the tumor margin, rather than in more centrally located areas in four of six tumors. Presumptive DIGCs isolated from the tumor margin have been shown to grow twice as fast as cells taken from the tumor mass,49 but the expected differences in migration velocities between the two types of cells could not be established in the same monolayer migration study. In contrast, the use of the ISIS seemed to validate the hypothesis of a center-to-periphery gradient of invasiveness in diffuse gliomas. The ISIS may, therefore, help to isolate (by microdissection) DIGCs from the rest of the tumor cells and perform a specific analysis of their properties. In this way, the ISIS could allow us to decipher the driving forces at work in glioma invasion and, eventually, to design targeted antiinvasive therapeutics.

Conclusions

Organotypical cultures of brain slices support the invasive growth of both clonal and freshly isolated human glioma cells. Using the ISIS, gliomas invade as single cells that follow preferential pathways of dissemination, similar to those present in the human brain. Measurements of single invasive cells allow detailed characterization of individual glioma invasiveness. These advantageous features of the ISIS seem to set it apart as the assay of choice for studies on the mechanisms, clinical and pathological correlates, and pharmacological modification of glioma invasion.

Acknowledgments

We thank Dr. A. Perzelova for providing the 8-MG-BA GBM cell line. In addition we thank Drs. P. Decp and M. Djindjian for providing tumor specimens. Doctor Christov and Ms. de Boüard contributed equally to this study.

References

  • 1.

    Antunes LAngioi-Duprez KSBracard SRet al: Analysis of tissue chimerism in nude mouse brain and abdominal xenograft models of human glioblastoma multiforme: what does it tell us about the models and about glioblastoma biology and therapy? J Histochem Cytochem 48:8478582000J Histochem Cytochem 48:

  • 2.

    Barth RF: Rat brain tumor models in experimental neuro-oncology: the 9L, C6, T9, F98, RG2 (D74), RT-2 and CNS-1 gliomas. J Neurooncol 36:911021998J Neurooncol 36:

  • 3.

    Benda PLightbody JSato Get al: Differentiated rat glial cell strain in tissue culture. Science 161:3703711968Science 161:

  • 4.

    Bernstein JJ: Local invasion and intraparenchymal metastasis of astrocytomas. Neuropathol Appl Neurobiol 22:4214241996Bernstein JJ: Local invasion and intraparenchymal metastasis of astrocytomas. Neuropathol Appl Neurobiol 22:

  • 5.

    Bjerkvig RTonnesen ALaerum ODet al: Multicellular tumor spheroids from human gliomas maintained in organ culture. J Neurosurg 72:4634751990J Neurosurg 72:

  • 6.

    Bocchini VCasalone RCollini Pet al: Changes in glial fibrillary acidic protein and karyotype during culturing of two cell lines established from human glioblastoma multiforme. Cell Tissue Res 265:73811991Cell Tissue Res 265:

  • 7.

    Bouterfa HPicht TKess Det al: Retinoids inhibit human glioma cell proliferation and migration in primary cell cultures but not in established cell lines. Neurosurgery 46:4194302000Neurosurgery 46:

  • 8.

    Chicoine MSilbergeld DL: Assessment of brain tumor cell motility in vivo and in vitro. J Neurosurg 82:6156221995J Neurosurg 82:

  • 9.

    Chicoine MRSilbergeld DL: The in vitro motility of human gliomas increases with increasing grade of malignancy. Cancer 75:290429091995Cancer 75:

  • 10.

    Cohen-Haguenauer ORestrepo LMMasset Met al: Efficient transduction of haemopoietic CD34+ progenitors of human origin using an original retrovirus vector derived from Fr-MuLV-FB29: in vitro assessment. Hum Gene Ther 9:2072161998Hum Gene Ther 9:

  • 11.

    Coons SWJohnson PCPearl DK: Prognostic significance of flow cytometry deoxyribonucleic acid analysis of human astrocytomas. Neurosurgery 35:1191261994Neurosurgery 35:

  • 12.

    Coons SWJohnson PCScheithauer BWet al: Improving diagnostic accuracy and interobserver concordance in the classification and grading of primary gliomas. Cancer 79:138113931997Cancer 79:

  • 13.

    de Ridder L: Screening for invasion of the individual human brain tumor in an autologous confrontation system in vitro. Pathobiology 67:1331391999de Ridder L: Screening for invasion of the individual human brain tumor in an autologous confrontation system in vitro. Pathobiology 67:

  • 14.

    de Ridder LBruyneel ECalliauw L: The invasiveness in vitro of brain tumor derived-cells depends on their micro-eco system. Acta Neurochir 130:1401431994Acta Neurochir 130:

  • 15.

    de Ridder LCalliauw L: Invasiveness of primary and secondary brain tumors in vitro correlated with clinical results. Neurosurgery 31:104310481992Neurosurgery 31:

  • 16.

    Engebraaten OBjerkvig RLund-Johansen Met al: Interaction between human brain tumor biopsies and fetal rat brain tissue in vitro. Acta Neuropathol 81:1301401990Acta Neuropathol 81:

  • 17.

    Engebraaten OHjortland GHirschberg Het al: Growth of pre-cultured human glioma specimens in nude rat brain. J Neurosurg 90:1251321999J Neurosurg 90:

  • 18.

    Escallona-Zapata J: Tissue culture in the diagnosis of brain tumorsGarcia JHEscallona-Zapata JSandbank Uet al (eds): Diagnostic Neuropathology. Berlin: Springer-Verlag19902346Diagnostic Neuropathology.

  • 19.

    Farr-Jones MAParney IFPetruk KC: Improved technique for establishing short term human brain tumor cultures. J Neurooncol 43:1101999J Neurooncol 43:

  • 20.

    Friedlander DRZagzag DShiff Bet al: Migration of brain tumor cells on extracellular matrix proteins in vitro correlates with tumor type and grade and involves αv and β1 integrins. Cancer Res 56:193919471996v and β1 integrins. Cancer Res 56:

  • 21.

    Gähwiler BCapogna MDebanne Det al: Organotypic slice cultures: a technique has come of age. Trends Neurosci 20:4714771997Trends Neurosci 20:

  • 22.

    Giese AHagel CKim ELet al: Thromboxane synthase regulates the migratory phenotype of human glioma cells. Neurooncol 1:3131999Neurooncol 1:

  • 23.

    Giese ALaube BZapf Set al: Glioma cell adhesion and migration on human brain sections. Anticancer Res 18:243524471998Anticancer Res 18:

  • 24.

    Giese ALoo MATran Net al: Dichotomy of astrocytoma migration and proliferation. Int J Cancer 67:2752821996Int J Cancer 67:

  • 25.

    Giese AWestphal M: Glioma invasion in the central nervous system. Neurosurgery 39:2352521996Neurosurgery 39:

  • 26.

    Guillamo JSLisovoski FChristov Cet al: Migration pathways of human glioblastoma cells xenografted into the immunosuppressed rat brain. J Neurooncol 52:2052152001J Neurooncol 52:

  • 27.

    Hartmann CKluwe LLucke Met al: The rate of homozygous CDKN2A/p16 deletions in glioma cell lines and in primary tumors. Int J Oncol 15:9759821999Int J Oncol 15:

  • 28.

    Hegedus BCzirok AFazekas Iet al: Locomotion and proliferation of glioblastoma cells in vitro: statistical evaluation of video-microscopic observations. J Neurosurg 92:4284342000J Neurosurg 92:

  • 29.

    Horten BCBasler GAShapiro WR: Xenograft of human malignant glial tumors into brains of nude mice. A histopatholgical study. J Neuropathol Exp Neurol 40:4935111981J Neuropathol Exp Neurol 40:

  • 30.

    Janka MFischer UTonn JCet al: Comparative amplification analysis of human glioma tissue and glioma derived fragment spheroids using reverse chromosome painting (RCP). Anticancer Res 16:260126061996Anticancer Res 16:

  • 31.

    Jung SAckerley CIvanchuk Set al: Tracking the invasiveness of human astrocytoma cells by using green fluorescent protein in an organotypical brain slice model. J Neurosurg 94:80892001J Neurosurg 94:

  • 32.

    Kaczarek EZapf SBouterfa Het al: Dissecting glioma invasion: interrelation of adhesion, migration and intercellular contacts determine the invasive phenotype. Int J Dev Neurosci 17:6256411999Int J Dev Neurosci 17:

  • 33.

    Khoshyomn SPenar PLRossi Jet al: Inhibition of phospholipase C-gamma1 activation blocks glioma cell motility and invasion of fetal rat brain aggregates. Neurosurg 44:5685781999Neurosurg 44:

  • 34.

    Kleihues PScheithauer BBurger P: Histological Typing of Tumors of the Central Nervous Systemed 2. Berlin: Springer-Verlag1993

  • 35.

    Louis DN: A molecular genetic model of astrocytoma histopathology. Brain Pathol 7:7557641997Louis DN: A molecular genetic model of astrocytoma histopathology. Brain Pathol 7:

  • 36.

    Matsumura HOhnishi TKanemura Yet al: Quantitative analysis of glioma cell invasion by confocal laser scanning microscopy in a novel brain slice model. Biochem Biophys Res Commun 269:5135202000Biochem Biophys Res Commun 269:

  • 37.

    McDonough WTran NGiese Aet al: Altered gene expression in human astrocytoma cells selected for migration: I. Thromboxane synthase. J Neuropathol Exp Neurol 57:4494551998J Neuropathol Exp Neurol 57:

  • 38.

    McDonough WSJohansson AJoffee Het al: Gap junction intercellular communication in gliomas is inversely related to cell motility. Int J Dev Neurosci 17:6016111999Int J Dev Neurosci 17:

  • 39.

    Noble M: Can neural stem cells be used to track down and destroy migratory brain tumor cells while also providing a means of repairing tumor-associated damage? Proc Natl Acad Sci USA 97:12393123952000Noble M: Can neural stem cells be used to track down and destroy migratory brain tumor cells while also providing a means of repairing tumor-associated damage? Proc Natl Acad Sci USA 97:

  • 40.

    Nygaard SJHaugland HKLaerum ODet al: Dynamic determination of human glioma invasion in vitro. J Neurosurg 89:4414471998J Neurosurg 89:

  • 41.

    Ohnishi TMatsumura HIzumoto Set al: A novel model of glioma cell invasion using organotypic brain slice culture. Cancer Res 58:293529401998Cancer Res 58:

  • 42.

    Paulus WHuettner CTonn JC: Collagens, integrins and the mesenchymal drift in glioblastomas: a comparison of biopsy specimens, spheroid and early monolayer cultures. Int J Cancer 58:8418461994Int J Cancer 58:

  • 43.

    Paulus WTonn JC: Basement membrane invasion of glioma cells mediated by integrin receptors. J Neurosurg 80:5155191994J Neurosurg 80:

  • 44.

    Pedersen PHNess GOEngebraaten Oet al: Heterogeneous response to the growth factors [EGF, PDGF (bb), TGF-alpha, bFGF, IL-2] on glioma spheroid growth, migration and invasion. Int J Cancer 56:2552611994Int J Cancer 56:

  • 45.

    Penar PLKhoshyomn SBhushan Aet al: Inhibition of glioma invasion of fetal brain aggregates. In Vivo 12:75841998In Vivo 12:

  • 46.

    Perzelova AMacikova IMraz Pet al: Characterization of two new permanent glioma cell lines 8-mG-BA and 42-mG-BA. Neoplasma 45:25291998Neoplasma 45:

  • 47.

    Pilkington GJBjerkvig RDe Ridder Let al: In vitro and in vivo models for the study of brain tumor invasion. Anticancer Res 17:410741091997Anticancer Res 17:

  • 48.

    Rooprai HKLiyanage KRobinson SFet al: Extracellular matrix-modulated differential invasion of human meningioma cell lines in vitro. Neurosci Lett 263:2142161999Neurosci Lett 263:

  • 49.

    Silbergeld DChicoine M: Isolation and characterization of human malignant glioma cells from histologically normal brain. J Neurosurg 86:5255311997J Neurosurg 86:

  • 50.

    Stoppini LBuchs PAMuller D: A simple method for organotypic cultures of nervous tissue. J Neurosci Methods 37:1731821991J Neurosci Methods 37:

  • 51.

    Stromblad LGBrun ASalford LGet al: A model for xenotransplantation of human malignant astrocytomas into the brain of normal adult rats. Acta Neurochir 65:2172261982Acta Neurochir 65:

  • 52.

    Wester KBjerkvig RCressey Let al: Organ culture of a glioblastoma from a patient with an unusually long survival. Neurosurgery 35:4284331994Neurosurgery 35:

This work was supported by Institut Nationale de la Santé et de la Recherche Médicale, Association pour la Recherche sur le Cancer (to Dr. Lefrançois, 9032), Ligue Nationale contre le Cancer, and fellowships from Association Française de Recherche Génétique (to Ms. de Boüard).

Article Information

Address reprint requests to: Marc Peschanski, M.D., Ph.D., INSERM, Unité 421, IM3, Faculté de Médecine, 8 rue du Général Sarrail, 94010 Créteil, France. email: peschanski@im3.inserm.fr.

© AANS, except where prohibited by US copyright law."

Headings

Figures

  • View in gallery

    Photomicrographs demonstrating a differential type of expansion following implantation of two distinct tumor cell lines into mouse brain slices. The C6 rodent glioma cells, identified by vimentin immunofluorescence, form a rounded tumor mass 8 days after implantation (A) that expands as a solid mass up to 29 days (B), without a significant number of cells leaving it to invade surrounding neural parenchyma. In sharp contrast, GL15 human glioma cells, identified by spontaneous green fluorescence due to their preimplantation transfection with a GFP retrovirus vector, expand essentially because of a massive migration of cells out of the tumor mass, into surrounding parenchyma, between 8 (C) and 29 (D) days. The cell density of the tumor mass even seems to decrease at the expense of the infiltrating area. Bar = 400 µm (A and B) and 200 µm (C and D).

  • View in gallery

    Interaction between migrating GL15 tumor cells and blood vessels. A: Results of confocal analysis of the thickness of the slice at the level of the tumor mass showing GFP-fluorescing GL15 cells embedded in the slice, the upper and lower limits of which are visualized through their content of GFAP-immunofluorescent astrocytes. B: The GFP-fluorescing GL15 cells (green) 29 days after implantation. Tumor cells migrating away from the mass (located at the top of the picture) are observed along blood vessels identified by laminin immunofluorescence (red). The most distant cells are at least 2 mm away from the tumor mass (framed area at bottom of the picture, which is analyzed and enlarged in C). C: Confocal analysis of the framed area of the brain slice in panel B showing bipolar tumor cells in tight contact with the laminin-immunostained basal lamina of the blood vessel. Bar = 50 µm (A), 200 µm (B), and 100 µm (C).

  • View in gallery

    Photomicrographs demonstrating representative tumors 15 days after implantation of fresh tumor fragments from an LGG (A) and a GBM (B) obtained at biopsy. Note the striking differences between the two types of tumor in both the density of cells that have migrated out of the tumor mass and the distance of invasion (LGG [C] and GBM [D]). Bar = 800 µm (A and B) and 100 µm (C and D).

  • View in gallery

    Graphs showing quantitative comparison between invasive capacities of GBMs (12 specimens) and LGGs (seven specimens). A: The maximum distance of invasion from the tumor margin and the maximum cell density in the invasion area, as assessed in both cases by the 90th percentiles of distribution of all migrating cells (P90) are plotted for each tumor, and the median values were calculated (horizontal bars) separately for GBMs and LGGs. Despite a relative intragroup heterogeneity, statistical comparisons (Mann—Whitney test) show highly significant intergroup differences. B: Mean cell densities for the two groups of tumors plotted against the distance of invasion from the tumor margin. Data represent the means ± standard errors of the mean (Wilcoxon test).

  • View in gallery

    Graph depicting the results of a comparative analysis between invasion distances recorded for two tumor fragments, one taken from the center of a GBM and the other from the periphery of the same lesion. For each fragment, the invasion distances have been plotted as progressively increasing percentiles of the population of migrating cells, against the relevant axis of distances from the tumor margin. Note the overlap of distribution up to the 70th percentile (P70) and the left shift for higher values, indicating that the overall wider spread of tumor cells from the peripheral fragment relates to the fact that the 30% of the most invasive cells have migrated significantly farther away than their counterparts from the central one (Wald—Wolfowitz test).

References

1.

Antunes LAngioi-Duprez KSBracard SRet al: Analysis of tissue chimerism in nude mouse brain and abdominal xenograft models of human glioblastoma multiforme: what does it tell us about the models and about glioblastoma biology and therapy? J Histochem Cytochem 48:8478582000J Histochem Cytochem 48:

2.

Barth RF: Rat brain tumor models in experimental neuro-oncology: the 9L, C6, T9, F98, RG2 (D74), RT-2 and CNS-1 gliomas. J Neurooncol 36:911021998J Neurooncol 36:

3.

Benda PLightbody JSato Get al: Differentiated rat glial cell strain in tissue culture. Science 161:3703711968Science 161:

4.

Bernstein JJ: Local invasion and intraparenchymal metastasis of astrocytomas. Neuropathol Appl Neurobiol 22:4214241996Bernstein JJ: Local invasion and intraparenchymal metastasis of astrocytomas. Neuropathol Appl Neurobiol 22:

5.

Bjerkvig RTonnesen ALaerum ODet al: Multicellular tumor spheroids from human gliomas maintained in organ culture. J Neurosurg 72:4634751990J Neurosurg 72:

6.

Bocchini VCasalone RCollini Pet al: Changes in glial fibrillary acidic protein and karyotype during culturing of two cell lines established from human glioblastoma multiforme. Cell Tissue Res 265:73811991Cell Tissue Res 265:

7.

Bouterfa HPicht TKess Det al: Retinoids inhibit human glioma cell proliferation and migration in primary cell cultures but not in established cell lines. Neurosurgery 46:4194302000Neurosurgery 46:

8.

Chicoine MSilbergeld DL: Assessment of brain tumor cell motility in vivo and in vitro. J Neurosurg 82:6156221995J Neurosurg 82:

9.

Chicoine MRSilbergeld DL: The in vitro motility of human gliomas increases with increasing grade of malignancy. Cancer 75:290429091995Cancer 75:

10.

Cohen-Haguenauer ORestrepo LMMasset Met al: Efficient transduction of haemopoietic CD34+ progenitors of human origin using an original retrovirus vector derived from Fr-MuLV-FB29: in vitro assessment. Hum Gene Ther 9:2072161998Hum Gene Ther 9:

11.

Coons SWJohnson PCPearl DK: Prognostic significance of flow cytometry deoxyribonucleic acid analysis of human astrocytomas. Neurosurgery 35:1191261994Neurosurgery 35:

12.

Coons SWJohnson PCScheithauer BWet al: Improving diagnostic accuracy and interobserver concordance in the classification and grading of primary gliomas. Cancer 79:138113931997Cancer 79:

13.

de Ridder L: Screening for invasion of the individual human brain tumor in an autologous confrontation system in vitro. Pathobiology 67:1331391999de Ridder L: Screening for invasion of the individual human brain tumor in an autologous confrontation system in vitro. Pathobiology 67:

14.

de Ridder LBruyneel ECalliauw L: The invasiveness in vitro of brain tumor derived-cells depends on their micro-eco system. Acta Neurochir 130:1401431994Acta Neurochir 130:

15.

de Ridder LCalliauw L: Invasiveness of primary and secondary brain tumors in vitro correlated with clinical results. Neurosurgery 31:104310481992Neurosurgery 31:

16.

Engebraaten OBjerkvig RLund-Johansen Met al: Interaction between human brain tumor biopsies and fetal rat brain tissue in vitro. Acta Neuropathol 81:1301401990Acta Neuropathol 81:

17.

Engebraaten OHjortland GHirschberg Het al: Growth of pre-cultured human glioma specimens in nude rat brain. J Neurosurg 90:1251321999J Neurosurg 90:

18.

Escallona-Zapata J: Tissue culture in the diagnosis of brain tumorsGarcia JHEscallona-Zapata JSandbank Uet al (eds): Diagnostic Neuropathology. Berlin: Springer-Verlag19902346Diagnostic Neuropathology.

19.

Farr-Jones MAParney IFPetruk KC: Improved technique for establishing short term human brain tumor cultures. J Neurooncol 43:1101999J Neurooncol 43:

20.

Friedlander DRZagzag DShiff Bet al: Migration of brain tumor cells on extracellular matrix proteins in vitro correlates with tumor type and grade and involves αv and β1 integrins. Cancer Res 56:193919471996v and β1 integrins. Cancer Res 56:

21.

Gähwiler BCapogna MDebanne Det al: Organotypic slice cultures: a technique has come of age. Trends Neurosci 20:4714771997Trends Neurosci 20:

22.

Giese AHagel CKim ELet al: Thromboxane synthase regulates the migratory phenotype of human glioma cells. Neurooncol 1:3131999Neurooncol 1:

23.

Giese ALaube BZapf Set al: Glioma cell adhesion and migration on human brain sections. Anticancer Res 18:243524471998Anticancer Res 18:

24.

Giese ALoo MATran Net al: Dichotomy of astrocytoma migration and proliferation. Int J Cancer 67:2752821996Int J Cancer 67:

25.

Giese AWestphal M: Glioma invasion in the central nervous system. Neurosurgery 39:2352521996Neurosurgery 39:

26.

Guillamo JSLisovoski FChristov Cet al: Migration pathways of human glioblastoma cells xenografted into the immunosuppressed rat brain. J Neurooncol 52:2052152001J Neurooncol 52:

27.

Hartmann CKluwe LLucke Met al: The rate of homozygous CDKN2A/p16 deletions in glioma cell lines and in primary tumors. Int J Oncol 15:9759821999Int J Oncol 15:

28.

Hegedus BCzirok AFazekas Iet al: Locomotion and proliferation of glioblastoma cells in vitro: statistical evaluation of video-microscopic observations. J Neurosurg 92:4284342000J Neurosurg 92:

29.

Horten BCBasler GAShapiro WR: Xenograft of human malignant glial tumors into brains of nude mice. A histopatholgical study. J Neuropathol Exp Neurol 40:4935111981J Neuropathol Exp Neurol 40:

30.

Janka MFischer UTonn JCet al: Comparative amplification analysis of human glioma tissue and glioma derived fragment spheroids using reverse chromosome painting (RCP). Anticancer Res 16:260126061996Anticancer Res 16:

31.

Jung SAckerley CIvanchuk Set al: Tracking the invasiveness of human astrocytoma cells by using green fluorescent protein in an organotypical brain slice model. J Neurosurg 94:80892001J Neurosurg 94:

32.

Kaczarek EZapf SBouterfa Het al: Dissecting glioma invasion: interrelation of adhesion, migration and intercellular contacts determine the invasive phenotype. Int J Dev Neurosci 17:6256411999Int J Dev Neurosci 17:

33.

Khoshyomn SPenar PLRossi Jet al: Inhibition of phospholipase C-gamma1 activation blocks glioma cell motility and invasion of fetal rat brain aggregates. Neurosurg 44:5685781999Neurosurg 44:

34.

Kleihues PScheithauer BBurger P: Histological Typing of Tumors of the Central Nervous Systemed 2. Berlin: Springer-Verlag1993

35.

Louis DN: A molecular genetic model of astrocytoma histopathology. Brain Pathol 7:7557641997Louis DN: A molecular genetic model of astrocytoma histopathology. Brain Pathol 7:

36.

Matsumura HOhnishi TKanemura Yet al: Quantitative analysis of glioma cell invasion by confocal laser scanning microscopy in a novel brain slice model. Biochem Biophys Res Commun 269:5135202000Biochem Biophys Res Commun 269:

37.

McDonough WTran NGiese Aet al: Altered gene expression in human astrocytoma cells selected for migration: I. Thromboxane synthase. J Neuropathol Exp Neurol 57:4494551998J Neuropathol Exp Neurol 57:

38.

McDonough WSJohansson AJoffee Het al: Gap junction intercellular communication in gliomas is inversely related to cell motility. Int J Dev Neurosci 17:6016111999Int J Dev Neurosci 17:

39.

Noble M: Can neural stem cells be used to track down and destroy migratory brain tumor cells while also providing a means of repairing tumor-associated damage? Proc Natl Acad Sci USA 97:12393123952000Noble M: Can neural stem cells be used to track down and destroy migratory brain tumor cells while also providing a means of repairing tumor-associated damage? Proc Natl Acad Sci USA 97:

40.

Nygaard SJHaugland HKLaerum ODet al: Dynamic determination of human glioma invasion in vitro. J Neurosurg 89:4414471998J Neurosurg 89:

41.

Ohnishi TMatsumura HIzumoto Set al: A novel model of glioma cell invasion using organotypic brain slice culture. Cancer Res 58:293529401998Cancer Res 58:

42.

Paulus WHuettner CTonn JC: Collagens, integrins and the mesenchymal drift in glioblastomas: a comparison of biopsy specimens, spheroid and early monolayer cultures. Int J Cancer 58:8418461994Int J Cancer 58:

43.

Paulus WTonn JC: Basement membrane invasion of glioma cells mediated by integrin receptors. J Neurosurg 80:5155191994J Neurosurg 80:

44.

Pedersen PHNess GOEngebraaten Oet al: Heterogeneous response to the growth factors [EGF, PDGF (bb), TGF-alpha, bFGF, IL-2] on glioma spheroid growth, migration and invasion. Int J Cancer 56:2552611994Int J Cancer 56:

45.

Penar PLKhoshyomn SBhushan Aet al: Inhibition of glioma invasion of fetal brain aggregates. In Vivo 12:75841998In Vivo 12:

46.

Perzelova AMacikova IMraz Pet al: Characterization of two new permanent glioma cell lines 8-mG-BA and 42-mG-BA. Neoplasma 45:25291998Neoplasma 45:

47.

Pilkington GJBjerkvig RDe Ridder Let al: In vitro and in vivo models for the study of brain tumor invasion. Anticancer Res 17:410741091997Anticancer Res 17:

48.

Rooprai HKLiyanage KRobinson SFet al: Extracellular matrix-modulated differential invasion of human meningioma cell lines in vitro. Neurosci Lett 263:2142161999Neurosci Lett 263:

49.

Silbergeld DChicoine M: Isolation and characterization of human malignant glioma cells from histologically normal brain. J Neurosurg 86:5255311997J Neurosurg 86:

50.

Stoppini LBuchs PAMuller D: A simple method for organotypic cultures of nervous tissue. J Neurosci Methods 37:1731821991J Neurosci Methods 37:

51.

Stromblad LGBrun ASalford LGet al: A model for xenotransplantation of human malignant astrocytomas into the brain of normal adult rats. Acta Neurochir 65:2172261982Acta Neurochir 65:

52.

Wester KBjerkvig RCressey Let al: Organ culture of a glioblastoma from a patient with an unusually long survival. Neurosurgery 35:4284331994Neurosurgery 35:

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 68 68 68
PDF Downloads 24 24 24
EPUB Downloads 0 0 0

PubMed

Google Scholar