β1,4-Galactosyltransferase V Functions as a Positive Growth Regulator in Glioma*

β1,4-galactosyltransferase V (GalT V; EC 2.4.1.38) can effectively galactosylate the GlcNAcβ1→6Man arm of the highly branched N-glycans that are characteristic of glioma. Previously, we have reported that the expression of GalT V is increased in the process of glioma. However, currently little is known about the role of GalT V in this process. In this study, the ectopic expression of GalT V could promote the invasion and survival of glioma cells and transformed astrocytes. Furthermore, decreasing the expression of GalT V in glioma cells promoted apoptosis, inhibited the invasion and migration and the ability of tumor formation in vivo, and reduced the activation of AKT. In addition, the activity of GalT V promoter could be induced by epidermal growth factor, dominant active Ras, ERK1, JNK1, and constitutively active AKT. Taken together, our results suggest that GalT V functioned as a novel glioma growth activator and might represent a novel target in glioma therapy.

␤1,4-galactosyltransferase V (GalT V; EC 2.4.1.38) can effectively galactosylate the GlcNAc␤13 6Man arm of the highly branched N-glycans that are characteristic of glioma. Previously, we have reported that the expression of GalT V is increased in the process of glioma. However, currently little is known about the role of GalT V in this process. In this study, the ectopic expression of GalT V could promote the invasion and survival of glioma cells and transformed astrocytes. Furthermore, decreasing the expression of GalT V in glioma cells promoted apoptosis, inhibited the invasion and migration and the ability of tumor formation in vivo, and reduced the activation of AKT. In addition, the activity of GalT V promoter could be induced by epidermal growth factor, dominant active Ras, ERK1, JNK1, and constitutively active AKT. Taken together, our results suggest that GalT V functioned as a novel glioma growth activator and might represent a novel target in glioma therapy.
The carbohydrate moieties of cell surface glycoconjugates play an important role in cell adhesion and metastasis (1). One of the most prominent transformation-associated changes in the sugar chains of glycoproteins is an increase in the large N-glycans of cell surface glycoprotein (2). ␤1,4-galactosyltransferase (GalT) 2 family are the enzymes responsible for the biosynthesis of N-acetyllactosamine on N-glycans by transferring UDP-galactose to the terminal N-acetylglusamine (N-Glc-NAc) residues, and this family consist of seven members, from GalT I to GalT VII (3,4).
GalT V, a member of ␤1,4-galactosyltransferase (GalT) family, could effectively galactosylate the GlcNAc␤136 branch (5), which is a marker of glioma (6). The expression change of GalT V has been investigated using NIH3T3 and the highly malignant transformed cell line MTAg. Northern blot analysis has revealed that the transcript of GalT V gene increases by 2-3-fold in the transformed cells (7). Similar results have been obtained in several human cancer cell lines (8). Consistently, our previous study has shown that the expression of GalT V is increased in the process of glioma development, with the highest level in grade IV glioma (9). Despite this knowledge, currently little is known about the role of GalT V in the process of glioma formation.
The experiments reported here were undertaken to further study the role of GalT V in glioma malignancy, including cell migration, invasion, growth, and survival. Our results indicate that GalT V functioned as a novel glioma growth activator and could represent a novel target in glioma therapy.
Cell Culture and Transfection-Human glioma cell lines SHG44 (10), U87, and U251 were cultured in RPMI 1640 medium or DMEM containing 10% bovine calf serum, 100 units/ml penicillin, and 50 g/ml streptomycin at 37°C in a humidified CO 2 incubator (5% CO 2 , 95% air). Transformed astrocytes C8-D30 (American Type Culture Collection) were cultured in DMEM containing 10% fetal bovine serum, 1.5 g/liter sodium bicarbonate and 4.5 g/liter glucose. SHG44 cells stably transfected with pcDNA3.0 or GalT V antisense cDNA have been described previously (11). HA-GalT V-and HA-tagged mutant format of GalT Vor mock-transfected stable cells were generated by transfection with pcDNA3.0 or HA-GalT V or the HA-tagged mutant format of GalT V, followed by selection in G418. Individual clones were picked and analyzed. Cell transfection was performed with Lipofectamine (Invitrogen) according to the manufacturer's instructions. Cells were harvested and measured by flow cytometry 24 -48 h after transfection, as reported previously (12).
Invasion and Migration Assay-Boyden chamber invasion assay was performed basically as described previously by Albini et al. (16). Cells were added to the upper compartment of the chamber, and 800 l of medium (containing 0.1% bovine serum albumin) was added into the lower chamber. Cells were incubated and allowed to migrate for 24 h. After removal of non-migrated cells, cells that had migrated through the filter were counted under a microscope in five fields at a magnification of ϫ400. Wound healing assay was performed as described previously (17). Briefly, subconfluent cells in 6-well plates were serum-starved overnight. Over 20 wounds were made on the cell monolayer by scratching with a 200-l sterile tip. After being rinsed with phosphatebuffered saline (PBS) three times, the medium was replaced with complete growth medium. Cells were photographed at 0 and 24 h after scraping, and the wound-induced migration of cells was measured after 24 h.
Survival Assay-Cells were plated onto 6-well dishes. After 24 h, the medium was removed, cells were washed twice with Dulbecco's phosphate-buffered saline (DPBS), and the serum-free medium was added. Cultures were visually inspected daily. Cell numbers were determined after trypan blue staining of viable cells in parallel plates (18).
Clonogenic Survival Assay-To determine IR sensitivity, proliferating cells were x-irradiated at the indicated doses using x-ray generator oper- ating at 320 kV/12 mA and allowed to recover for 4 h before trypsinization. The trypsinized cells (5 ϫ 10 3 ) were replanted in duplicate 100-mm dishes and undisturbed for 12 days. The plates were then stained with crystal violet, and the numbers of colonies that were Ͼ1 mm in diameter were counted.
Flow Cytometry of Glycan Level of Membrane Protein and Lectin Blotting-After grown to subconfluence, cells were treated with sialidase for 5 h. Endogenous peroxidase activity was blocked with 0.3% H 2 O 2 for 30 min. After being washed with PBS, cells were harvested after treatment with EGTA (2 mM) and incubated with biotinylated PHA-L or RCA-I (2 g/l) for 45 min at 4°C. Then, the cells were washed with PBS and probed with FITC-conjugated streptavidin (1:128) for 30 min at 4°C. Next, the cell samples were subjected to flow cytometry. Lectin blotting assay was performed as reported previously (13).
Analysis of Apoptosis by Flow Cytometry-Adherent and non-adherent cells were collected, washed twice in PBS, and fixed with ice-cold 70% ethanol for at least 1 h. The fixed cells were washed and stained with propidium iodide. After incubation for 45 min at 37°C, the DNA content was determined by quantitative flow cytometry with standard optics of FACScan flow cytometer (BD Biosciences FACStar). The percentage of apoptotic cell was quantified from sub-G 1 events.
Implantation of Tumor Cells in Mice-Tumorigenicity assay was performed as described previously by Dong Xie et al. (19). At confluence, the cells were harvested, centrifuged, and then resuspended in a sterile solution of PBS at a final concentration of about 1.0 ϫ 10 7 cells/ml. A 100-l aliquot of resuspended cells (about 1.0 ϫ 10 6 cells) was injected subcutaneously between the shoulder blade ϳ3 cm from the tail. After 3 weeks, photographs were taken, and tumors were harvested and individually weighed after mice were anesthetized. Data were presented as tumor weight (mean Ϯ S.D.). Statistical analysis was performed by computer program software using the Student's t test.
Statistics and Presentation of Data-All experiments were repeated three times. All numerical data were expressed as mean Ϯ S.D. Data were analyzed using the two-tailed t test.

RESULTS
We adopted a semiquantitative RT-PCR to analyze the expression of GalT V mRNA in normal brain tissues and glioma tissues. Consistent with our previous study (9), GalT V mRNA expression was correlated with staging to the glioma tumor (Fig. 1A). To elucidate the biological significance of GalT V in glioma, the GalT V antisense cDNA construct was stably transfected into glioma cell line SHG44 (Fig. 1B) (11). In culture, the control SHG44 cells showed invasive growth with spindleshaped morphology (Fig. 1C, left panel) and grew in an actinomorphic manner (Fig. 1D, left panel). However, antisense-transfected SHG44 cells exhibited a round morphology (Fig. 1C, right panel) and grew in a ramble way (Fig. 1D, right panel). To examine the property of GalT V in glioma, we studied the effects of reduction in GalT V expression on the motility of SHG44 cells, assayed by wound healing and Boyden chamber assays. As depicted in Fig. 1, E and F, reduction in GalT V expression resulted in a significant decrease in cell migration in vitro and the ability to migrate through Matrigel-coated 8-m pore size membranes. Similar results were obtained in agarose drop explant assay (data not shown). Moreover, reduction in the expression of GalT V resulted in a total suppression of glioma growth in vivo, compared with the controls (Fig. 1G).
To evaluate the relationship between GalT V overexpression and tumor behavior, transformed astrocytes C8-D30 and glioma cell lines U87 and U251 were stably transfected with HA-GalT V ( Fig. 2A). GalT V overexpression in glioma cells U87 and U251 and transformed astrocytes C8-D30 resulted in a striking increase of cell migration (Fig. 2B), an almost 3-fold increase in vitro invasiveness through a reconstituted Matrigel basement membrane (Fig. 2C), a great increase in colony number (data not shown) and greater numbers of viable cells in serum-free conditions relative to the control cell lines (Fig. 2D). To examine the effect of GalT V overexpression on the ability of tumor formation in vivo, HA-GalT V-or vector-transfected glioma cells were injected subcutaneously into nude mice. As depicted in Fig. 2E, the HA-GalT V-transfected cells developed tumors with a markedly large size during the 3 weeks of observation compared with the control cells. Collectively, these experiments demonstrated that GalT V could promote glioma cell invasiveness and survival.
The GalT V protein consists of a short NH 2 -terminal cytoplasmic domain, a stem region, and a catalytic domain, which contains two conserved residues (Tyr 268 /Trp 294 ) which are important for the galactosylation activity of GalT I (5, 20 -22). To investigate the contribution of these residues in the galactosylation activity of GalT V, HA-tagged point mutants of GalT V (Y268G/W294G) were constructed and transiently transfected into U87 cells (Fig. 3A). As depicted in Fig. 3B, Trp 294 was involved in the galactosyltransferase activity of GalT V. To investigate the contribution of the GalT V galactosylation activity in its tumorigenic effects, the mutation construct W294G was stably transfected into transformed astrocytes C8-D30 and glioma cell lines U87 and U251 (Fig. 3C). As shown in Fig. 3, D and E, the point mutation (W294G) abolished the ability of GalT V to promote the migration ability and invasive potential of glioma cells, indicating that an intact catalytic domain might be essential for GalT V tumorigenic function in glioma. This conclusion was further supported by clonogenic assay and tumorigenicity assay in vivo (data not shown).
The expression of ␤1,6-linked GlcNAc-bearing N-glycans has been reported as a marker of tumor progress in glioma (6). To assess the contribution of GalT V in the expression of N-glycan of cell membrane protein, flow cytometry analysis was performed using FITC-conjugated RCA-I which interacts with oligosaccharides terminating with the Gal␤13 4GlcNAc group or FITC-conjugated PHA-L, which interacts with highly branched N-Glycans with the Gal␤13 4GlcNAc␤13 6-(Gal␤13 4GlcNAc␤13 2)Man branch (15). As expected, reduction in the expression of GalT V decreased the binding with RCA-I or PHA-L on the cell surface (Fig. 4A). Consistent with this, a significant decrease of the binding of total glycoprotein with RCA-I (Fig. 4B, left panel) or PHA-L (Fig. 4B, right panel) was observed for 50 -80-kDa and 80 -100-kDa protein bands in the antisense-transfected cells compared with the mock-transfected cells.
Furthermore, reduction in the expression of GalT V inhibited cell cycle progression (Fig. 5A, left panel) and reduced the endogenous expression of Cyclin D1, Cyclin D3, and E2F1 (Fig. 5A, right panel), which are important regulators of cellular proliferation and highly expressed in glioma (23)(24)(25). In addition, suppression of GalT V expression promoted apoptosis induced by chemotherapeutic agent etoposide (VP16), which was widely used for treatment of malignant gliomas (26) (Fig. 5B, left panel), and inhibited the ability to form invisible clones with x-irradiated treatments (Fig. 5B, right panel). We next examined the contribution of GalT V in the activation of AKT and MAPK kinase, which have shown to associate with glioma invasiveness (18,(27)(28)(29). As shown in Fig. 5C, reduction in the expression of GalT V led to a reduction of the levels of phospho-AKT (Ser 473 /Thr 308 ) and phospho-JNK1/2 (Thr 183 /Tyr 185 ) status. In summary, these data strongly suggest that GalT V could represent a novel target in glioma therapy.
We next explored the possible relationship between Ras/MAPK and PI3K/AKT signaling pathways and GalT V expression. GalT V mRNA expression levels in serum-starved and serum-or EGF-stimulated SHG44 cells were assessed by RT-PCR analysis. As shown in Fig. 6A, endogenous GalT V mRNA expression was markedly induced by serum (Fig. 6A, upper panel) or EGF (Fig. 6A, lower panel), indicating the contribution of serum or EGF in the transcription regulation of the GalT V gene. To adequately address this question, we constructed the GalT V reporter construct pGL3 (Ϫ200/ϩ120), which retained relative strong promoter activity in cancer cells and contained one Sp1-binding site at nucleotide positions Ϫ81/Ϫ69 (15). As expected, transient transfection of the GalT V reporter plasmid pGL3 (Ϫ200/ϩ120) into SHG44 cells showed a dose-dependent reporter gene activity in response to serum (Fig. 6B) or EGF stimulation (Fig. 6C). To determine whether Ras signaling pathway was involved in the GalT V transcription activation, we transiently cotransfected reporter plasmid pGL3 (Ϫ200/ϩ120) and either the dominant negative expression construct Ras-DN or the constitutively active expression construct Ras-DA into SHG44 cells. As expected, expression of Ras-DN decreased the GalT V promoter activity in a dose-dependent manner, whereas expression of Ras-DA caused a similarity dependent activation (Fig. 6D), indicating a role of Ras/ MAPK signaling pathway in the transcription regulation of GalT V. Consistent with this, transient overexpression of ERK1 or JNK1 into SHG44 cells led to a significant increase in the GalT V promoter activity (Fig. 6E, lane 1-5). As the PI3K-AKT pathway represents one of the most potently pro-survival signaling pathways frequently activated in malignant gliomas (27), we explored the effect of AKT on the GalT V promoter activity. As expected, cotransfection of pGL3 (Ϫ200/ϩ120) and constitutively active AKT (Gag-AKT) construct increased the promoter activity in a dose-dependent manner (Fig. 6E, lanes 1, 6, and 7). Sp1, which plays an essential role in the GalT V transcription in cancer cells (15), has been well defined as a nuclear target factor of pro-growth and pro-survival signal transduction pathways in tumor cells, including Ras/MAPK and PI3K/AKT signaling pathways (30,31). To elucidate the contribution of this Sp1-binding site in the activation of the GalT V promoter by Ras/MAPK and PI3K/AKT signaling pathways, we introduced site-directed mutagenesis into this Sp1-binding site (M (Sp1)). It was found that the mutagenesis of this Sp1-binding site abolished the effects of Ras-DA, Ras-DN, ERK1, JNK1, or Gag-AKT on the GalT V promoter activity (Fig. 6F).

DISCUSSION
Malignant gliomas are the most common primary brain tumors and one of the deadliest human cancers (32). They develop as the result of stepwise accumulations of multiple genetic alterations, which result in the activation of oncogenes and/or the inactivation of tumor suppressor The effect of GalT V on the sensitivity of SHG44 cells after exposure to x-rays is shown. Cells were plated, x-irradiated at the indicated doses, allowed to recover, and then replanted and scored for colony formation, as described under "Experimental Procedures." The colony number was standardized with that of cells without x-ray treatment (right panel). C, equal amounts of proteins from mock-(Control) or antisense-transfected SHG44 cells were immunoblotted with the indicated antibodies. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
genes (33). Our early studies have demonstrated that GalT V is highly expressed in glioma. Moreover, the expression of GalT V increases upon malignant transformation of cells and is correlated with staging to the glioma tumor (9). These findings motivate us to examine the function of GalT V protein in glioma.
Results from previous studies have shown that the expression of GlcNAc␤13 6Man-branched N-glycans play a major role in glioma malignant (6). This GlcNAc␤13 6Man structure is synthesized by N-acetylglucosaminyltransferase V, a key enzyme in the processing of asparagine-linked glycans during the synthesis of glycoproteins (34). As the reduction in the expression level of the GalT V gene in SH-SY5Y cells resulted in the decreased galactosylation of highly branched N-glycans, and the expression of GalT V in several human cancer cell lines was highly correlated with that of the N-acetylglucosaminyltransferase V gene (8,15), the GalT V was considered to be involved in the expression of highly branched N-glycans. Consistently, reduction in the GalT V expression in SHG44 cells decreases the galactosyltransferase of the highly branched N-glycans, which leads to the reduced binding to L-PHA. Similar results are obtained in other cancer cell lines including U87 and HeLa cells (data not shown). Therefore, GalT V might be very important in glioma for expressing the highly branched N-glycans that are involved in glioma growth and metastasis (6).
In this study, the role of GalT V in glioma development has been evaluated in several experimental models. Its forced expression in U87 and U251 cells could markedly stimulate their growth and invasion and migration and significantly enhance their tumorigenicity in vivo. These cells developed larger tumors in nude mice. To further address the physiological functions of GalT V in glioma, we investigated the effect of The GalT V promoter construct pGL3 (Ϫ200/ϩ120) was transiently transfected into SHG44 cells. After transfection, cells were treated with the indicated concentration of FBS (B) or EGF (C) for 24 h. The luciferase activity was determined as described above. The values were presented as fold activation over those observed in 1% FBS-treated or non-treated samples. D, SHG44 cells were transiently cotransfected with pGL3 (Ϫ200/ϩ120) and increasing amounts of plasmids expressing the constitutively active form of Ras (Ras DA) or dominant negative form of Ras (Ras DN), and the luciferase activity was determined as described above. E, SHG44 cells were transiently cotransfected with pGL3 (Ϫ200/ ϩ120) and increasing amounts of ERK1, JNK1, or Gag-AKT constructs, and the luciferase activity was determined as described above. F, the promoter constructs pGL3 (Ϫ200/ϩ120) (WT) or M (Sp1) were cotransfected into SHG44 cells with the plasmids expressing the constitutively active Ras (Ras DA), dominant negative Ras (Ras DN), ERK1, JNK1, or Gag-AKT. The luciferase activity was determined as described above. reduction in the GalT V expression on glioma malignant. Reduction in the expression of GalT V resulted in the changes in cell morphology, and decreasing the GalT V expression in glioma cells promoted apoptosis and decreased the migration potential and glioma growth in vitro and in vivo. These findings indicate that GalT V functioned as a positive growth regulator in glioma.
The regulation of glioma cell growth is a complex and intricately orchestrated process involving multiple interrelated signaling pathways. Studies from a number of laboratories have demonstrated that signaling pathways downstream of Ras are critical for facilitating tumor cell proliferation, inhibiting cell death, and maintaining the tumor phenotype (35,36). Similarly, activation of the AKT pathway provides a survival advantage by inhibiting programmed cell death or apoptosis and is important for the maintenance of the tumor phenotype by promoting cell motility and cell shape changes for tumorigenesis (37,38). In an effort to define the mechanism by which the GalT V protein plays a positive growth regulator role in glioma, we analyzed the contribution of GalT V in mitogenic signaling pathways and demonstrated that decreasing the expression of GalT V reduced the AKT activity, indicating that GalT V functioned as a novel glioma growth activator by modulating signaling pathways such as AKT.
Furthermore, Ras/MAPK and PI3K/AKT signaling pathways were involved in the transcription regulation of the GalT V gene. The activity of GalT V promoter could be induced by epidermal growth factor, dominant active Ras, ERK1, JNK1, and constitutively active AKT, suggesting that GalT V might be involved in the pro-survival role of AKT or Ras/ MAPK signaling pathways in glioma. However, the mechanism of the activation of the GalT V promoter medicated by AKT or Ras/MAPK signaling pathways remained unknown and should be explored.
Another important finding of this study was the characterization of GalT V as a novel target in glioma therapy, which was suggested by quantitative evidences. (a) The GalT V mRNA expression was correlated with glioma grade. (b) Reduction in the GalT V expression resulted in a decrease in colony number (data not shown) and glioma growth in vivo. (c) The suppression of GalT V expression inhibited cell cycle progression and reduced the expression of Cyclin D1 and E2F1. (d) The impact of GalT V expression in glioma decreased the relative resistance to apoptosis induced by etoposide or x-ray. (e) Reduction in the GalT V expression decreased the AKT activity which has been inversely correlated with survival in patient glioma specimens (39). These results indicate that decreasing the GalT V expression resulted in the changes in the intracellular pro-survival signaling pathways and decreased glioma cell invasion potential, suggesting that manipulating the expression of GalT V might have therapeutic potential for the treatment of glioma.
The finding that GalT V performed a positive growth regulator in glioma and could represent a novel target in glioma therapy expands our understanding of the proteins involved in gliomagenesis. The molecular mechanism of the transcription regulation of GalT V in glioma should be explored next.