Rat C6 glioma cell motility and glioma growth are regulated by netrin and netrin receptors unc5B and DCC

Abstract 
Background: Cell migration plays a key role in tumor invasion and metastasis. Deleted in colorectal cancer (DCC) and the unc5 homologues are receptors for secreted netrins that regulate cell and axon migration,


Introduction
Malignant gliomas are primary brain tumors characterized by rapid and invasive growth that progress to highly aggressive glioblastoma multiform (GBM) [1,2]. Their diffuse and infiltrative growth throughout the brain parenchyma makes these tumors difficult to manage by surgery and chemotherapy. An important therapeutic aim is to understand the molecular mechanism responsible for aggressive cell migration, in order to develop treatments that can limit invasion [2].
Tumor development toward aggressive, invasive, and metastasizing phenotypes is a multi-stage process [3] and lost or diminished DCC expression is associated with tumor progression. DCC, a putative tumor suppressor, was identified on the basis of allelic deletions and diverse somatic mutations in this gene in human colorectal neoplasia [4,5]. DCC is down-regulated or lost by various genetic and mutational events in many advanced tumors of different origin, including neuroblastoma [6,7], ovarian cancer [8], breast cancer, prostate cancer, melanoma [9,10], and highly invasive glioblastoma multiform [9,11,12]. Experimental studies support DCC tumor suppressive abilities. For example, ectopic expression of full-length DCC in transformed cells inhibited tumorigenicity [13,14]. Loss of function point mutations in Frazzled, the Drosophila orthologue of DCC, resulted in a cancer-like phenotypes in flies [15]. DCC has been proposed to function as a dependence receptor, activating apoptosis in the absence of netrin-1 [16]. Introducing a point mutation in DCC that disrupts this pro-apoptotic function in mouse increased tumor severity in a genetic background predisposed to the development of intestinal neoplasia [17]. Furthermore, conditional deletion of DCC promoted metastasis in a mouse model of mammary carcinoma [18]. Despite such studies, the underlying mechanism by which DCC inactivation contributes to cancer progression and an invasive phenotype remains poorly understood.
Early studies carried out in C. elegans and the embryonic vertebrate nervous system demonstrated that DCC functions as a netrin receptor to direct cell and axon migration [19][20][21][22]. In vertebrates, DCC, the DCC paralogue neogenin, and the Unc5 homologues, are netrin receptors that regulate chemotropic migration during development [22]. Netrin-1 and netrin receptors are also expressed in many adult tissues, but their contribution to function in adult tissues remains poorly understood [23][24][25][26].
We reported that netrin and DCC influence human glioblastoma cell migration, and that DCC slows the rate of spontaneous cell migration [47]. Here, using the rat C6 glioma cell line, we provide evidence for an autocrine function of netrin acting via Unc5B that promotes cell dispersion, whereas ectopic DCC expression restrains cell movement and intracerebral glioma growth.
Rat C6lacZ glioma cells were modified by double transfection with pUT535 plasmid containing the Escherichia coli lacZ gene encoding β-galactosidase (β-gal) together with pRCRSV encoding the neomycin phosphotransferase gene conferring resistance to G418 [49]. Cells were selected using Geneticin (600μg/ml) (G418; Invitrogen) and periodically monitored using X-gal to ensure continuous lacZ expression. LacZ expression facilitated tumor cell detection in vivo.

Transfection, cell Lysates, conditioned media
Full-length rat DCC from pBS-DCC (clone D4-1, GenBank U68725) was subcloned into pCEP4 (Invitrogen) [50]. For in vivo experiments, C6lacZ cells were transfected with the control pCEP4 vector or the DCC expression construct pCEP4-DCC. For in vitro cell migration assays, C6 cells were transfected with the control pGFP vector encoding green fluorescent protein or the expression construct pDCC-GFP encoding DCC tagged at its C-terminus with GFP [50]. Transfections were carried out using lipofectamine (Invitrogen, Burlington, ON) according to the manufacturer's instruction. Stable cell transfectants (C6lacZ pCEP4 and C6lacZ pCEP4-DCC) were selected using hygromycin at 400 µg/ml. C6 cells transfected with pGFP or pDCC-GFP were selected using Geneticin (G418; Invitrogen). For conditioned media, cells were grown to 80% confluence, rinsed, and replaced with serum-free DMEM that was collected after 48 hrs. For cell lysates, cells were grown in 100 mm tissue culture dishes to 80% confluence, rinsed with PBS, and lysed in 0.6 ml of hot sample buffer (60mM Tris/HCl, pH 6.8, 2% SDS, 10% glycerol, 100 mM DTT) and denatured at 98°C for 5min. Lysates were stored at -20°C until used.

Transfilter chemotaxis assay
Cells were plated on polycarbonate transwell cell culture inserts (6.5 mm diameter with 8μm pore size, Corning) at 5x10 4 cells/100 μl per filter and the filters were placed in 24well plate over 0.6 ml of medium. DMEM with 0.2 % BSA, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM glutamax was the base medium used for all assay conditions. Where applicable, laminin-1 (10μg/ml; BD Biosciences, Bedford, MA), netrin function -blocking antibody (100 μg/ml netrin FB , PN3, rabbit, [23,51]), control rabbit pre-immune IgG (100 μg/ ml, RbIgG; Invitrogen, ON), recombinant human Unc5H3/Fc chimera (5 μg/ml, R&D Systems Inc., Minneapolis, MN), or recombinant netrin-1 protein purified from a 293-Epstein-Barr virus nuclear antigen (EBNA) cell line secreting netrin-1, as described [29], was added to the base medium. Following migration, cells on the upper side of filter were scraped off, and the cells attached to the lower side of the filter fixed with 0.125% glutaraldehyde (25 min, at 4 0 C). Filters were rinsed with PBS and cell nuclei stained with Hoechst dye. Three transwell inserts were used per condition. Four images of each filter were captured using a 10 X objective and nuclei counted using Northern Eclipse software (Empix Imaging, TO). The results are presented as the mean number of cells migrating per condition.

Intracranial injection
Studies using animals were carried out according to the guidelines of the Canadian Council on Animal Care. Stereotactic injections were made as previously described [49]. Six weekold female Sprague-Dawley rats (Charles River, Canada) were anesthetised with sodium pentobarbital (25 mg/ kg) administrated intraperitoneally and the animal's head immobilized in stereotactic apparatus (David Kopf Instruments). An incision was made in the skin to identify bregma on the exposed skull and a burr hole drilled 1mm anterior and 3mm lateral to bregma. Cell suspensions (1x10 4 cells in 5 μl of HBSS) of parental C6lacZ, C6lacZpCEP4 -empty vector control, or C6lacZpCEP4-DCC (expressing DCC) were injected stereotactically over a 10 min period using a Hamilton syringe at a depth of 3.5 mm. Six animals were used per cell type.

Brain tumor analysis
Two weeks after tumor cell injections, rats were euthanatized, brains removed, and quickly frozen in isopentane chilled with liquid nitrogen. Ten μm-thick coronal sections were mounted on glass slides and stained for β-gal activity in the presence of x-gal substrate, as described [49]. Tumor volumes were calculated using the formula a x b 2 x 0.4 (if b < a) or a 2 x b x 0.4 (if a < b), in which a represents the length of tumor from the beginning to the end and b represents the length of the biggest tumor section.

Statistical analysis
All data are reported as mean ± SE. Unpaired Student's t-tests were used for statistical comparisons of the mean values between the data. A value of p < 0.05 denotes statistical significance. Statistical analysis was performed with Prism software (Graph pad).

C6 glioma cells express netrin and netrin receptors
Netrin and netrin receptors are essential for normal neural development [22], and are also implicated in cancer progression in various tumors, including tumors of the brain [6,40,48]. To address the possibility that netrin and netrin receptors regulate glioma cell migration, we first characterized their expression in C6 glioma. Cell culture medium conditioned by C6 glioma cells was analyzed for netrin expression by Western blot using polyclonal PN2 antibodies that bind netrin-1 and netrin-3 [51]. An ~75 kDa band was detected, corresponding to full length netrin protein (Figure 1Aa). RT-PCR analysis detected C6 cell expression of netrin-1 and netrin-3 mRNAs ( Figure 1B). The single band on the Western blot likely corresponds to netrin-1 and netrin-3 due to their similar molecular mass [52]. Both netrin-1 and netrin-3 bind DCC and the Unc5 homologues and exhibit similar functions [22]. DCC protein was not detected in parental C6lacZ cell extracts analyzed by Western blot using a monoclonal antibody against the intracellular domain of DCC (DCC IN ) (Figure 1Ac). This antibody detected a band of ~195 kDa corresponding to DCC in C6 and C6lacZ cell lysates that were transfected with the DCC expressing constructs pDCC-GFP and pCEP4-DCC, respectively. DCC immunoreactivity was not detected in cells transfected with control vectors pGFP and pCEP4 (Figure 1Ac). Anti-DCC IN also detected DCC immunoreactivity in cell extracts of U87 glioblastoma cells, which served as a positive control for DCC expression (Figure 1Ac) [47]. The DCC paralogue neogenin was not detected in C6 glioma cell extracts assayed using Western blotting and the Neo C-20 antibody (not shown). Consistent with this, RT-PCR did not detect neogenin or DCC mRNA expression in C6 glioma cells (Figure 1B).
Western blot analysis of C6 cell lysate with an antibody against Unc5B detected an ~135 kDa band (Figure 1Ab), consistent with previous studies [53]. RT-PCR analysis detected C6 cell expression of Unc5B mRNA ( Figure 1B). Using the respective primers, netrin and netrin receptor expression was demonstrated in adult and embryonic rat brain homogenates as positive controls (Figure 1B).

Netrin and netrin receptors regulate C6 cell migration
These findings establish that rat C6 glioma cells express A. Western blot analyses of proteins in conditioned media and cell lysates: a. Full-length netrin protein (N) (~75 kDa) was detected in conditioned medium with rabbit PN2 antibody, which binds netrin-1 and netrin-3. b. A band corresponding to Unc5B protein (~135 kDa) was detected in whole cell lysate using a goat polyclonal antibody against Unc5B. c. Anti-DCC IN detected full-length DCC protein (~195 kDa) in whole cell lysates prepared from either C6LacZ cells transfected with a plasmid carrying the DCC expression cassette pCEP4-DCC or from C6 cells transfected with a plasmid encoding DCC tagged with GFP at its C-terminus (pDCC-GFP). DCC protein was not detected in whole cell lysates prepared from parental C6 LacZ cells (-) or in the two control cell lines: C6 LacZ pCEP4 and C6 pGFP. DCC protein was readily detected in extracts of U87 human glioblastoma cells, which served as a positive control. Molecular mass markers (kDa) are indicated on the right. B. RT-PCR analysis of total RNA: Total cellular RNA isolated from C6 cells and embryonic rats was subjected to RT-PCR using primers specific for netrins and netrin receptors. Netrin netrin-1, netrin-3 and unc5B, but not DCC or neogenin. To assess the possibility that autocrine expression of netrins might affect C6 motility, cells were allowed to migrate in a transfilter chemotaxis assay as described [47], in the absence of an exogenous cue. Briefly, cells were plated on the upper surface of the filter and allowed to migrate through the porous membrane. Following migration, the residual cells were scraped off the top surface of the membrane and the cells that had migrated to the opposite side of the membrane were fixed, stained and counted. In the presence of PN3 netrin function-blocking antibody (N FB ) added to the top and bottom compartments, C6 migration was significantly slower than cells migrating in medium alone or in the presence of control IgG (Figure 2A). Thus, disrupting endogenous netrin-1 and netrin-3 function decreased the rate of spontaneous migration, consistent with netrin expression by C6 cells exerting an autocrine effect that promotes movement.
C6 cells express Unc5B and were therefore challenged in the transfilter assay by having them migrate in the presence of an unc5H3/Fc chimera (a fusion of a human IgG constant region with Unc5C ectodomain), added to both the top and bottom compartments. This chimeric protein functions as a competitive antagonist of netrin binding to Unc5 homologue netrin receptors. Addition of the Unc5H3/Fc chimera to both the top and bottom compartments of the transfilter assay resulted in significantly slower C6 cell migration (Figure 2B), consistent with autocrine expression of netrin promoting cell motility in an Unc5 homologue dependent manner.
Unc5 homologues have been proposed to function as dependence receptors that trigger cell death in the absence of netrin-1 [54,55]. This raises the possibility that the changes in cell migration detected in the transfilter assay could result from altered cell survival rather than changes in cell motility. However, following disruption of netrin function, cell survival, as measured by cell viability, was not affected, indicating that the results of the transfilter assays are due to altered cell migration (Supplement figure S1).
Transfilter migration assays were also used to assess the chemotropic directional migratory response of C6 glioma cells to a gradient of netrin-1. Addition of netrin-1 to the bottom compartment of the migration chamber significantly reduced the number of cells migrating from the top to bottom through the porous membrane (Figure 2C), consistent with a chemorepellent migratory response. Netrin-1 in the top compartment resulted in an increase in the mean number of cells migrating from top to bottom, albeit a difference that did not reach statistical significance. When exogenous netrin-1 was added to both compartments (top and bottom), the rate of cell migration returned to the control level. These findings are consistent with C6 glioma cells responding to a gradient of netrin-1 as a chemotropic repellent.
To investigate the effect of ectopic DCC expression on C6 glioma cell migration, we introduced the DCC gene into C6 cells by stable transfection with an expression vector encoding DCC-GFP [50]. DCC expression was confirmed by Western blot analysis of cell extracts (Figure 1Ac, clone pDCC-GFP) using anti-DCC IN . Assaying transfilter migration in the absence of an imposed gradient, C6 cells expressing pDCC-GFP migrated significantly more slowly than control C6 parental cells (Figure 2D). Furthermore, directional migration up a gradient of laminin-1 was also significantly reduced compared to the respective controls (Figure 2D), consistent with DCC inhibiting rat glioma cell motility.

Ectopic DCC expression in C6 cells controls glioma formation in rats
Glioma cell invasion resembles glial and neuronal migration during nervous system development, during which DCC guides migrating cells and axons in response to gradients of netrin-1 [1,20,51,56]. DCC and netrin-1 are also expressed in the adult mammalian CNS, but their contribution to adult CNS function is incompletely understood [24, 26,57]. Of interest, reduced expression of DCC and netrin-1 have been observed in highly malignant brain tumors [11,48].
Highly invasive rat C6 cells do not express DCC and we therefore wanted to assess the influence of DCC expression on brain tumor growth. For this purpose, we used C6 cells that were modified to express the E. coli lacZ gene (C6lacZ cells) encoding β-gal to facilitate cell tracking and monitor glioma growth [49]. DCC was introduced into C6lacZ cells by stable transfection with plasmid pCEP4-DCC. Cells were maintained with drug selection and clones expanded and analyzed by Western blot for DCC expression. A clone C6lacZpCEP4-DCC was selected in which an ~195 kDa band corresponding to full length DCC was detected (Figure 1Ac).
To monitor glioma growth in vivo, three cell populations (parental C6lacZ, vector-transfected C6lacZpCEP4, and DCC expressing C6lacZpCEP4-DCC cells) were intracerebrally implanted in rats. All animals developed tumors. Two weeks after cell implantation, sections containing the brain tumors were histochemically stained with X-gal to visualize the tumor cells, allowing evaluation of glioma growth and the extent of spreading into adjacent brain tissue (Figure 3). C6lacZpCEP4-DCC cells generated significantly smaller tumors than the control C6lacZ and C6lacZpCEP4 cells. Tumors composed of DCC-expressing cells exhibited a smooth border and the tumor tissue itself was composed of relatively densely packed cells (Figure 3A). In contrast, tumors arising from control cells that lack DCC grew as a mass of loosely associated cells with diffuse infiltrating borders exhibiting cell migration into brain parenchyma (Figure 3B and 3C). Measurement of tumor volume indicated that the mass of tumors arising from C6lacZpCEP4-DCC cells was reduced by 2-and 4.5-fold, compared to that of control tumors arising from C6lacZpCEP4 and C6lacZ, respectively (Figure 4). These findings provide evidence that DCC expression by C6 glioblastoma cells results in significantly reduced tumor mass and non-invasive borders.

Discussion
Cell migration plays an important role in normal biological processes including embryogenesis, immunity, and wound healing. Pathological changes in neoplasia cause dysfunction in the mechanisms that control cell migration, allowing tumor cells to colonize new sites. Hence, metastatic growth becomes unmanageable by surgery and therapy. Malignant gliomas are typically confined to the CNS and seldom metastasize, but they resist therapeutic treatment and are neurologically destructive due to diffuse single-cell invasion [1]. C6 rat glioma is a highly invasive tumor that is morphologically similar to human glioblastoma multiforme, the most common aggressive and infiltrative brain tumor causing neurological dysfunction and death [56]. Parenchymal invasion is a hallmark of gliomas and a therapeutic intervention that limits cell dissemination could be clinically beneficial. Here we provide evidence that chemotropic guidance molecules regulate C6 glioma cell migration. We have established that C6 rat glioma cells are deficient in DCC expression, but express netrin-1, netrin-3, and the Unc5B netrin receptor. We demonstrate that the inherent motility of C6 cells is influenced by autocrine secretion of netrin, and that disruption of netrin and Unc5B function inhibits cell motility. Spontaneous migration, in the absence of an external gradient, was inhibited by interfering with netrin function, and in the presence of a gradient of netrin-1, C6 cells were repelled.
Netrin-1 can function as a chemorattractant or chemorepellent, in part depending on the receptors expressed by target cells. While DCC contributes to both attraction and repulsion, the Unc5 homologues are only known to signal repulsion [22,[42][43][44]58]. Our findings are consistent with Unc5B being sufficient to evoke a chemorepellent response by C6 Tumors were established by stereotactic injection of cell suspension (1x10 4 cells in 5 μl HBSS) and allowed to grow for 14 days. Brain tissue sections were stained histochemically for β-gal activity (blue) using X-gal followed by H&E staining. Gliomas were derived from C6lacZpCEP4-DCC cells (expressing DCC) (pCEP4-DCC), C6lacZpCEP4 cells (empty vector transfected) (pCEP4), and C6lacZ (parental cells) (parental). Tumors grown from DCC-expressing cells (pCEP4-DCC) appeared as a condensed mass of cells with smooth edges in contrast to control tumors (pCEP4 and parental), which exhibited diffuse growth with cells migrating into surrounding normal brain tissue. Arrows point to the margin of the tumor mass. Stars mark brain parenchyma. (100 X objective lens).
cells to a gradient of netrin-1. We also show that ectopic expression of DCC by C6 glioma cells slowed cell motility. These data are consistent with our previous report that DCC expression by human glioblastoma cells reduced motility [47]. In that study, we provided evidence that DCC promotes the formation of focal adhesions: however, the detailed mechanism underlying the capacity of DCC to restrain cell movement remains to be elucidated.
Here we show that rat C6 glioma cells are attracted up a gradient of laminin-1 and that DCC expression slows the rate of tropic migration. Laminins are vital for tumor propagation and are an effective regulator of cell adhesion and migration [59,60]. As such, DCC may contribute to inhibiting the spread of an invading glial tumor. C6 cells secrete laminins, among other ECM components [56]. Glioma cell migration in the human brain typically follows pathways of ECM proteins that are enriched in the perivascular space and brain parenchyma [59]. Normal brain parenchyma, which is free of laminin [59], can produce laminin in response to invading glioma cells, and laminin deposits are present in the parenchyma in the vicinity of invading glioma cells [60]. Laminin-stimulated migration is integrin-based, and C6 cells express integrin α 3 β 1 [56], a laminin receptor that is frequently expressed in human glioma cells [60]. Interestingly, integrins α 3 β 1 and α 6 β 4 are reported to bind netrin-1 and function as netrin receptors [33]. Furthermore, netrin-1 signaling through DCC activates the integrin-associated kinase FAK to promote cell adhesion and migration [22,61,62], raising the possibility that DCC and integrin signaling may interact.
C6 glioma cells express netrin-1 and netrin-3. Both netrin-1 and netrin-3 bind DCC and Unc5B [22], and it is likely that both proteins similarly contribute to the autocrine influence of netrin on C6 migration. Our findings support the hypothesis that netrins promote the growth of DCC-deficient C6 glioma, perhaps similarly to the growth-supportive activity of netrin-1 in DCC-lacking colon cancer cells [13]. In contrast, DCC expression suppressed C6 cell migration and tumor growth in vivo, inhibiting glioma cell dissemination and identifying DCC as a potential therapeutic target in glioma patients.
DCC signaling regulates cell migration, adhesion, and cytoskeletal dynamics [22]. Here, we have shown that DCCexpressing gliomas grow in vivo as a relatively compact cell mass, consistent with increased cell-cell interactions. Restoration of DCC function may re-engage a mechanism that normally acts to restrain cell motility by promoting cell adhesion. Receptors for ECM adhesion proteins, together with cytoskeletal-linked proteins, are responsible for organized cell migration [63]. DCC association with FAK, a principal tyrosine kinase of focal adhesions, suggests roles for DCC influencing cell attachment and detachment. Netrin-1 signaling through DCC also regulates src family kinases and the Rho family of small GTPases, that direct the dynamic organization of adhesions, the cytoskeleton, and thereby influence cell movement [22,50,61,62,64,65]. During glioma migration, activation of the src family kinase fyn and the GTPase Rac1 by integrin engagement, is regulated by the tyrosine phosphatase PTEN, a tumor suppressor that dephosphorylates FAK and is frequently deregulated in highly malignant tumors [66].
The Unc5 family of netrin receptors signal chemorepulsion in response to netrin-1 [22]. Consistent with the findings we present here, they have also been shown to promote migration and cancer invasion in response to netrin-1 [13,37,38]. Expression of Unc5 homologues has been reported to be lost or reduced in several types of cancer [38][39][40]67], but not others [38]. Unc5B expression is detected in many tumor cell lines, including particularly strong expression in glioma [68]. Bearing in mind that DCC is lost in highly invasive glioblastoma multiform [7,9,12], the response of glioma cells expressing Unc5B may become deregulated, causing them to migrate within the brain, but blocked from disseminating beyond the CNS by the repellent action of netrin-1 and laminin associated with the vasculature.
Major roles for netrins and netrin receptors have been identified during development and tissue morphogenesis [21,31,34,43,69], during tissue regeneration after injury [37], and also in cell survival [34,37,46,54,55]. Coincidental loss of DCC with tumor progression suggests a role for DCC in the maintenance of tissue structure in adulthood. Our findings suggest that loss of DCC expression may destabilize a mechanism that normally restrains cell migration.

Conclusions
In summary, we provide evidence that autocrine secretion of netrins promotes C6 glioma cell migration, and that activation of Unc5 homologue chemorepellent function may direct cells away from an expanding tumor mass. Our findings provide the first demonstration that the expression of DCC in highly malignant glioma cells can limit tumor growth in vivo. DCC reconstitution reduces cell motility and inhibits them from invading the surrounding brain tissue. Glioma treatment is compromised by tumor cell mobility and invasion. Our findings identify possible targets for the development of therapeutics to reduce glioma cell motility.