Gap junctions are intercellular channels that connect the interiors of coupled cells. We sought to determine the extent to which malignant glioma cells form gap junction channels with astrocytes from either adult human brain or rat forebrain. The astrocytic gap junction protein, connexin 43 (Cx43), was identified in immunoreactive plaques at areas of cell-to-cell contact between cocultured glioma cells and astrocytes. These gap junction plaques were composed of functional channels, because extensive dye coupling was evident between the glioma cells and astrocytes from both human and rat brain. Calcium signaling was also readily transmitted from glioma cells to astrocytes and vice versa. In live rat brain, injection of glioma cells prelabeled with the gap junction tracer, dicarboxy-dichlorofluorescein, revealed extensive dye transfer to host cells, demonstrating that malignant glioma cells directly couple with normal brain cells. These observations suggest that intercellular communication via gap junctions may play a role in regulating cellular interactions during tumor invasion. In fact, the presence of gap junctions between astrocytes and glioma cells was sufficient to induce a transformation of astrocytic phenotype. Astrocytes cocultured with C6 glioma cells overexpressing Cx43 were significantly smaller and expressed a lower level of glial fibrillary acidic protein than astrocytes cocultured with otherwise identical mock-transfected, gap junction-deficient C6 cells. Thus, direct cellular coupling with glioma cells result in a phenotypic transformation of astrocytes that may contribute to the susceptibility of surrounding tissue to glioma invasion.

Even in early stages, two hallmarks of malignant gliomas are diffuse infiltration and lack of clear demarcation between tumor and adjacent normal brain tissue. The pattern of glioma cell invasion and spread through host brain differs from that of other tumors. Whereas local infiltration of most types of tumor cells occurs primarily through the blood vessels or lymphoid systems, glioma cells can migrate directly into surrounding brain tissue without the guidance of defined anatomical structures (1). Systemic metastasis by gliomas is exceptionally rare (2). Invasion requires adhesion to the intercellular matrix and subsequent degradation by matrix-degrading proteases. An increased level of metalloproteases has been linked to glioma invasiveness (3). However, interaction of malignant glioma cells with surrounding astrocytes has not been established. It is known that astrocytes in contact with gliomas often show reactive changes, but to what extent this is a result of direct interactions between the two cell types remains poorly understood (4). Also, seizure activity has been linked to phenotypic transformation of astrocytes located in the vicinity of glioma cells (5, 6). We sought to determine whether there are direct cellular interactions mediated by gap junctions between malignant glioma cells and astrocytes.

Gap junctions are a subset of cell membrane channels that link the coupled cells (7). Gap junctions are permeable to molecules ≤1.2 kDa, making it possible for second messengers such as cAMP, inositol 1,4,5-trisphosphate and calcium to affect not only their cell of origin but also neighboring cells. As such, gap junctions play a critical role in differentiation, morphogenesis, and pattern formation (8). Gap junctions are composed of homologous proteins called connexins that are encoded by a highly conserved multigene family and expressed in a tissue-specific manner (9). More than 14 connexin genes have been cloned and characterized from rodents, and homologues have been identified in humans, chicks, and frogs (10). In brain, astrocytes are extensively coupled by gap junctions. Electron microscopic studies suggest that single astrocytes express more than 30,000 gap junction channels. Astrocytes may therefore be regarded as a syncytium rather than as individual cells (11, 12). Gap junctions in both astrocytes and glioma cells are composed primarily of the gap junction protein, Cx433(7, 13). In contrast, neurons and oligodendrocytes are coupled sparsely by gap junctions. Gap junctions are required for calcium signaling between astrocytes and possibly between astrocytes and neurons. As such, gap junctions play a central role in both physiological and pathophysiological conditions (14).

Our study demonstrates that gap junctions form readily between malignant glioma cells and astrocytes from both adult human and rat brains in cocultures. The formation of gap junctions between the two cell types was not a culture artifact; glioma cells injected into live rat brain quickly established functional coupling with host cells. Furthermore, gap junction coupling to glioma cells appears to be a regulator of astrocytic phenotype. Astrocytes cocultured with glioma cells expressing Cx43 were significantly smaller and expressed lower levels of GFAP than did astrocytes cocultured with gap junction-deficient, but otherwise identical, glioma cells.

Primary Cultures.

All human specimens were obtained from the Department of Neurosurgery, Westchester Medical Center, New York with approval of the Institutional Review Board. Human malignant glioma cells were cultured by standard procedure. Briefly, the specimens were immersed in HBSS, cut into ∼0.3 mm3 pieces, and trypsinized with 0.25% trypsin-EDTA solution (Sigma Chemical Co., St. Louis, MO) at 37°C for 30 min. An equal volume of culture medium (DMEM/F12, containing 10% fetal bovine serum, 8 mg/ml d-glucose, 20 units/ml penicillin-G, 20 mg/ml streptomycin, and 50 ng/ml amphotericin, all from Life Technologies, Bethesda, MD) was added, and the tissue was triturated to homogeneity. After removal of tissue debris by centrifugation, the cell suspensions were recentrifuged, and the pellet was diluted in 2 ml of warm culture medium before plating. The cultures were kept at 37°C in a 5% CO2 humidified incubator, and the medium was changed every third day.

Human astrocytes were cultured from adult temporal lobes obtained during therapeutic lobectomies for drug-refractory epilepsy. Resected lobes were dissected into neocortical samples. Each tissue sample was cut into roughly 0.3 mm3 pieces, trypsinized, and plated on murine laminin-coated culture dishes as described previously (15). When the outgrowth reached confluence, the cultures were replated and processed for assays. Primary cultures of both astrocytes and glioma cells were discarded after five passages.

Rat mixed forebrain cultures were derived from 16-day-gestation embryos and prepared by standard primary culture procedure as described previously (16). Briefly, embryos were removed from pregnant rats anesthetized with pentobarbital (50 mg/kg; Anpro Pharmaceutical) and decapitated. The forebrains were removed and immersed in HBSS at 37°C. The tissue was then trypsinized and triturated to homogeneity. An equal volume of 10% fetal bovine serum-DMEM/F12 medium was added, and the cell suspension was centrifuged for 10 min at 1000 rpm. A total of 8 × 105 cells were plated on poly-l-lysine- and gelatin-coated 35-mm dishes (Corning) and incubated at 37°C as described above. After 14 days, >95% of the substrate cells stained positively for GFAP.

Cell Lines, Transfection and Selection, and Cocultures.

Rat C6 glioma cells and human malignant glioma cell lines, U87-MG and U373-MG, were obtained from American Type Culture Collection (Rockville, MD). The rat gliosarcoma cell line 9 L was obtained from the Neurosurgical Tissue Bank, UC San Francisco Brain Tumor Research Center (San Francisco, CA).

Cx43 cDNA in expression vector pcDNA1 containing sequence for geneticin resistance was kindly provided by K. Willecke (17). C6 cells were transfected by Clonfectin (Clontech, Palo Alto, CA) according to manufacturer’s instructions, and stable transfectants were selected with 2 mg/ml geneticin. Expression of Cx43 was confirmed by immunolabeling and functional dye transfer assays.

For mixed cultures, glioma cells or astrocytes were prelabeled with the fluorescent cell tracer, 10 μm DiIC18, or 2 μm CMTMR (both from Molecular Probes, Eugene, OR), according to the manufacturer’s instructions. Labeled and unlabeled cells were mixed at a ratio of 1:100 just before plating unless otherwise specified. CMTMR contains a thiol-reactive chloromethyl group which, after reaction with intracellular thiols, becomes membrane and gap junction impermeable. Both cell tracers remain detectable for several cell generations and are resistant to fixation and immunohistochemical procedures (18).

Immunocytochemistry.

A polyclonal antibody against the cytoplasmic COOH-terminal of Cx43 was kindly provided by Dr. Bruce Nicholson (SUNY, Buffalo, NY), and the anti-GFAP polyclonal antibody was from Sigma. Immunocytochemical staining for Cx43 and GFAP was performed as described previously (19). Cells were plated on 12-mm diameter uncoated coverslips put into 24-well culture plates (0.5–1 × 105 cells/well) and cultured for 1–3 days. After fixation in 4% paraformaldehyde for 10 min at room temperature, cells were permeabilized with 0.1% Triton X-100 and blocked with 10% normal goat serum. Incubation with primary antibodies was for 2 h at room temperature or overnight at 4°C and was for 1 h at room temperature with FITC-conjugated goat anti-rabbit antibodies. After several washes in PBS, the coverslips were mounted in Slow Fade (Molecular Probes). Immunofluorescence was visualized by confocal microscopy (MRC1000; Bio-Rad, Hercules, CA). Cell size and immunofluorescence intensity were quantified using an imaging software, Comos (version 7.0a, Bio-Rad). The level of GFAP expression was measured in a representative cytosolic area.

Dye Transfer Assay.

The method of Goldberg et al.(20) was followed with minor modifications. Cells were loaded with 5 μm CDCF (excitation, 488 nm) for 5 min, washed, and trypsinized. After centrifugation, the CDCF-loaded cells were labeled in suspension with 10 μm DiIC18 (excitation, 547 nm) for 10 min and mixed with unlabeled cells at a ratio of 1:250. One hour after plating on polylysine-coated dishes, dye transfer from the CDCF/DiIC18-labeled (donor) cells to unlabeled (recipient) cells was evaluated using confocal scanning microscopy. The coupling index was calculated as the fraction of donor cells that transfer dye to their surroundings times the mean number of recipient cells.

Intercellular Ca2+ Signaling.

Calcium signaling was measured according to the method described by Wang et al.(21). Confluent cultures were loaded for 1 h with 10 μm Fluo-3 acetomethoxyester (Bio-Rad). All experiments were performed at room temperature. A calcium wave was initiated by mechanical stimulation in an astrocyte in the center of the viewing field with a glass micropipette (tip diameter, <1 μm) mounted on a micromanipulator (MMO-220; Narishige). Excitation was provided by the 488-nm line of the krypton-argon laser of a Bio-Rad confocal microscope. Images were acquired every 3–4 s and recorded on an optical disc (LM-D702W; Panasonic). The radii of calcium waves were measured as the maximum distance traveled by a wave from the point of initiation. Velocity was calculated by dividing the maximum distance of wave propulsion (μm) by time (seconds). A calcium wave was defined as a 50% increase (ΔF/Fo) that propelled for a minimum of 50 μm in at least one direction. Background counts were subtracted from all measurements.

Surgical Procedure and Injection of Prelabeled Glioma Cells.

Male Wistar rats (7–180 days) were anesthetized with pentobarbital (50 mg/kg body weight i.p.). Supplemental pentobarbital doses of 15 mg/kg were administered hourly. Rectal temperature was kept close to 37°C by means of a thermostatically controlled heating lamp. The spontaneously breathing animals were placed in a stereotactic head holder. Two burr holes positioned 2.0 mm anterior and 2.5 mm lateral to the bregma were made on the right and left sides of the parietal bone for microinjection of glioma cells. The tip of a microsyringe was inserted 200 μm into the cortical tissue on either side by a micromanipulator. Rat glioma C6 cells (C6-mock transfected, gap junction-deficient), C6-Cx43 cells (Cx43-transfected), or 9 L cells (high expression of endogenous Cx43) were prelabeled with CDCF and DiIC18 and resuspended at a concentration of 5 × 107 cells/ml in serum-free culture medium. The cells were then injected into rat neocortex (3 μl). One hour after cell injection, the rats were decapitated, and their brains were sectioned on a vibratome in 700-μm-thick slices. The extent of dye transfer was evaluated by confocal microscopy as described above. In vitro dye transfer was evaluated concurrently with the in vivo injections. Cell preparations that had <98% labeling efficiency with either of the dyes were discarded.

Cx43 Immunoreactivity and Functional Coupling of Astrocytes and Glioma Cell Cultures.

Astrocytes from both rat and human brains are strongly immunoreactive to Cx43 with a large number of immunopositive plaques at boundaries between adjacent cells (Fig. 1; Table 1). Primary human malignant glioma cell cultures also expressed Cx43, although the expression levels varied as reported by others (13). Established glioma cell lines also varied in their levels of Cx43 expression; the rat gliosarcoma cell line 9 L and the human malignant glioma cell line U87-MG displayed intense immunoreactivity against Cx43, whereas another human glioma cell line, U373-MG, essentially lacked Cx43 immunoreactivity. As shown in Fig. 2, C6 cells derived from rat brain glioma expressed very low levels of Cx43, in agreement with an earlier report (22). After stable transfection with cDNA for Cx43, >95% of cells contained several Cx43-immunoreactive plaques. We used the dye transfer assay developed by Naus and co-workers (8, 20) to assess the extent of functional gap junction coupling. The technique is based on the principle of preloading cells in suspension with both a gap junction-passable tracer, CDCF, and a membrane dye, DiIC18. The double-labeled cells were then mixed with unlabeled identical cells, and the extent of CDCF diffusion from labeled donor cells to unlabeled recipient cells was evaluated after 1 h by confocal microscopy (Figs. 1 and 2). Extensive dye transfer was evident among primary astrocytes from both human and rat brains. All of the primary cultures of human malignant gliomas were coupled, and the extent of coupling was a direct function of Cx43 expression. Both human and rat glioma cell lines varied in the extent of coupling, which was also in direct proportion to Cx43 immunoreactivity (Table 1).

Direct Gap Junction Coupling between Astrocytes and Malignant Glioma Cells.

To test whether glioma cells and normal astrocytes couple directly, we prelabeled glioma cells with either the membrane dye, DiIC18, or CMTMR before mixing with astrocytes at a ratio of 1:100. This approach allowed easy identification of cell types in both live cultures and after fixation. Cx43 immunoreactive plaques were readily identified at areas of cell-to-cell contact between astrocytes and glioma cells (Fig. 3). Connexin-expressing glioma cells, either from rats or humans, were extensively coupled with surrounding astrocytes at varied dye transfer rates (Fig. 3; Table 2).

Gap Junction-dependent Ca2+ Signaling between Astrocytes and Malignant Glioma Cells.

One of the functions of the gap junction is transduction of calcium signaling. To study calcium signaling expressed as long-distance calcium waves, we loaded astrocytic and glioma cell cultures with the calcium indicator, Fluo-3. As reported by others, astrocytes from both rat and human brain produced robust calcium waves after mechanical stimulation, although the maximum radius of wave production was larger in rat cultures than in human cultures (Fig. 4; Table 1). By contrast, calcium waves were produced only within short distances in the primary cultures of human gliomas and in the glioma cell lines studied. Furthermore, as shown in Fig. 5, calcium signal can travel from human astrocytes to glioma cells. Of note, intercellular calcium signaling was, in most cases, communicated bidirectionally between glioma and normal astrocytes with the exception of Cx43-negative U373-MG. U373-MG, despite a marked increase in the calcium content of stimulated cells, failed to induce a significant extent of calcium signaling in neighboring astrocytes (data not shown).

Gap Junction-mediated Transformation of Astrocytic Phenotype in Glioma Cocultures.

To test whether gap junction-coupled gliomas can change the characteristics of cocultured astrocytes, we measured cell size as well as GFAP expression in rat astrocytes cocultured at a ratio of 1:100 with either mock-transfected C6 cells (Cx-deficient) or Cx43-transfected C6 cells (C6-Cx43) or 9 L wild-type cells. C6-Cx43 cells and 9 L wild-type cells expressed a high level of Cx43 and were extensively coupled to primary astrocytes (Figs. 2 and 3; Table 2). In three independent sets of experiments performed in triplicates, astrocytic cells were smaller and their GFAP levels lower when they were cocultured with, and thereby coupled by gap junctions to, C6-Cx43 cells or 9 L cells, compared with astrocytes cocultured with Cx-deficient, C6-mock cells (Fig. 6). These observations indicate that glioma cells, by virtue of gap junction coupling, can transform the phenotype of adjacent astrocytes. Transfer of conditioned medium from C6-Cx43 cells to astrocytes cocultured with Cx-deficient C6 cells (mock-transfected) had no significant effect on either cell size or GFAP expression (data not shown).

Injected Glioma Cells Are Coupled to Host Cells in Live Rat Brain.

We injected Cx-expressing glioma cells (C6-Cx43 cells or wild-type 9L cells) preloaded with CDCF and DiIC18 into the cortex of live rats (Fig. 7) to determine whether glioma cells can form functional gap junctions in intact brain. Burr holes were performed over both hemispheres, and 3 μl of pre-labeled cells in suspension (5 × 107 cells/ml) were injected into the cortical tissue. One hour later, the animal was decapitated, and the extent of dye diffusion was assessed by confocal microscopy. We consistently found extensive CDCF diffusion from prelabeled Cx-expressing donor cells to surrounding unlabeled host cells (C6-Cx43, n = 8; 9L, n = 8), but no CDCF diffusion when Cx-deficient C6-mock cells were injected (n = 4, Fig. 8). Similar CDCF transfer was observed when prelabeled C6-Cx43 or 9 L wild-type cells were injected into acutely prepared cortical brain slices (data not shown).

These observations confirm that functional intercellular communication between astrocytes and malignant glioma cells occurs readily via gap junctions, both in cocultures and in live brain (Figs. 3 and 8). Importantly, gap junction coupling with glioma cells have a profound effect upon astrocytic phenotypes in that astrocytes cocultured with Cx-expressing glioma cells were smaller and expressed lower levels of GFAP than in Cx-deficient cocultures (Fig. 6). Gap junctions are a significant means of intercellular exchange of electrolytes, secondary messengers, and other low molecular weight metabolites. During embryogenesis and tissue differentiation, expression of specific gap junction proteins establishes well-defined compartments and thereby produces complex patterns of communication. Within each compartment, groups of cells are joined by gap junctions to each other and not to cells in other compartments (9). Gap junctions provide a pathway of cytoplasmic continuity so that all of the coupled cells have common access to shared pools of small ions and signaling molecules (23).

Considerable evidence suggests that loss of gap junction communication may be essential in neoplastic transformation. Loewenstein and Kanno (24) originally reported that cancer cells lack gap junction communication, and others subsequently found that the down-regulation of gap junction-mediated intercellular communication might lead to uncontrolled growth of malignant cells as well as metastasis (25, 26, 27, 28, 29). Communication-deficient fibroblast cell lines isolated from Cx43 knock-out mouse embryos showed characteristics of transformed cells including increased growth rate (30). Transfection with various viral oncogenes leads to reduced gap junction coupling that, in several studies, has been linked to phosphorylation of either serine, tyrosine, or threonine on the gap junction protein. For example, the src oncogene product, pp60v-src, reduced coupling by phosphorylating Cx43 at tyrosine 265. By contrast, transfection with the c-erbB2/neu oncogene (neu+) is most likely associated with phosphorylation of serine/threonine residues (31). Conversely, the up-regulation of gap junction communication in communication-deficient cells has been associated with a decrease in cellular growth rates. Growth rates in vitro and in vivo were reduced in rat glioma C6 cells after transfection with a Cx43 cDNA (22, 32). Of note, conditioned medium from Cx43-transfected C6 cells reduced the growth rate of mock-transfected, gap junction-deficient C6 cells (33). Later studies have shown that high levels of connexin expression may not always be associated with decreased proliferation (34). Another group examined Cx43 expression in surgical specimens of astrocytoma in primary cell cultures of the same tumors and in several glioblastoma cell lines. The authors concluded that astrocytomas express Cx43 almost universally but at widely varied levels, especially in high-grade astrocytomas (13). Recently, the growth properties and gap junctional communication of two glioma cell lines, 9 L and C6, were compared (35). The level of Cx43 was an order of magnitude higher in 9L cells, as compared with C6 cells, but no significant differences in their growth rates were noted. Our study supports both these reports in that we found highly variable levels of Cx43 in primary glioma cell cultures and lower levels of Cx43 in C6 versus 9L cells. Most recently, Warn-Cramer et al.(36) reported that Cx43 gap junction communication is regulated in vivo by an EGF-induced signaling pathway that leads to mitogen-activated protein kinase activation which in turn phosphorylates Cx43 at Ser279 and/or Ser282 site(s). Such a phosphorylation is sufficient to disrupt gap junction communication. Malignant gliomas express high levels of EGF receptors and coupling among glioma cells as well as between glioma and astrocytes might thus be regulated by EGF activation of mitogen-activated protein kinase.

Production of matrix-degrading proteases, such as MMPs enables malignant cells to break out of the tumor entity and invade an environment that is regulated by host cells. Most of the increased production of MMPs is from host cells and not from the invading cells (37). The increased production of MMPs has been ascribed to activation of host cells by tumor-produced cytokines and growth factors. Gap junction coupling to glioma cells, or lack thereof, was identified as a potent modulator of astrocytes in this study. In this regard, reactive gliosis, a process by which cells become hypertrophic and increase their expression of GFAP in astrocytes is a prominent feature of peritumor tissue (38). It is note worthy that Cx-deficient C6 cells induced changes in cocultured astrocytes that mimicked reactive gliosis. Astrocytes surrounded by Cx-deficient C6 cells were hypertrophic and displayed increased GFAP expression compared with astrocytes cocultured with Cx-expressing glioma cells. Astrocytes in peritumor tissue are in direct cellular contact with migrating glioma cells. In future studies, we plan to determine whether Cx expression is down-regulated in migrating versus stationary glioma cells. In the developing nervous system, for example, progenitor cells in the proliferating subventricular zone are extensively coupled by gap junctions, but as their newly formed neuronal progenies migrate into the brain parenchyma along radial guide fibers, coupling to surroundings diminishes or disappears. Once their destination is reached, the new neurons establish gap junction connections with their surroundings, as well as synaptic contact (7).

Our tests of calcium signaling disclosed that waves of elevated cytosolic calcium travel within as well as between individual astrocytes that constitutes a newly discovered form of nonsynaptic, long-range signaling in the brain. Several studies have shown that intercellular wave propagation is critically dependent on the coupling of adjoining astrocytes by functional gap junctions, but that an extracellular pathway also may participate (14, 39). The constant velocity of the waves suggests that their production involves a short-range autocatalytic reaction, rather than the long-range diffusion of Ca2+ ions. Regenerative release of Ca2+ from internal stores triggered by Ca2+ itself or inositol 1,4,5-trisphophate has been shown to mediate wave propagation. In our study, calcium waves were elicited in confluent cultures loaded with the calcium-indicator dye fluo-3 by mechanical stimulation. In general, calcium waves propagated over shorter distance in glioma cell cultures compared with astrocytes from either human or rat brain. Nevertheless, calcium signals readily traverse from astrocytes to glioma cells and vice versa, demonstrating that bidirectional communication exists between the two cell types. The extent of such communication was a direct function of the Cx43 level in glioma cells (Table 1). These findings suggest that calcium might be an important signaling molecule involved in the cell-to-cell communication between astrocytes and glioma cells. In fact, Naus et al.(5) found that Cx43 mRNA in peritumor tissue was elevated in patients who have seizures as compared with patients who have no seizures. It is tempting to speculate that calcium signaling between glioma cell and astrocytes plays a role in aberrant electrical activities.

Gap junctions in malignant gliomas have recently been implicated in the “bystander effect.” Proliferating tumor cells expressing herpes thymidine kinase (tk+) die when exposed to acyclovir (40, 41). Several groups have demonstrated that tk tumor cells die after exposure to acyclovir when connected by gap junctions to the tk+-expressing cells but not when gap junctions are absent. Recently, it has been demonstrated that gap junctions play a role in “natural bystander effect.” In a coculture system containing resistant bcl2-expressing cells (bcl2+) and less resistant bcl2-deficient (bcl2) cells, gap junctions propagated and amplified injury (18). Bcl2+ cells died after otherwise nonlethal injury when connected by gap junction to bcl2 cells. bcl2+ cells survived similar injury when cocultured with gap junction-deficient bcl2 cells. A significant distinction is that, whereas acyclovir is an exogenous prototoxin, the mediators of bystander death in a later study were generated by the dying cells as intermediates in the apoptotic process.

The recent explosion of information has established that cell-to-cell coupling is a dynamically regulated pathway for intercellular communication and that gap junctions, on the other hand, control basic cellular functions (42). In cancer research, the traditional concept is that the primary function of gap junctions is to regulate proliferation (43, 44, 45). We have presented data supporting the notion that tumor cells are directly coupled with normal host cells and that this coupling had profound effects on the phenotypic characteristics of host cells. The significance of these observations awaits further studies, but the lesson from other systems is that gap junction coupling has the potential to both shape and regulate the pattern of cellular interactions that may take place during tumor cell invasion.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

      
1

This study was supported by NIH Grants NS130007, NS135011 (to M. N.), and NS01672 (to W. T. C.), and a New York Medical College Intramural grant (to W. Z.).

            
3

The abbreviations used are: Cx43, connexin 43; GFAP, glial fibrillary acidic protein; HBSS, calcium- and magnesium-free Hank’s balanced salt solution; DiIC18, 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocardocarbocyanine perchlorate; CMTMR, 5-(and-6-)-(((4-chloromethyl)benzoyl)amino)tetramethylrhodamine; CDCF, dicarboxy-dichlorofluorescein diacetate; EGF, epidermal growth factor; MMP, matrix metalloproteinase.

Fig. 1.

Cx43 expression in astrocytes and glioma cells from human brains. Left panel, primary cultures were immunoreacted with an anti-Cx43 antibody. Right panel, gap junction permeability visualized by dye transfer assay. A few cells were prelabeled with a fluorescent gap junction-permeable dye (CDCF, green), and a nondiffusable membrane dye, DiIC18 (red) before mixing with unlabeled cells. Transfer of CDCF to unlabeled cells indicates functional coupling. A and B, human astrocytes express abundant Cx43-immunoreactive plaques and are dye-coupled. C and D, an example of one human glioblastoma with high Cx43 expression and extensive dye coupling. E and F, another human glioblastoma culture displaying low level of Cx43 expression and limited functional coupling.

Fig. 1.

Cx43 expression in astrocytes and glioma cells from human brains. Left panel, primary cultures were immunoreacted with an anti-Cx43 antibody. Right panel, gap junction permeability visualized by dye transfer assay. A few cells were prelabeled with a fluorescent gap junction-permeable dye (CDCF, green), and a nondiffusable membrane dye, DiIC18 (red) before mixing with unlabeled cells. Transfer of CDCF to unlabeled cells indicates functional coupling. A and B, human astrocytes express abundant Cx43-immunoreactive plaques and are dye-coupled. C and D, an example of one human glioblastoma with high Cx43 expression and extensive dye coupling. E and F, another human glioblastoma culture displaying low level of Cx43 expression and limited functional coupling.

Close modal
Fig. 2.

Functional gap junction in glioma cell lines, C6 and 9 L. Wild-type 9 L cells have high endogenous expression of Cx43. Transfection of C6 glioma cells with cDNA for Cx43 result in a dramatic increase in gap junction coupling. Left panel, Cx43 expression in C6 cells transfected with Cx43 (C6-Cx43; A), mock-transfected C6 cells (C6-Mock; C), and 9L gliosarcoma cells (wild-type; E). Right panel, C6-Cx43 cells and 9L cells are extensively dye-coupled (B and F), whereas C6-mock cells are not (D). Arrows, prelabeled (donor) glioma cells.

Fig. 2.

Functional gap junction in glioma cell lines, C6 and 9 L. Wild-type 9 L cells have high endogenous expression of Cx43. Transfection of C6 glioma cells with cDNA for Cx43 result in a dramatic increase in gap junction coupling. Left panel, Cx43 expression in C6 cells transfected with Cx43 (C6-Cx43; A), mock-transfected C6 cells (C6-Mock; C), and 9L gliosarcoma cells (wild-type; E). Right panel, C6-Cx43 cells and 9L cells are extensively dye-coupled (B and F), whereas C6-mock cells are not (D). Arrows, prelabeled (donor) glioma cells.

Close modal
Fig. 3.

Cx43 expression and dye coupling in mixed cultures. Left panel, cocultures of malignant glioma cells and astrocytes immunoreacted with an anti-Cx43 antibody. The glioma cells were labeled with DiIC18 and mixed with astrocytes at a ratio of 1:100. Cx43 immunopositive plaques are present at areas of cell-to-cell contact between glioma cells and astrocytes (small arrows). Right panel, gap junction permeability visualized by dye transfer assay in the same set of cocultures as in the left panel. A and B, rat gliosarcoma cell line 9L surrounded by rat astrocytes; C and D, human glioblastoma cell line U87-MG surrounded by human astrocytes; E and F, primary glioblastoma cultures surrounded by human astrocytes.

Fig. 3.

Cx43 expression and dye coupling in mixed cultures. Left panel, cocultures of malignant glioma cells and astrocytes immunoreacted with an anti-Cx43 antibody. The glioma cells were labeled with DiIC18 and mixed with astrocytes at a ratio of 1:100. Cx43 immunopositive plaques are present at areas of cell-to-cell contact between glioma cells and astrocytes (small arrows). Right panel, gap junction permeability visualized by dye transfer assay in the same set of cocultures as in the left panel. A and B, rat gliosarcoma cell line 9L surrounded by rat astrocytes; C and D, human glioblastoma cell line U87-MG surrounded by human astrocytes; E and F, primary glioblastoma cultures surrounded by human astrocytes.

Close modal
Fig. 4.

Calcium signaling in human astrocyte cultures (upper panel), human glioblastoma culture (middle panel), and in a human malignant glioma cell line, U373-MG (lower panel). A calcium wave was initiated by mechanical stimulation (red arrow) and a sequence of images collected 1, 7, and 15 s after stimulation. The color scale indicates relative changes in Fluo-3 signal (ΔF/Fo).

Fig. 4.

Calcium signaling in human astrocyte cultures (upper panel), human glioblastoma culture (middle panel), and in a human malignant glioma cell line, U373-MG (lower panel). A calcium wave was initiated by mechanical stimulation (red arrow) and a sequence of images collected 1, 7, and 15 s after stimulation. The color scale indicates relative changes in Fluo-3 signal (ΔF/Fo).

Close modal
Fig. 5.

Calcium signaling in mixed cultures of glioblastoma (prelabeled with DiIC18) and human astrocytes at a ratio of 1:100. Culture was loaded with Fluo-3 and imaged by confocal microscopy. In the first panel, the DiIC18-labeled glioblastoma cell is visualized in the field of interest (small arrow). The following panels map the Fluo-3 signals in the same field. Images were collected 1, 7, and 15 s after focal mechanical stimulation (arrow). The color scale is the same as in Fig. 4 and indicates relative changes in Fluo-3 signal (ΔF/Fo).

Fig. 5.

Calcium signaling in mixed cultures of glioblastoma (prelabeled with DiIC18) and human astrocytes at a ratio of 1:100. Culture was loaded with Fluo-3 and imaged by confocal microscopy. In the first panel, the DiIC18-labeled glioblastoma cell is visualized in the field of interest (small arrow). The following panels map the Fluo-3 signals in the same field. Images were collected 1, 7, and 15 s after focal mechanical stimulation (arrow). The color scale is the same as in Fig. 4 and indicates relative changes in Fluo-3 signal (ΔF/Fo).

Close modal
Fig. 6.

Gap junction-mediated transformation of astrocytic phenotype. Rat astrocytes were prelabeled with CMTMR and cocultured with: A and D, mock-transfected, gap junction-deficient C6 glioma cells (C6-mock, −Cx); B and E, Cx43-transfected C6 glioma cells (C6-Cx43, +Cx); or C and F, 9 L gliosarcoma cells (wild-type express high level of Cx43, +Cx). After 2 days in vitro, the cultures were fixed and assayed for immunoreactivity against GFAP. Astrocytes were significantly larger and expressed more intense staining against GFAP in gap junction-deficient cultures (A and D), compared with astrocytes in cocultures connected by gap junctions (B and E; C and F). G, histogram summarizing three independent experiments (each in triplicate). More than 200 astrocytes were evaluated in each set of cocultures; Images are 200 × 200 nm. *, P < 0.001, **, P < 0.01.

Fig. 6.

Gap junction-mediated transformation of astrocytic phenotype. Rat astrocytes were prelabeled with CMTMR and cocultured with: A and D, mock-transfected, gap junction-deficient C6 glioma cells (C6-mock, −Cx); B and E, Cx43-transfected C6 glioma cells (C6-Cx43, +Cx); or C and F, 9 L gliosarcoma cells (wild-type express high level of Cx43, +Cx). After 2 days in vitro, the cultures were fixed and assayed for immunoreactivity against GFAP. Astrocytes were significantly larger and expressed more intense staining against GFAP in gap junction-deficient cultures (A and D), compared with astrocytes in cocultures connected by gap junctions (B and E; C and F). G, histogram summarizing three independent experiments (each in triplicate). More than 200 astrocytes were evaluated in each set of cocultures; Images are 200 × 200 nm. *, P < 0.001, **, P < 0.01.

Close modal
Fig. 7.

Protocol used for injection of prelabeled glioma cells in live rat brain.

Fig. 7.

Protocol used for injection of prelabeled glioma cells in live rat brain.

Close modal
Fig. 8.

In vivo dye transfer from glioma cells to host cells. Left panel, cell suspension of CDCF (green)- and DiIC18 (red)-labeled glioma cells observed under confocal microscope before injection. All of the cells were double-labeled. Right panel, extensive transfer of CDCF from prelabeled 9 L gliosarcoma cells to surrounding host cells. In contrast, C6-mock (gap junction-deficient) did not transfer CDCF to host cells. The stellate morphology and organization of labeled host cells (recipient cells) are characteristics of cortical astrocytes. All of the cells were double-labeled and appear yellow because of the merging of red and green.

Fig. 8.

In vivo dye transfer from glioma cells to host cells. Left panel, cell suspension of CDCF (green)- and DiIC18 (red)-labeled glioma cells observed under confocal microscope before injection. All of the cells were double-labeled. Right panel, extensive transfer of CDCF from prelabeled 9 L gliosarcoma cells to surrounding host cells. In contrast, C6-mock (gap junction-deficient) did not transfer CDCF to host cells. The stellate morphology and organization of labeled host cells (recipient cells) are characteristics of cortical astrocytes. All of the cells were double-labeled and appear yellow because of the merging of red and green.

Close modal
Table 1

Gap junction coupling and calcium signaling in glioma and astrocyte culturesa

Cell culturesCx43 plaques/cell (mean ± SE)Dye transfer index (mean ± SE)Calcium wave radius (μm, mean ± SE)
Human    
 Astrocytes 21.1 ± 1.0 5.96 ± 0.52 148.0 ± 2.3 
 U87-MG 5.0 ± 0.6 0.88 ± 0.14 95.0 ± 4.3 
 U373-MG 0.2 ± 0.1 74.5 ± 11.5 
 GBM1 8.7 ± 1.6 1.47 ± 0.26 116.9 ± 9.6 
 GBM2 5.3 ± 0.6 2.95 ± 0.07 114.0 ± 11.0 
 GBM3 19.3 ± 4.0 3.11 ± 0.14 148.2 ± 12.2 
Rat    
 Astrocytes 20.9 ± 1.1 4.55 ± 0.33 229.5 ± 6.8 
 9L gliosarcoma 4.0 ± 0.6 3.43 ± 0.71 71.3 ± 7.26 
 C6 glioma (Mock) 0.1 ± 0.1 0.21 ± 0.13 59.2 ± 6.8 
 C6 glioma (Cx43) 22.3 ± 2.2 4.53 ± 0.65 152.6 ± 5.6 
Cell culturesCx43 plaques/cell (mean ± SE)Dye transfer index (mean ± SE)Calcium wave radius (μm, mean ± SE)
Human    
 Astrocytes 21.1 ± 1.0 5.96 ± 0.52 148.0 ± 2.3 
 U87-MG 5.0 ± 0.6 0.88 ± 0.14 95.0 ± 4.3 
 U373-MG 0.2 ± 0.1 74.5 ± 11.5 
 GBM1 8.7 ± 1.6 1.47 ± 0.26 116.9 ± 9.6 
 GBM2 5.3 ± 0.6 2.95 ± 0.07 114.0 ± 11.0 
 GBM3 19.3 ± 4.0 3.11 ± 0.14 148.2 ± 12.2 
Rat    
 Astrocytes 20.9 ± 1.1 4.55 ± 0.33 229.5 ± 6.8 
 9L gliosarcoma 4.0 ± 0.6 3.43 ± 0.71 71.3 ± 7.26 
 C6 glioma (Mock) 0.1 ± 0.1 0.21 ± 0.13 59.2 ± 6.8 
 C6 glioma (Cx43) 22.3 ± 2.2 4.53 ± 0.65 152.6 ± 5.6 
a

All values represent mean ± SE of a minimal of three (range, 3–28) experiments.

Table 2

Gap junction coupling between glioma cells and astrocyte culturesa

Cell culturesDye transfer index (mean ± SE)
Human  
 Astrocytes/Astrocytes 5.96 ± 0.52 
 U87-MG/Astrocytes 0.88 ± 0.14 
 GBM/Astrocytes 2.55 ± 0.20 
Rat  
 Astrocytes/Astrocytes 4.55 ± 0.33 
 9L gliosarcoma/Astrocytes 3.06 ± 0.85 
 C6 glioma (Mock)/Astrocytes 
 C6 glioma (Cx43)/Astrocytes 3.08 ± 0.50 
Cell culturesDye transfer index (mean ± SE)
Human  
 Astrocytes/Astrocytes 5.96 ± 0.52 
 U87-MG/Astrocytes 0.88 ± 0.14 
 GBM/Astrocytes 2.55 ± 0.20 
Rat  
 Astrocytes/Astrocytes 4.55 ± 0.33 
 9L gliosarcoma/Astrocytes 3.06 ± 0.85 
 C6 glioma (Mock)/Astrocytes 
 C6 glioma (Cx43)/Astrocytes 3.08 ± 0.50 
a

All values represent mean ± SE of a minimal of three (range, 3–10) experiments.

1
Giese A., Westphal M. Glioma invasion in the central nervous system.
Neurosurgery
,
39
:
235
-252,  
1996
.
2
Russel D. S., Rubinstein L. J. Pathology of tumours of the nervous system Ed. 5
423
-429, Edward Arnold London  
1989
.
3
Nakano A., Tani E., Miyazaki K., Yamamoto Y., Furuyama J. Matrix metalloproteinases and tissue inhibitors of metalloproteinases in human gliomas.
J. Neurosurg.
,
83
:
298
-307,  
1995
.
4
Knott J. C. A., Mahesparan R., Garcia-Caberera I., Tysnes B. B., Edvardsen K., Ness G. O., Mork S., Lund-Johansen M., Bjerkvig R. Stimulation of extracellular matrix components in the normal brain by invading glioma cells.
Int. J. Cancer
,
75
:
864
-872,  
1998
.
5
Naus C. C. G., Bechberger J. F., Paul D. L. Gap junction gene expression in human seizure disorder.
Exp. Neurol.
,
111
:
198
-203,  
1991
.
6
Lee S. H., Magge S., Spencer D. D., Sontheimer H., Cornell-Bell A. H. Human epileptic astrocytes exhibit increased gap junction coupling.
Glia
,
15
:
195
-202,  
1995
.
7
Dermietzel R., Spray D. C. Gap junctions in the brain: where, what type, how many and why?.
Trends Neurosci.
,
16
:
186
-192,  
1993
.
8
Naus C. C. G., Bechberger J. F., Bond S. L. Effect of gap junctional communication on glioma cell function Spray D. C. Dermietzel R. eds. .
Gap Junctions in the Nervous System
,
:
193
-202, R. G. Landes Austin, Texas  
1996
.
9
Pitts J. D., Finbow M. E., Kam E. Junctional communication and cellular differentiation.
Br. J. Cancer
,
58
:
52
-57,  
1988
.
10
El Aoumart A., Fromaget C., Dupont E., Reggio H., Durbec P., Briand J. P., Boller K., Kreitman B., Gros D. Conservation of a cytoplasmic carboxy-terminal domain of connexin 43, a gap junctional protein, in mammal heart and brain.
J. Membr. Biol.
,
115
:
229
-240,  
1990
.
11
Nadarajah B., Thomaidou D., Evans W., Parnavelas J. Gap junctions in the adult cerebral cortex: regional differences in their distribution and cellular expression of connexins.
J. Comp. Neurol.
,
376
:
326
-342,  
1996
.
12
Rohlmann A., Wolff J. R. Subcellular topography and plasticity of gap junction distribution on astrocytes Spray D. C. Dermietzel R. eds. .
Gap Junctions in the Nervous System
,
:
175
-192, R. G. Landes Austin, Texas  
1996
.
13
Shinoura N., Chen L., Wani M. A., Kim Y. G., Larson J. J., Warnick R. E., Simon M., Menon A. G., Bi W. L., Stambrook P. J. Protein and messenger RNA expression of connexin43 in astrocytomas: implications in brain tumor gene therapy.
J. Neurosurg.
,
84
:
839
-846,  
1996
.
14
Nedergaard M. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells.
Science (Washington DC)
,
263
:
1768
-1771,  
1994
.
15
Kirschenbaum B., Nedergaard M., Preuss A., Barami K., Frase R. A., Goldman S. A. In vitro neuronal production and differentiation by precursor cells derived from the adult human forebrain.
Cereb. Cortex
,
4
:
576
-589,  
1994
.
16
Nedergaard M., Goldman S., Desai S., Pulsinelli W. A. Acid-induced death in neuron and glia.
J. Neurosci.
,
11
:
2489
-2497,  
1991
.
17
Elfgang C., Eckert R., Lichtenberg-Frate H., Butterweck A., Traub O., Klein R. A., Hulser D. F., Willecke K. Specific permeability and selective formation of gap junction channels in connexin transfected Hela cells.
J. Cell Biol.
,
129
:
805
-817,  
1995
.
18
Lin J. H-C., Weigel H., Cotrina M. L., Liu S., Bueno E., Hansen A. J., Hansen T. W., Goldman S., Nedergaard M. Gap-junction-mediated propagation and amplification of cell injury.
Nat. Neurosci.
,
1
:
494
-500,  
1998
.
19
Cotrina M. L., Kang J., Lin J. H-C., Bueno E., Hansen T. W., He L., Liu Y., Nedergaard M. Astrocytic gap junctions remain open during ischemic conditions.
J. Neurosci.
,
18
:
2520
-2537,  
1998
.
20
Goldberg G. S., Bechberger J. F., Naus C. C. G. A pre-loading method of evaluating gap junctional communication by fluorescent dye transfer.
Biotechnology
,
18
:
490
-497,  
1995
.
21
Wang Z., Tymianski M., Jones O. T., Nedergaard M. Impact of cytoplasmic calcium buffering on the spatial and temporal characteristics of intercellular calcium signals in astrocytes.
J. Neurosci.
,
17
:
7359
-7371,  
1997
.
22
Zhu D., Caveney S., Kidder G. M., Naus C. C. G. Transfection of C6 glioma cells with connexin 43 cDNA: analysis of expression, intercellular coupling, and cell proliferation.
Proc. Natl. Acad. Sci. USA
,
88
:
1883
-1887,  
1991
.
23
Bruzzone R., White T. W., Goodenough D. A. The cellular internet: on-line with connexins.
BioEssays
,
18
:
709
-718,  
1996
.
24
Loewenstein W. R., Kanno Y. Intercellular communication and the control of tissue growth: lack of communication between cancer cells.
Nature (Lond.)
,
209
:
1248
-1249,  
1966
.
25
Eghbali B., Kessler J. A., Reid L. M., Roy C., Spray D. C. Involvement of gap junctions in tumorigenesis: transfection of tumor cells with connexin 32 cDNA retards growth in vivo.
Proc. Natl. Acad. Sci. USA
,
88
:
10701
-10705,  
1991
.
26
El-Sabban M. E., Pauli B. U. Cytoplasmic dye transfer between metastatic tumor cells and vascular endothelium.
J. Cell Biol.
,
115
:
1375
-1382,  
1991
.
27
Mehta P. P., Bertram J. S., Loewenstein W. R. Growth inhibition of transformed cells correlates with their junctional communication with normal cells.
Cell
,
44
:
187
-196,  
1986
.
28
Rose B., Mehta P. P., Loewenstein W. R. Gap-junction protein gene suppresses tumorigenicity.
Carcinogenesis (Lond.)
,
14
:
1073
-1075,  
1993
.
29
Trosko J. E., Ruch R. J. Cell-cell communication in carcinogenesis.
Front. Biosci.
,
3
:
D208
-D236,  
1998
.
30
Martyn K. D., Kurata W. E., Warn-Cramer B. J., Burt J. M., TenBroek E., Lau A. F. Immortalized connexin 43 knockout cell lines display a subset of biological properties associated with the transformed phenotype.
Cell Growth & Differ.
,
8
:
1015
-1027,  
1997
.
31
Hofer A., Saez J. C., Chang C. C., Trosko J. E., Spray D. C., Dermietzel R. c-erbB2/neu transfection induces gap junctional communication incompetence in glial cells.
J. Neurosci.
,
16
:
4311
-4321,  
1996
.
32
Naus C. C. G., Elisevich K., Zhu D., Belliveau D. J., Del Maestro R. F. In vivo growth of C6 glioma cells transfected with connexin 43 cDNA.
Cancer Res.
,
52
:
4208
-4213,  
1992
.
33
Zhu D., Kidder G. M., Caveney S., Naus C. C. G. Growth retardation in glioma cells cocultured with cells overexpressing a gap junction protein.
Proc. Natl. Acad. Sci. USA
,
89
:
10218
-10221,  
1992
.
34
Naus C. C. G., Bechberger J. F., Zhang Y., Venance L., Yamasaki H., Juneja S. C., Kidder G. M., Giaume C. Altered gap junctional communication, intercellular signaling, and growth in cultured astrocytes deficient in connexin 43.
J. Neurosci. Res.
,
49
:
528
-540,  
1997
.
35
Singh M. V., Bhatnagar R., Price C. J., Malhotra S. K. Gap junctions in 9L and C6 glioma cells: correlation with growth characteristics.
Cytobios
,
89
:
209
-225,  
1997
.
36
Warn-Cramer B. J., Cottrell G. T., Burt J. M., Lau A. F. Regulation of connexin-43 gap junctional intercellular communication by mitogen-activated protein kinase.
J. Biol. Chem.
,
273
:
9188
-9196,  
1998
.
37
Edwards D. R., Murphy G. Proteases-invasion and more.
Nature (Lond.)
,
394
:
527
-528,  
1998
.
38
Yong V. W., Tejada-Berges T., Goodyer C. G., Antel J. P., Yong F. P. Differential proliferative response of human and mouse astrocytes to γ-interferon.
Glia
,
6
:
269
-280,  
1992
.
39
Steinhardt R. A., Bi G., Alderton J. M. Cell membrane resealing by a vesicular mechanism similar to neurotransmitter release.
Science (Washington DC)
,
263
:
390
-393,  
1994
.
40
Dilber M. S., Abedi M. R., Christensson B., Bjorkstrand B., Kidder G. M., Naus C. C. G., Gahrton G., Smith C. I. Gap junction promote the bystander effect of herpes simplex virus thymidine kinase in vivo.
Cancer Res.
,
57
:
1523
-1528,  
1997
.
41
Hamel W., Magnelli L., Chiarugi V. P., Israel M. A. Herpes simplex virus thymidine kinase/ganciclovir-mediated apoptotic death of bystander cells.
Cancer Res.
,
56
:
2697
-2702,  
1996
.
42
Paul D. L. New functions for gap junction.
Curr. Opin. Cell Biol.
,
7
:
665
-672,  
1995
.
43
Holder J. W., Elmore E., Barrett J. C. Gap junction function and cancer.
Cancer Res.
,
53
:
3473
-3485,  
1993
.
44
Ruch R. J. The role of gap junctional intercellular communication in neoplasia.
Ann. Clin. Lab. Sci.
,
24
:
216
-231,  
1994
.
45
Yamasaki H., Naus C. C. G. Role of connexin genes in growth control.
Carcinogenesis (Lond.)
,
17
:
1199
-213,  
1996
.