Ciliary Neurotrophic Factor (CNTF) in Combination with Its Soluble Receptor (CNTFRα) Increases Connexin43 Expression and Suppresses Growth of C6 Glioma Cells¹ Free

Normally, astrocytes play a supportive role for neurons by providing crucial metabolites, maintaining favorable extracellular levels of neurotransmitters and ions, and helping constitute the blood-brain barrier (1). Unlike neurons, astrocytes also have the ability to readily divide, which may explain why they have the highest predisposition among CNS³ cells to malignantly transform. In fact, the glioma is the commonest tumor arising in the CNS and accounts for more than 65% of all primary brain tumors (2).

Despite advancements in diagnosis, radiation, and chemotherapy, the prognosis of gliomas has not considerably improved over the past two decades (3). A lack of physical barriers within the CNS likely permits the highly invasive glioma to diffusely infiltrate normal regions of the brain, rendering current therapy relatively ineffective (4). Therefore, management of brain cancer may rely on incorporating novel therapies, including gene therapy.

The C6 glioma cell, initially isolated from N-nitrosomethylurea-induced glial tumors in rats (5), is believed to be astrocytic in origin and displays typical histological features of human glioblastomas, including reduced gap junctional intercellular communication (6).

Gap junctions are proteineous channels that directly link the cytosol of adjacent cells and allow intercellular passage of molecules of M_r <1200 (7). The structural unit of the gap junction is the connexon that itself consists of protein subunits termed “connexins” (7). At least eight different Cxs have been identified in the CNS. Specifically, normal astrocytes are highly coupled by gap junctions and express primarily Cx43 and to a lesser extent Cxs 26, 30, 31.1, and 40 (8).

Loewenstein and Kanno (9) first demonstrated that down-regulation of gap junctional intercellular communication was linked to cancer. Since then, many studies have confirmed that most neoplastic cells have fewer and smaller gap junctions, and express less Cxs than their nonneoplastic counterparts (10, 11). Glioma cells frequently lack functional gap junctions and have reduced Cx43 expression (9, 12). Transfecting glioma cells with exogenous Cx43 has been shown to reverse the transformed phenotype and increase gap junctional communication (13, 14). A potentially less invasive and target-specific cancer therapy may involve increasing endogenous Cx43 levels. Recent studies have shown that endogenous Cx43 expression can be manipulated by cytokine exposure (reviewed in Ref. 15).

CNTF is a “brain injury” cytokine that may have implications in cancer therapy. It is normally found within the cytosol of astrocytes and is released only after a disturbance within the CNS (16). On its release, CNTF may bind its receptor (CNTFRα; soluble or membrane-bound form) and subsequently activate a cascade of signal transduction pathways. One plausible gene to be affected by this CNTF signaling pathway is Cx43. This rationale is based on the following two reasons: first, substantial evidence suggests that Cx43 is up-regulated after a CNS insult (17, 18); second, we have identified a CNTFR-element sequence in the Cx43 gene promoter region located 1523 bases upstream from the first exon site.⁴ Binding of CNTF-induced signal transducer molecules may drive an up-regulation of Cx43, gap junctions and intercellular coupling.

The purpose of this investigation was to determine whether CNTF can be used to increase endogenous Cx43 levels, enhance intercellular coupling, and retard the growth rate of C6 cells. It was found that only when CNTF was administered in the presence of its soluble receptor, CNTFRα, did it induce an increase in both Cx43 and intercellular communication as well as a reduction in the growth rate of C6 glioma cells.

C6 glioma cells, obtained from American Type Culture Collection, were grown in DMEM supplemented with 10% (v/v) fetal bovine serum, 10 μg/ml streptomycin, and 10 units/ml penicillin. Twenty-four h before, and throughout all of the experiments, cells were maintained in serum-reduced media (1% fetal bovine serum). When noted, cells were exposed to the following agents: vehicle (PBS), CNTF (20 ng/ml), CNTFRα (200 ng/ml), or Complex (CNTF + CNTFRα).

C6 cells were washed twice with PBS, and total RNA was extracted using the phenol-chloroform-isoamyl alcohol method outlined by Sambrook et al. (19). RT-PCR of the RNA was performed to create cDNA according to the method obtained with supplies by Invitrogen Corp. (Burlington, ON, Canada). Briefly, 3 μg of RNA was pretreated with DNase and then reverse transcribed in a thermal cycler (Perkin-Elmer, Norwalk, CT) using SuperScript II and oligo(dT) primers for 1 h at 42°C. The cDNA product (1 μl) was mixed with Tris-HCl (pH 8.4; 20 mm), KCl (50 mm), MgCl₂ (1.5 mm), dNTP (200 μm), platinum Taq (2 units), and one set of oligonucleotide primers (200 μm) given in final concentrations in a final volume of 40 μl. Samples were denatured for 5 min at 95°C and then amplified for 30 cycles of 94°C for 45 s, 58°C for 1 min, and 72°C for 1 min. Twenty μl of each PCR sample was run on a 1.8% agarose gel in parallel with a 1.5-kb DNA standard (Invitrogen Corp.).

Primers for CNTF, CNTFRα, and GAPDH (used as a positive RNA-loading control) were used for amplification of the cDNA. Primer sequences were as follows: CNTF forward primer: 5′-GGA AGA TTC GTT CAG ACC TGA C-3′, CNTF reverse primer: 5′-CCC ATC AGC CTC ATT TTC AGG G-3′, predicted product size of 350 bp (20); CNTFRα forward primer: 5′-AGC TGC CGT TCC AAC ACT TAC C-3′, CNTFRα reverse primer: 5′-CCA CAT GCT GCC ATT GAT CC-3′, predicted product size of 430 bp (20); GAPDH forward primer: 5′-AAT GCA TCC TGC ACC ACC AA-3′, GAPDH reverse primer: 5′-GTA GCC ATA TTC ATT GTC ATA-3′, product size of 515 bp (21).

False-positive reactions were ruled out by omitting Superscript II from the cDNA preparation. Primers that produced negative results were subjected to a positive control by performing the RT-PCR on brain tissue (cortex and cerebellum) isolated from an adult CD-1 mouse.

Cells were grown to 80% confluence and subsequently treated with agents every 24 h (with fresh media changes) for 3 days. Cells were then washed twice with PBS and fixed for 10 min with 70% ethanol containing 0.15 m NaCl. After blockage of nonspecific antibody binding using 10% normal goat serum in PBS for 1 h, cells were incubated with rabbit polyclonal anti-Cx43 antibody (1:400 dilution; Sigma-Aldrich, Oakville, ON, Canada) for 1 h, rinsed with PBS and subsequently incubated with Alexa-Fluor-conjugated goat antirabbit IgG secondary antibody (1:500 dilution; Molecular Probes, Eugene, OR). Cells were then washed with PBS, stained with Hoechst 33342 dye (20 ng/ml; Sigma-Aldrich), and mounted with Vectashield medium (Vector Laboratories, Inc., Burlingame, CA). Staining was visualized with a Zeiss Axiophot photomicroscope (Carl Zeiss, Thornwood, NY) and captured using Northern Exposure, version 2 (ImageExperts Inc., Mississauga, ON, Canada).

C6 cells were treated with agents every 24 h (with fresh media changes) for 3 days, rinsed twice with PBS, and scraped off the plates in lysis buffer [radioimmunoprecipitation assay (RIPA) buffer supplemented with protease inhibitors (complete, Mini; Roche, Indianapolis, IN)]. DNA in the lysate was sheared using a 22-gauge needle. Total cell lysate was collected after microcentrifugation at 10,000 × g for 10 min, and total protein concentration was determined using the BCA Protein Assay kit (Pierce-BioLynx, Brockville, ON, Canada). Samples (50 μg) were run in parallel with molecular weight markers (Bio-Rad Lab., Hercules, CA) on a 10% SDS-PAGE. Protein bands were transferred to a nitrocellulose membrane at 80 W for 1 h and subsequently blocked with 5% nonfat dry milk in PBS (with 1% Tween 20) for 1 h. After three PBS rinses (10 min each), the membrane was incubated in anti-Cx43 antibody (1:400 dilution; Sigma-Aldrich) for 1 h, rinsed again with PBS, and then bathed in secondary antibody tagged with horseradish peroxidase (1:20,000 dilution; CedarLane Lab. Ltd., Hornby, ON, Canada). After three PBS rinses, the membrane was incubated in Supersignal (Pierce-BioLynx) and exposed to X-ray film to visualize antibody binding. To normalize protein loading, the membranes were gently stripped of antibodies and immunoblotted for GAPDH (1:20,000 dilution; CedarLane Lab. Ltd.).

Gap junctional coupling of C6 cells was determined by the preloading method as described by Goldberg et al. (22). Briefly, C6 cells were grown to confluence in 12-well plates. Twenty-four h before preloading, a medium change containing the various agents was performed. The agents were present in all of the solutions throughout the experiment. Donor C6 cells were preloaded with dye solution [5 μm calcein-AM (Molecular Probes) and 10 μm DiI (Sigma-Aldrich) in an isotonic (0.3 m) glucose solution] for 20 min in a humidified incubator (37°C, 5% CO₂/95% air). Subsequently, donor cells were rinsed twice with glucose solution, trypsinized, suspended in growth medium, and seeded onto recipient (unlabeled) cells at a 1:500 ratio. After being maintained in the incubator for 3 h, cells were examined with a photomicroscope. Gap junctional communication was assessed by the passage of calcein from donor cells to the underlying recipient cells. Only cells that were coupled to at least one other cell were examined.

C6 cells were seeded into 12-well plates at 2 × 10⁴ cells/well (designated day −1). Cells received fresh media containing the various agents every 24 h commencing on day 0. The number of cells in each well were counted on days 0–4, 6, and 8 using a hemocytometer and trypan blue as a dilutant.

Results are expressed as means ± SE of the means of four or more independent experiments. Statistical comparisons were performed using one-way ANOVA with a P of <0.05 considered significant.

To assess the effects of CNTF and CNTFRα in C6 glioma cells, it was first necessary to characterize endogenous expression of this cytokine and its receptor. To determine whether C6 cells express the mRNA for either CNTF or CNTFRα, RT-PCR was performed on total RNA. Compared with GAPDH levels, a low amount of CNTF mRNA was present in C6 cells, and CNTFRα mRNA could not be detected (Fig. 1). Specificity of the CNTFRα primers was confirmed on adult mouse brain samples (data not shown). Thus, C6 glioma cells are similar to nonreactive astrocytes in their expression of CNTF in the absence of its receptor.

Because many cytokines influence Cx43 expression and gap junctional communication, we sought to determine whether Cx43 could be influenced by the CNTF pathway in C6 cells. Confluent cultures were exposed to vehicle, CNTF, CNTFRα, or Complex every 24 h for 3 days.

Compared with vehicle, CNTF and CNTFRα alone did not alter Cx43 protein expression (Fig. 2,A). In contrast, immunoblot analysis indicated that Cx43 protein levels increased in Complex-treated cells (compared with vehicle; Fig. 2,A). This effect of Complex on Cx43 protein levels predominately occurred by causing an increase in the nonphosphorylated form of the Cx (Fig. 2,A). When protein loading was normalized against GAPDH levels, Complex increased Cx43 levels to 157% that of vehicle, CNTF, and CNTFRα (n = 4; P < 0.01; Fig. 2 B). No significant differences in Cx43 protein expression was detected between vehicle, CNTF, or CNTFRα treatments.

C6 cell cultures are known to contain a mixed population of cells with regard to protein expression (23). Indeed, when examining Cx43 expression by immunocytochemistry, a heterogeneous population of Cx43-expressing cells was identified; some cells showed low Cx43 staining (∼70% of the population) and others showed very low staining (the remaining 30% of the population). Although no detectable changes in the ratios of this heterogeneous population were induced by CNTF, CNTFRα, or Complex, immunocytochemical analysis revealed that Complex induced an increase in Cx43 compared with vehicle, CNTF, or CNTFRα alone (Fig. 3). This increase in Cx43 was detected in both the cytoplasm and at the cell membrane.

Dye coupling of C6 cells was examined to determine whether the Complex-induced up-regulation of Cx43 also increased gap junctional communication. No apparent change in the initial coupling of the preloaded cells was detected between vehicle, CNTF, CNTFRα, or Complex treatments; ∼75% of donor cells were coupled to at least one recipient cell for all of the treatments in the allowed time. However, when the number of recipient cells coupled to a single donor cell was examined, a significant difference between the treatments was observed. Whereas neither CNTF nor CNTFRα alone affected coupling compared with vehicle, Complex significantly increased intercellular dye coupling based on the passage of the gap junction permeable dye calcein (Fig. 4,A). When quantified, Complex significantly increased the number of recipient cells coupled to one donor cell from 4–7 cells (vehicle, CNTF, CNTFRα) to 41 cells (n = 4; P < 0.001; Fig. 4 B).

Induction of Cx43 expression and gap junctional intercellular communication results in reduced growth of many tumor cells, including C6 gliomas. When C6 cells were grown in serum-reduced medium in the presence of the various agents, by day 4, Complex had significantly retarded the growth rate compared with that of vehicle, CNTF, and CNTFRα alone (Fig. 5). This reduction in proliferation induced by Complex became more evident by day 6. By this time, cultures treated with vehicle, CNTF, or CNTFRα alone had reached confluence, whereas Complex-treated cultures had not. This significant decrease in C6 cell growth rate induced by Complex could not be attributed to any toxic effects of the agent because no cell death was be detected by trypan blue exclusion. When cells were examined on day 8, all of the cultures had reached confluence.

C6 glioma cells typically express low amounts of Cx43, show limited gap junctional coupling and have an increased growth rate compared with astrocytes. Modification of aberrant intercellular gap junctional communication seems to be a potential route for cancer therapy. The use of various cytokines, including CNTF, is one method to increase endogenous gap junctions in various cells. We have been able to demonstrate that the C6 glioma cell line expresses the transcript for CNTF but not for CNTFRα. When CNTF was administered in the presence of the soluble CNTFRα, an increase in Cx43 and gap junctional communication was detected in C6 cells, accompanied by a significant decrease in C6 cell growth rate.

The expression of CNTF is almost exclusively restricted to nervous tissue, in which it is normally produced by astrocytes in the CNS and Schwann cells in the peripheral nervous system (16). Positive identification of CNTF mRNA in C6 glioma cells strengthens the theory that this glioma is astrocytic in origin. Furthermore, this finding suggests that examination for CNTF may provide an alternative method in determining whether other gliomas have an astrocytic lineage. This identification technique may be highly useful because the vast majority of astrocytomas, including both cell lines and in vivo biopsies, lack glial fibrillary acidic protein expression, a marker commonly used to identify and distinguish astrocytes from other CNS cell types (24).

In the CNS, CNTFRα expression is restricted to neurons (25) and reactive astrocytes (26). The lack of CNTFRα mRNA in C6 glioma cells reveals two things. First, C6 glioma cells are not immortalized reactive astrocytes, a point which is further strengthened by the diminished or lack of glial fibrillary acidic protein expression. Second, although C6 cells lack CNTFRα, they may still be able to respond to CNTF if coadministered with the soluble form of CNTFRα. Normally, CNTFRα lacks both transmembrane and cytoplasmic domains and is, instead, anchored to the cell surface by a glycosyl-phosphotidylinositol linkage (27). This glycosyl-phosphotidylinositol link can be cleaved by phospholipases and can release CNTFRα to act as a soluble protein. Although soluble CNTFRα is functional, it requires CNTF to be released into the extracellular space (by dying or compromised cells), to which it can bind and subsequently activate the ubiquitously expressed β components of the receptor. Accordingly, Davis et al. (28) have demonstrated that the administration of soluble CNTFRα with CNTF activates signaling in cells that normally do not respond to CNTF alone. Administration of CNTF in the presence of CNTFRα should activate the CNTF signaling pathway in C6 cells and subsequently activate genes containing the CNTF response element sequence in the promoter regions.

Only Complex, and not CNTF nor CNTFRα alone, caused a significant increase in Cx43 in the C6 glioma cells. Impaired expression and posttranslational processing of Cx43 has previously been identified in neoplastic cells (18), including gliomas (29). This increase of Cx43 in C6 cells induced by Complex implies that the CNTF signaling pathway can overcome the Cx43 expression deficiency in these cells. Immunofluorescent staining revealed that a large portion of this up-regulated Cx43 remained within the cell, which indicated that Complex may not be able to fully salvage deficient Cx43 trafficking in C6 cells. Furthermore, most of the increase in Cx43 was detected as the nonphosphorylated form, which further supports a trafficking problem; phosphorylation of Cx43 is believed to occur at the cell membrane after the connexon has been established (reviewed in Ref. 30). An increase in Cx43 by Complex was also identified at the cell membrane, which suggests that increased Cx43 expression enhanced Cx43 trafficking to the membrane and contributed to an increase in gap junctions. This increase in functional gap junctions induced by Complex in C6 cells was confirmed by the extended passage of calcein dye between the cells.

Reduction in C6 glioma cell proliferation by Complex treatment likely occurs by the Complex increasing functional gap junctions. It has been hypothesized that gap junctions allow the intercellular diffusion of growth arrest signals. Supporting this view, it has previously been demonstrated that transfection of C6 cells with exogenous Cx43 results in increased coupling and a concomitant decrease in cell growth (13, 14). It has also been suggested that Cx43 itself has tumor-suppressive properties that are not related to gap junction communication (29). In this study, Complex-induced reduction of the C6 growth rate was more apparent at higher culture densities. This finding suggests that gap junctional coupling played an important role in the decreased growth rate, as has been previously demonstrated (14). Furthermore, the overall findings of this study that Cx43 expression and increased gap junctional coupling decreases the growth rate of glioma cells strengthens the link of aberrant gap junctional communication and cancer.

Several cytokines have been shown to increase gap junctions in one cell type and yet have the opposite effect in another. For instance, whereas transforming growth factor β1 increases Cx43 expression in astrocytes, it decreases Cx43 in C6 glioma cells (31). Therefore, possible use of cytokines to manage gliomas must be considered in light of adverse effects on neighboring normal cells. An ideal cytokine candidate in the CNS would correct aberrant gap junctional communication and have beneficial effects on the sensitive neuronal population. The only known function of CNTF within the CNS is its ability to support the survival of neurons by a mechanism not yet determined. To our knowledge, this is the first study that demonstrates that the CNTF pathway can be used to increases both Cx43 expression and gap junctional coupling, as well as showing a correlation of these effects with a significant decrease in the proliferation of glioma cells.