Experimental and Molecular Pathology 93 (2012) 8–17
Contents lists available at SciVerse ScienceDirect
Experimental and Molecular Pathology
j ourna l homepage: www.e lsev ie r .com/ locate /yexmp
The role of N-acetylglucosaminyltransferases V in the malignancy of humanhepatocellular carcinoma
Ting Wei a,1, Qiulian Liu a,1, Fuli He b, Weiliang Zhu a, Lijuan Hu a, Linlang Guo a, Jian Zhang a,⁎a Department of Oncology, Zhujiang Hospital, Southern Medical University, Guangzhou 510282, Chinab Department of Oncology, Qianxinanzhou People Hospital, Xingyi, Guizhou Province, 562400, China
⁎ Corresponding author at: Department of OncologyMedical University, 253 Gongye Road, Guangzhou 5102
E-mail address: [email protected] (J. Zhang).1 Contributed equally to this work.
To investigate the role of N-acetylglucosaminyltransferases V (GnT-V) in the malignancy of human hepato-cellular carcinoma (HCC), the GnT-V stably suppressed cell line HepG2 GnT-V/1564 was constructed fromHepG2. The proliferation, migration, invasion, metastasis of HepG2 GnT-V/1564 was investigated both invitro and in vivo. The clinical pathological significance of GnT-V expression was also studied in 140 cases ofHCC tissues. This study showed that down-regulation of GnT-V inhibited the proliferation, migrationand invasion of the HepG2 cells. In addition, GnT-V expression was shown in 138 cases of 140 (98.6%)HCC samples, in 3 cases of 31 (9.7%) in liver cirrhosis cases and in 1 cases of 20 (5.0%) in normal livertissues. Besides, a higher level of GnT-V was observed more frequently in the advanced tumors withhigher T stage and histological grade. These data suggested that GnT-V expression was positively relatedwithmalignancy inHCC and GnT-Vmay be both a differentiationmarker and a potential target for the treatmentof HCC.
Most proteins in an organism exist in the form of a glycoprotein,and the sugar chains of a glycoprotein directly affect protein function(Apweiler, et al., 1999). It is well known that sugar chains have a va-riety of functions and play a key role in cell growth, differentiationand adhesion (Ohtsubo and Marth, 2006). Structural changes inN-acetylglucosamine on asparagine (N)-linked oligosaccharides(N-glycan) have been reported to be critical in tumor progression.Although the underlying mechanism is still indefinite, the activationof glycosyltransferases plays a major role in the structure of sugarchains, leading to “aberrant glycosylation” in cancer tissues (Asada,et al., 1997; Hakomori, 1989).
GnT-V, one of the most important glycosyltransferase members, isessential for N-linked glycoprotein synthesis and is highly involved incancer invasion and metastasis. GnT-V catalyzes the formation ofGlcNAc-β1-6 branches at the Man α1-6 side of the trimannosyl coreof N-glycans (Chakraborty and Pawelek, 2003; Taniguchi, et al.,1999). These branches are abnormal in cancer tissues, especially inthose with high metastatic potential (Fernandes, et al., 1991;Handerson and Pawelek, 2003; Pierce, et al., 1997). In breast cancer
, Zhujiang Hospital, Southern82, China.
rights reserved.
cells MDA-MB231, down-regulation of GnT-V could reduce the expres-sion of N-linked (1,6)-branching on epidermal growth factor receptors,resulting in suppressed invasion (Guo, et al., 2007).Mammary tumor pro-gression caused by the polyomavirus middle T oncogene was diminishedin null mice that lacked GnT-V (Granovsky, et al., 2000). The beta1,6-branched oligosaccharides catalyzed by GnT-V in primary breast tu-mors were a predictor of poor outcome (Handerson, et al., 2005).These results suggested that GnT-V gene may be an oncogene. How-ever, some studies showed that low GnT-V expression is associatedwith shorter survival and poor prognosis in non-small cell lung cancerand bladder cancer (Dosaka-Akita, et al., 2004; Ishimura, et al., 2006),which suggested that GnT-V may be a potential cancer suppressorgene. Thus, GnT-V expression and its function in human cancer clear-ly remain controversial.
HCC is one of the most malignant tumors in the world, with anestimation of 600,000 new cases per year (Tanaka and Arii, 2009).To further investigate the effect of GnT-V expression on malignancyincluding proliferation, migration, invasion, metastasis, malignanttransformation and progression of the HCC, the expression vectors ofshort hairpin RNA (shRNA) targeting the GnT-V gene were constructedand transfected into HCC HepG2 cells. The GnT-V stably suppressed cellline HepG2 GnT-V/1564 was developed, and its malignancy was investi-gated both in vitro and in vivo. Furthermore, the possible implications ofGnT-V expression in HCC tissueswere accessed from the clinicopatholog-ical point of view. The tissuemicroarrays (TMAs)was used in this study todetect the relationship between GnT-V and clinical pathological featuresin 140 cases of HCC tissues, 31 cases of liver cirrhosis and 20 cases ofnormal liver tissues.
9T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
Materials and methods
Construction of pGPU6/GFP/Neo GnT-V shRNA
GnT-V shRNA sequence designAccording to siRNA design principles and using the “siRNA Target
Finder and Design Tools” from Ambion company, two RNA fragmentsaimed at GnT-V cDNA (NM_002410.3) were designed. The sequenceswere sense 5′-GGAAGTGCATGCAACTGTTTA-3′ (GnT-V/1564) andsense 5′-CTCCTTTGACCCTAAGAAT-3′ (GnT-V/2224) respectively.The control sequence was sense 5′-GTTCTCCGAACGTGTCACGT-3′(GnT-V/NC), which lacked more than 70% homology to either GnT-V/1564 or GnT-V/2224 by Blast search.
shRNA transcriptional template DNA designAccording to the planned siRNA sequence, the template DNA was
designed as follows: BbsI restriction site, sense sequence, 9 nt loopsequence, antisense sequence, RNA polymerase III termination (6nucleotide poly T) and BamHI restriction site (Table 1).
Construction of the pGPU6/GFP/Neo GnT-V plasmidTo construct the recombinant plasmid, the two shRNA template
oligonucleotides of each groupwere annealed followed by connectionto the pGPU6/GFP/Neo vector. Then the E. coli DH5αwas transformedwith the recombinant plasmid and the Kanamycin resistant colonieswere selected. The plasmid was purified for transfection, named aspGPU6/GFP/Neo GnT-V/1564, pGPU6/GFP/Neo GnT-V/2224 andpGPU6/GFP/Neo GnT-V/NC respectively. To confirm whether plasmidswere constructed successfully, the recombinant plasmid would bedigested by BbsIand BamHI, and the products were checked by electro-phoresis. The identification of recombinant plasmid sequence analysiswas undertaken by Shanghai GenePharma Co., Ltd.
Cell culture and transfection
The high metastatic potential human hepatocellular carcinomaHepG2 cells (cell bank of Sun Yat-Sen University) were cultured inRPMI-1640 containing 10% newborn cow serum at 37 °C with 5% CO2.All of the constructed plasmids were transfected into HepG2 cells byLipofectamine 2000 (Invitrogen, California, USA). The oligofectaminereagent was composed of 0.5 μg/μl plasmids 2 μl, Lipofectamine 20002 μl and Opti-MEM 96 μl. Transfected cells were selected for 15 daysunder G418 (500 μg/ml).
Examination of interference effects on GnT-VmRNA and protein expression
The expression of GnT-VmRNAwasmeasured by semi-quantitative RT-PCRTotal RNA was isolated from cells and reversely transcribed into
cDNA by using random hexamer primers. GnT-V cDNA was amplifiedby using primers (Shanghai Sangon Biotech Co Ltd) as follows: forward5′-AACTCTTGGACCATCCTGGGTTC-3′, reverse: 5′-TTGCTGCTTTTGGGT-GGGTT-3′, which generated a 555 bp product. To normalize the efficiencyof the amplification, β-actin cDNA was amplified by using primers asfollows: forward: 5′-GAAACTACCTTCAACTCCATC-3′, reverse: 5′-CGA-GGCCAGGATGGAGCCGCC-3′, which generated a 219 bp product. Bothproducts were amplified under the conditions: 36 cycles of 94 °C for
Table 1Information on GnT-V siRNA sequence.
Sequence information
1564 Sense 5′CACCGAntisense 3′CC
2224 Sense 5′CACCGAntisense 3′CG
NC Sense 5′CACCGAntisense 3′CA
30 s, 60 °C for 30 s and 72 °C for 30 s. The bands were quantified byQuantity One software. The experiment was repeated for three times.
GnT-V protein expression was measured by western blot analysisBriefly, proteins were electrophoresed on an 8% polyacrylamide gel
and transferred to immobilon polyvinyldifluoride (PVDF) membranes.The membranes were blocked with 3% BSA+7% fat-free dry milk for1 h at room temperature and then treated with goat anti-human anti-body (Santa Cruz Biotechnology, Inc, California, USA) against GnT-V(1:200) and mouse anti-human antibody against GAPDH (1:1000,Beijing Biosynthesis Biotechnology Co. Ltd) for 1 h at room tempera-ture, respectively. Then, the membranes were subsequently incubatedwith a peroxidase conjugated rabbit anti-goat secondary antibody(1:1000) and a peroxidase conjugated rabbit anti-mouse secondary an-tibody (1:5000) for 1 h at room temperature, respectively. The proteinbands were visualized by enhanced chemiluminescence using Fujisuper RX film (Fuji film, Tokyo, Japan) and quantified by Quantity Onesoftware. The experiment was repeated for three times.
The effects of pGPU6/GFP/Neo GnT-V shRNA expression on HepG2proliferation, migration and invasion of the HepG2 cell line in vitro
Cell Counting Kit-8 (CCK-8) assayCell proliferation was evaluated by CCK-8 assay. Cells were seeded
onto the 96-well plates. CCK-8 solution was added to the 96-wellplates to analyze cell activity at different times (0 h, 24 h, 48 h, 72 h,and 96 h). The OD450 nm value of each group was measured to cal-culate the cell growth rate and growth inhibition rate. Growthrate=(absorbance at 450 nm at x h−absorbance at 450 nm at0 h) / (absorbance at 450 nm at 0 h)×100. Cell growth inhibitoryrate (IR)=(OD450 value of HepG2 GnT-V/NC−OD450 value ofHepG2 GnT-V/1564)/OD450 value of HepG2 GnT-V/NC×100. Thegrowth inhibitory rate represented the decreased proliferation abilityof HepG2 GnT-V/1564 relevant to HepG2 GnT-V/NC. The experimentwas repeated for three times and the number of replication at eachtime was six.
Cell wound healing assayThe cells were grown to monolayer and scraped by 100-μl plastic
tips. Pictures of the scraped wound were taken in 5 randomly chosenfields under light microscopy at 0 h, 12 h, 24 h and 36 h. Wound curerate=(distance of wounds before healing−distance of wounds afterhealing)/distance of wounds before healing×100. The experimentwas repeated for three times and the number of replication at eachtime was six.
Cell invasion assayTheMatrigel gel (50 μl) as a basementmembranewas spread on the
polycarbonate membrane of 24-well transwell units. The cells wereadded to each upper compartment of the transwell units. The numbersof cells that had penetrated through theMatrigel gel and polycarbonatemembranes were calculated by counting 5 randomly chosen fieldsunder light microscopy. The experiment was repeated for three timesand the number of replication at each time was six.
Fig. 1. (A) The recombinant plasmids were transfected into the HepG2 cells by Lipofectamine 2000TM. The transfected cells were screened in RPMI-1640 containing G418 at 500 μg/ml.The stably transfected cells emitted a green fluorescence that could be detected by fluorescence microscopy (400×). (B, C) The mRNA expression of GnT-V in the three cell groups wasevaluated by semi-quantitative RT-PCR. As shown in the pictures, GnT-V shRNA reduced the expression of GnT-V mRNA in HepG2 GnT-V/1564 and HepG2 GnT-V/2224 by (81.67±1.52) % and (62.00±3.00) %, respectively. There was no statistical difference in GnT-V mRNA expression between HepG2 and HepG2 GnT-V/NC. (D, E) The protein expression of GnT-Vin the three cell groups was analyzed by Western blot assay. As illustrated in the pictures, GnT-V shRNA inhibited GnT-V protein expression in HepG2 GnT-V/1564 and HepG2 GnT-V/2224 cells by (72.67±1.53) % and (55.33±4.16) %, respectively. There was no statistical difference in GnT-V protein expression between HepG2 and HepG2 GnT-V/NC.
10 T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
Tumorigenicity and tumor metastasis in vivo
Tumorigenicity in the nude miceThe experimental protocol was approved by the International
Ethics Review Committee for Animal Experimentation. HepG2 GnT-V/NC cells (2.5×106) were subcutaneously injected into the left glutealregion of 6 male BALB/c nude mice (6-week-old), and HepG2 GnT-V/1564 cells (2.5×106) were subcutaneously injected into the right glutealregion of the same 6 mice. The width and length of tumors weremeasured with calipers and the tumor size was calculated using thefollowing equation: tumor volume=width2×length/2. The mice werekilled after injection at the 21st day and the tumors were dissected and
weighed. Then, the transplanted tumors were studied by hematoxylinand eosin staining (H&E) and immunohistochemistry staining as below.
Establishment of animal models of HepG2 xenografts transplanted in thespleen and metastasized to the liver in the nude mice
The nude mice were anesthetized with methoxyfluorane andplaced in the right lateral decubitus position. A transverse incisionwas made in the left flank through the skin and peritoneum, exposingthe medial aspect of the spleen. 2.5×106 cells were injected into themedial splenic tip with a 30-gauge needle. The spleen was returned tothe abdominal cavity, and the wound was closed with wound clips.Three weeks after the injection of tumor cells, the animals were killed
Fig. 2. Cell proliferation was evaluated by CCK-8 assay. Cells were seeded onto 96-wellplates. CCK-8 solution was added to 96-well plates to analyze cell activity at differenttimes (0 h, 24 h, 48 h, 72 h, and 96 h). (A) The cell growth curve and (B) growth inhibitionrate curve showed that the proliferation ability of HepG2 GnT-V/1564 was significantlyinhibited compared with HepG2 GnT-V/NC.
11T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
by cervical dislocation and the organs were kept in formaldehyde solu-tion. The character of metastatic nodules in the liver was identified byH&E and immunohistochemistry as below. The numbers of metastatictuberculums large enough to be visible to the naked eye on the surfaceof livers were measured. Six mice were used in each group.
Human tissue microarrays (TMAs)
TMAs were obtained from Cybrdi, Shanxi ChaoYing BiotechnologyCo., Ltd. Two random blocks of each cancer tissue were selectedaccording to products information from Cybrdi Co., Ltd. The TMAsconsisted of 140 HCC cases with TNM stage (26 cases of T1, 46 casesof T2, 48 cases of T3, 20 cases of T4) and histological grade (36 welldifferentiated, 70 moderate differentiated, 34 poor differentiated),31 liver cirrhosis cases and 20 normal liver tissues. The expression ofGnT-V in TMAs was detected by Immunohistochemistry staining asbelow.
Histopathology
Hematoxylin and eosin staining (H&E)All the tissues were initially fixed with 10% formalin, dehydrated
in a graded alcohol series, and then embedded in paraffin. Sectionswere cut from each paraffin block, deparaffinized, stained with hema-toxylin (2 min), washed (H2O), differentiated in HCL/ethanol, rinsed(H2O), turned blue in warm tap water, stained with eosin (1 min),shortly rinsed (H2O), dehydrated in ascending ethanol concentrations,transferred to xylenes and finally embedded with Pertex® routineembedding medium (Medite, Germany) under cover slides.
Immunohistochemistry stainingSpecimens were deparaffinized and rehydrated routinely. Before
adding the primary antibody, antigens were retrieved by heatingsections in 10 mM citrate buffer (pH 6.0) in a microwave oven for10 min followed by 10 min of cooling. After blocking nonspecificbinding with 0.3% H2O2 and goat serum, the slides were incubatedwith a primary antibody directed against GnT-V (1:200, Santa CruzBiotechnology, Inc, California, USA), and subsequently incubated with aperoxidase conjugated rabbit anti-goat secondary antibody (1:1000,Beijing Biosynthesis Biotechnology Co., Ltd). Reaction products werevisualized by 3′-diaminobenzidine (DAB), and slides were subsequentlycounterstained with hematoxylin. Negative controls were performed byreplacing the primary antibody with PBS. Positive control sample wasfrom colorectal cancer for GnT-V. Brown–yellow granules in cytoplasmwere considered positive staining. The percent positivity was scored as0 if b5% (negative), 1 if 5–30% (sporadic), 2 if 30–70% (focal) and 3 if>70% (diffuse) of cells stained, whereas staining intensity was scoredrelative to the known positive and negative controls as 0 if no staining,1 if weakly to moderately stained and 2 if strongly stained. The GnT-Vexpression levels were classified semiquantitatively based on the totalscores of the percent positivity and the staining intensity as follows:‘negative’ if the sum scored 0, ‘GnT-V low’ if the sum scored 1 (+),‘GnT-V moderate’ if the sum scored 2–3 (++) and ‘GnT-V high’ if thesum was 4–5 (+++) (Takahashi, et al., 2009). The scoring procedurewas taken by two independent observers without any knowledge ofthe clinical data.
Statistical analysis
Results were analyzed by SPSS13.0 software. Statistical analysisbetween two sampleswas performedusing the Student's t-test. Statisticalcomparisons of more than two groups were performed using one-wayanalysis of variance (ANOVA). The association betweenGnT-V expressionand clinicopathological parameters was evaluated using the x 2-test.Pb0.05 was considered as significant difference.
Results
Construction of recombinant plasmid
To determine whether the recombinant plasmids were successfullyconstructed, the BbsIand BamHIdouble digestion assay was used. Theresult showed that GnT-V shRNA expression plasmidswere constructedsuccessfully. The above result was also confirmed by sequence analysis.
Development of down-expressing GnT-V HepG2 cells
The stably transfected HepG2 cells containing the green fluorescentprotein (GFP) gene sequence could emit a green fluorescence by fluo-rescence microscopy detection, and named as HepG2 GnT-V/NC,HepG2 GnT-V/2224, and HepG2 GnT-V/1564, respectively (Fig. 1A).
The GnT-V mRNA expressions in the three groups were evaluatedby semi-quantitative RT-PCR. The expression of GnT-V mRNA inHepG2 GnT-V/1564 and HepG2 GnT-V/2224 cells was decreased by(81.67±1.52) % and (62.00±3.00) % comparedwith that of the controlgroup, respectively (Figs. 1B, C). The protein expression of GnT-V in thethree cells was analyzed by Western blot assay. The GnT-V expressionin HepG2 GnT-V/1564 and HepG2 GnT-V/2224 cells reducedsignificantly by (72.67±1.53) % and (55.33±4.16) % compared withthat in control group, respectively (Figs. 1D, E). There was nostatistical difference of GnT-V expression between HepG2 and HepG2GnT-V/NC (P>0.05) (Figs. 1B, C, D, E). The HepG2 GnT-V/1564 cellexhibited the higher knockdown of GnT-V and was chosen for the fur-ther assays.
Fig. 3. Cells were grown in monolayer. Linear scrape wounds were made on the cell monolayer. Pictures of wound healing were taken at 0 h, 12 h, 24 h and 36 h, respectively(200×). The pictures (A) and cure curve (B) showed that the down-regulation of GnT-V expression can significantly inhibit the wound healing rate for HepG2 cells.
AHepG2GnT- HepG2 GnT-
B
/1564 /NC
Fig. 4. (A, B) The cells were added to each upper compartment of the transwell units. The number of cells that penetrated through the Matrigel gel and polycarbonate membraneswas calculated by counting 5 randomly chosenfields under lightmicroscopy (400×). The results showed that the numbers of penetrating cells were (11.00±1.86) and (47.35±1.98) forHepG2 GnT-V/1564 and HepG2 GnT-V/NC cells (t=59.770, Pb0.001), respectively.
12 T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
Fig. 5. HepG2 GnT-V/NC cells (2.5×106) were subcutaneously injected into the left gluteal region of 6 nude mice, and HepG2 GnT-V/1564 cells (2.5×106) were subcutaneouslyinjected into the right gluteal region of the same 6 mice. The mice were killed after 21 days and the transplanted tumors in the left and right gluteal region were dissected andweighed. (A) The left arrows represented transplanted tumors from HepG2 GnT-V/NC, the right arrows represented transplanted tumors from HepG2 GnT-V/1564. (B) The estab-lishment of transplanted tumors was confirmed by H&E (200×). As arrows showed that cores or nest of cells and hepatic sinusoid was observed in most fields of photomicrographs,and the nucleus was big, round or oval, with pathological karyokinesis. (C) The tumors from HepG2 GnT-V/NC grew quickly than those in HepG2 GnT-V/1564.
13T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
Effect of GnT-V shRNA expression on cell proliferation,migration and invasionin vitro
As described previously, the GnT-VmRNA and protein expression inHepG2 GnT-V/1564 were both reduced, then the proliferation, migra-tion and invasion abilities in vitrowas investigated in the further assays.
Cell growth rate of both groups and growth inhibition rate curves ofHepG2 GnT-V/1564 showed that the growth rate of HepG2 GnT-V/1564 was significantly inhibited compared with HepG2 GnT-V/NC(Pb0.001) (Figs. 2 A, B). Wound healing assay also revealed that down-regulation of GnT-V significantly reduced the wound healing rates ofthe HepG2 cell line at 12 h, 24 h and 36 h (Fig. 3). Cell invasion assayshowed that the numbers of penetrating cells were (11.00±1.86) and(47.35±1.98) for HepG2 GnT-V/1564 and HepG2 GnT-V/NC cells(t=59.770, Pb0.001), respectively (Figs. 4 A, B). These data suggested
that down-regulation of GnT-V can inhibit the proliferation, migrationand invasion of HepG2 in vitro.
Down-regulation of GnT-V suppresses tumor growth andmetastasis in vivo
The mean weights of transplanted tumors from HepG2 GnT-V/NCand HepG2 GnT-V/1564 were (1.48±0.10) g and (0.48±0.07) grespectively (t=13.724, Pb0.001, Fig. 5A). The establishment oftransplanted tumorswas confirmed by H&E (Fig. 5B). The tumor volumein the left gluteal region from HepG2 GnT-V/NC was larger than thetumor in the right gluteal region from HepG2 GnT-V/1564 (Fig. 5C),which showed that down-regulation of GnT-V suppressed the HepG2cell growth in vivo.
The numbers of metastatic liver tuberculums for HepG2 GnT-V/NCand HepG2 GnT-V/1564 were (120.67±11.00) and (29.33±9.50)
A
BHepG2 GnT- /NC HepG2 GnT- /1564
CHepG2 GnT- /NC HepG2 GnT- /1564
Fig. 6. 2.5×106 cells were injected into the medial splenic tip with a 30-gauge needle. Three weeks after the injection of tumor cells, the mice were killed. The spleen and livertissues were dissected. (A) The white tuberculums on the liver tissues were metastatic tumors and were marked by black arrows. The number of metastatic liver tuberculumsfor HepG2 GnT-V/NC and HepG2 GnT-V/1564 was (120.67±11.00) and (29.33±9.50). (B) The establishment of liver metastasis tumor was confirmed by H&E (200×). As arrowsshowed that cores or nest of cells and hepatic sinusoid was observed in most fields of photomicrographs, and the nucleus was big, round or oval, with pathological karyokinesis. (C)Immunohistochemical staining showed that the GnT-V expression in HepG2 GnT-V/NC (++) was stronger than that in HepG2 GnT-V/1564 (+) in metastatic liver tuberculums(400×). Brown–yellow granules in cytoplasm were considered positive staining for GnT-V.
14 T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
(Fig. 6A, t=10.803, Pb0.001) respectively. GnT-V expression wasclosely associated with metastasis in vivo. The establishment of livermetastasis tumorwas confirmed byH&E (Fig. 6B). Immunohistochemicalstaining showed that the GnT-V expression in HepG2 GnT-V/NC (++)was stronger than that in HepG2 GnT-V/1564 (+) in metastatic livertuberculums (Fig. 6C).
Correlations of GnT-V expression with clinicopathological factors inhepatocellular carcinoma tissue and liver cirrhosis cases
The association between GnT-V protein expression and histologicalgradewas detected using tissuemicroarrays. GnT-V positive expressionwas shown in 138 cases of 140 (98.6%) HCC samples, in 3 cases of 31
Fig. 7. The GnT-V protein expression in 140 HCC cases was detected using tissue microarrays technology. The representative photomicrographs showed that brown–yellow granules incytoplasm were considered positive staining for GnT-V. (A) high expression (100×); (B) high expression (400×); (C) moderate expression (100×); (D) moderate expression(400×); (E) low expression (100×); (F) low expression (400×); (G) negative (tissues treated with the GnT-V antibody did not show positive staining) (100×); (H) negative(tissues treated with the GnT-V antibody did not show positive staining) (400×).
Table 2Correlation of GnT-V expression with histological grade.
15T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
(9.7%) liver cirrhosis cases, and in 1 case of 20 (5.0%) normal liver tis-sues. The “high GnT-V” expression was found in 55.9% (19/34) of poordifferentiated cases, 25.7% (18/70) in moderate cases, and 0.0% (0/36)in well-differentiate cases. The “moderate GnT-V” expression rates
Table 3Correlation of GnT-V expression with TNM grade.
16 T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
were 19.4% (7/36), 51.4% (36/70), and 32.3% (11/34) in well, moderate,and poor cases, respectively. The “low expression” for GnT-V was 75.0%(27/36) in well differentiated cases, 22.9% (16/70) in moderate cases,11.8% (4/34) in poor differentiated cases, 9.7% (3/31) in liver cirrhosisand 5.0% (1/20) in normal liver tissues (Fig. 7, Table 2). A significantlyincreased expression of GnT-V in poorly and moderately differentiatedcases was observed compared with that in well-differentiated cases.
The correlation between GnT-V expression and TNM stage wasalso investigated. The TMAs consisted of 140 HCC cases with TNMstage (26 cases of T1, 46 cases of T2, 48 cases of T3, 20 cases of T4).The “high expression” for GnT-V was observed in 0.0% (0/26) of T1stage, 8.7% (4/46) of T2 stage, 29.2% (14/48) of T3 stage and 50%(10/20) of T4 stage cases. The “moderate GnT-V” expression rateswere 0.0% (0/26) in T1, 39.1% (18/46) in T2, 43.8% (21/48) in T3and 40% (8/20) in T4, respectively. In addition, the “low expression”for GnT-V was 92.3% (24/26) in T1, 52.2% (24/46) in T2, 27.1% (13/48)in T3 and 10% (2/20) in T4 (Fig. 7, Table 3). The GnT-V expression wasrising gradually from stages T1 to T4 . These data indicated that theGnT-V expression was positively correlated with TNM stage.
Discussion
The relationship between GnT-V expression and malignancy hasbeen studied in many tumors, such as in human fibrosarcoma cellHT1080 and Neuroblastoma (Fernandes, et al., 1991; Guo, et al., 2002;Inamori, et al., 2006; Przybylo, et al., 2008). In human fibrosarcomacell HT1080, increased expression of GnT-V could result in less adhesiveand more motile ability (Guo, et al., 2002). In Neuroblastoma, theexpression of GnT-V was related with the apoptosis of tumor (Inamori,et al., 2006). To investigate the role of GnT-V in HCC, the HepG2 GnT-V/1564 cell expressing lower GnT-V was developed from HepG2. Theproliferation, migration and invasion abilities were suppressed inHepG2 GnT-V/1564, indicating that down-regulation of GnT-V sup-pressed malignancy in vitro. This observation was consistent with previ-ous results reported by Guo et al. that over expression of GnT-V inhuman hepatocarcinoma 7721 cell resulted in the increased metastasis-associated phenotypes (Guo, et al., 2001). In our study, the relationshipbetweenGnT-V andHCCwas further analyzed in vivo. The results showedthat the mean weight of tumors from HepG2 GnT-V/NC was more thanthat fromHepG2GnT-V/1564 (t=13.724, Pb0.001), andmore livermet-astatic tuberculums were found in HepG2 GnT-V/NC compared withHepG2 GnT-V/1564. The tumor growth and metastasis ability in vivowas also decreased together with suppression of GnT-V. In summary,our study revealed sufficiently that GnT-V expressionwas positively cor-related with malignancy in HCC both in vitro and in vivo.
To further investigate the association between GnT-V and clinicalpathological features, the GnT-V expression was studied in humanhepatocellular carcinoma tissues. Yao et al. demonstrated that the activityof GnT-V in stage T1was significantly lower than that in stages T2–4 in 33cases of HCC by immunohistochemical study (Yao, et al., 1998). On theother hand, another study showed that high GnT-V expression wasmore frequently found in caseswith small size,well ormoderate differen-tiation (Ito, et al., 2001). To clarifywhether GnT-V positively or negativelycorrelated with malignancy, 140 cases of HCC and 31 liver cirrhosis caseswere studied by TMAs technology from the clinicopathological point ofview. A higher level of GnT-V was observed more frequently in theadvanced tumors with higher T stage, consistent with the resultsreported by Yao et al. To sufficiently access the role of GnT-V on HCC,the association betweenGnT-V and the histological grade of HCC tissueswas further investigated. The results showed that therewas a graded in-crease in GnT-V expression from liver cirrhosis to well differentiatedand then to poorly differentiated HCC. All the results indicated thatGnT-V expression was positively correlated with histological gradeand TNM classifications in hepatocellular carcinoma tissues.
Glycosylation is one of the most common posttranslational proteinmodifications, and nearly half of all known proteins in eukaryotes
are glycosylated (Saxon and Bertozzi, 2001). Changes in glycan areassociated with many physiological and pathological events, includingcell adhesion,migration and invasion (Dennis, et al., 1987). In the gastriccancer cells, increased GnT-V caused matriptase aberrant glycosylation,associated with malignant transformation (Ihara, et al., 2002). In thehuman colon cancer cell WiDr, aberrant glycosylation of tissue inhibitorof metalloproteinase-1 induced by up-regulated GnT-V resulted inmalignant cell behavior (Kim, et al., 2008). However, the mechanism asto how GnT-V regulates HCC malignant transformation is still unknown.Whether GnT-V influences HCC malignancy by glycosylation of cell-surface protein needs to be further discussed.
In this study, the proliferation, migration, invasion and metastasisof HepG2 cells both in vitro and in vivo were found to be inhibitedwith down-regulation of GnT-V. In addition, GnT-V expression wasmore frequently observed in HCC compared with that in liver cirrhosisand normal liver tissues, implying that GnT-V possibly contributed tomalignant transformation from noncancerous tissues to HCC. Further-more, the GnT-V expression was gradually increasing from early HCCto advanced HCC, indicating that it may be associated with the pro-gression of HCC. These data from HepG2 cell line and clinical tissueall implied thatGnT-V expressionwas positively relatedwithmalignancyin HCC. Down-regulation of GnT-Vmay suppress themalignancy of HCC,so GnT-V may be both a differentiation marker and a potential target forthe treatment of HCC.
Conflict of interest statement
The authors declare that there are no conflicts of interest.
Acknowledgments
This paper was supported by Science and Technology PlanningProject of Guangdong Province, China (2007B030703001).
References
Apweiler, R., et al., 1999. On the frequency of protein glycosylation, as deduced fromanalysis of the SWISS-PROT database. Biochimica et Biophysica Acta 1473, 4–8.
Asada, M., et al., 1997. Increased expression of highly branched N-glycans at cell sur-face is correlated with the malignant phenotypes of mouse tumor cells. Cancer Re-search 57, 1073–1080.
Chakraborty, A.K., Pawelek, J.M., 2003. GnT-V, macrophage and cancer metastasis: acommon link. Clinical & Experimental Metastasis 20, 365–373.
Dennis, J.W., et al., 1987. Beta 1–6 branching of Asn-linked oligosaccharides is directlyassociated with metastasis. Science 236, 582–585.
Dosaka-Akita, H., et al., 2004. Expression of N-acetylglucosaminyltransferase v is asso-ciated with prognosis and histology in non-small cell lung cancers. Clinical CancerResearch 10, 1773–1779.
Fernandes, B., et al., 1991. Beta 1–6 branched oligosaccharides as a marker of tumorprogression in human breast and colon neoplasia. Cancer Research 51, 718–723.
Granovsky, M., et al., 2000. Suppression of tumor growth and metastasis in Mgat5-de-ficient mice. Nature Medicine 6, 306–312.
Guo, H.B., et al., 2001. Relationship between metastasis-associated phenotypes and N-glycan structure of surface glycoproteins in human hepatocarcinoma cells. Journalof Cancer Research and Clinical Oncology 127, 231–236.
Guo, H.B., et al., 2002. Aberrant N-glycosylation of beta1 integrin causes reducedalpha5beta1 integrin clustering and stimulates cell migration. Cancer Research62, 6837–6845.
Guo, H.B., et al., 2007. Inhibition of a specific N-glycosylation activity results in attenu-ation of breast carcinoma cell invasiveness-related phenotypes: inhibition of epi-dermal growth factor-induced dephosphorylation of focal adhesion kinase.Journal of Biological Chemistry 282, 22150–22162.
Hakomori, S., 1989. Aberrant glycosylation in tumors and tumor-associated carbohy-drate antigens. Advances in Cancer Research 52, 257–331.
Handerson, T., Pawelek, J.M., 2003. Beta1,6-branched oligosaccharides and coarse ves-icles: a common, pervasive phenotype in melanoma and other human cancers.Cancer Research 63, 5363–5369.
Handerson, T., et al., 2005. Beta1,6-branched oligosaccharides are increased in lymphnodemetastases and predict poor outcome in breast carcinoma. Clinical Cancer Re-search 11, 2969–2973.
Ihara, S., et al., 2002. Prometastatic effect of N-acetylglucosaminyltransferase V is dueto modification and stabilization of active matriptase by adding beta 1–6 GlcNAcbranching. Journal of Biological Chemistry 277, 16960–16967.
17T. Wei et al. / Experimental and Molecular Pathology 93 (2012) 8–17
Inamori, K., et al., 2006. High expression of N-acetylglucosaminyltransferase V in favor-able neuroblastomas: Involvement of its effect on apoptosis. FEBS Letters 580,627–632.
Ishimura, H., et al., 2006. N-acetylglucosaminyltransferase V and beta1-6 branching N-linked oligosaccharides are associated with good prognosis of patients with blad-der cancer. Clinical Cancer Research 12, 2506–2511.
Ito, Y., et al., 2001. Elevated expression of UDP-N-acetylglucosamine: alphamannosidebeta1,6 N-acetylglucosaminyltransferase is an early event in hepatocarcinogenesis.International Journal of Cancer 91, 631–637.
Kim, Y.S., et al., 2008. Functional proteomics study reveals that N-Acetylglucosaminyl-transferase V reinforces the invasive/metastatic potential of colon cancer throughaberrant glycosylation on tissue inhibitor of metalloproteinase-1. Molecular & Cel-lular Proteomics 7, 1–14.
Ohtsubo, K., Marth, J.D., 2006. Glycosylation in cellular mechanisms of health and dis-ease. Cell 126, 855–867.
Pierce, M., et al., 1997. Regulation of N-acetylglucosaminyltransferase V and Asn-linkedoligosaccharide beta(1,6) branching by a growth factor signaling pathway and
effects on cell adhesion and metastatic potential. Glycoconjugate Journal 14,623–630.
Przybylo, M., et al., 2008. Beta1-6 branching of cell surface glycoproteins may contrib-ute to uveal melanoma progression by up-regulating cell motility. Molecular Vi-sion 14, 625–636.
Saxon, E., Bertozzi, C.R., 2001. Chemical and biological strategies for engineering cellsurface glycosylation. Annual Review of Cell and Developmental Biology 17, 1–23.
Taniguchi, N., et al., 1999. Implication of N-acetylglucosaminyltransferases III and V incancer: gene regulation and signaling mechanism. Biochimica et Biophysica Acta1455, 287–300.
Yao, M., et al., 1998. Elevated activity of N-acetylglucosaminyltransferase V in humanhepatocellular carcinoma. Journal of Cancer Research and Clinical Oncology 124,27–30.
