Purpose and Experimental Design: Little is known about the molecular events leading to the development and progression of pathological tumor stage 2 (pT2) gallbladder carcinoma. An alteration in the site of O-glycosylation may be associated with malignant behavior of carcinoma cells by modulation of the biological properties of the target mucin. The UDP-N-acetyl-α-d-galactosamine-polypeptide N-acetylgalactosaminyltransferase isozyme 3 (GalNAc-T3) has the epithelial gland-specific expression and catalyzes mucin-type O-glycosylation. In this study, immunohistochemistry was performed to determine the expression level of GalNAc-T3 in 34 cases of pT2 gallbladder carcinoma to determine the correlation of the GalNAc-T3 expression level with mode of recurrence and postsurgical survival.
Results: The expression levels of GalNAc-T3 protein and mRNA were increased in gallbladder carcinomas compared with the levels in adjacent noncancerous tissues and in intact gallbladders. Immunostaining of GalNAc-T3 was recognized in the cancerous epithelia, and the subcellular localization was classified into granular and diffuse types. In the 34 cases of pT2 carcinoma, the localization of GalNAc-T3 was granular type in 50% and diffuse type in 50% of the cases at the deepest invading sites in the subserosal layer. Postsurgical recurrence was significantly more frequent in cases showing diffuse-type localization of GalNAc-T3 at the deepest invading sites (65%) than in those showing granular-type localization (23%; P < 0.05). Postsurgical survival was significantly poorer in cases showing diffuse-type localization than in those showing granular-type localization (P = 0.033)
Conclusions: In pT2 gallbladder carcinoma, the presence of diffuse-type localization of GalNAc-T3 in the subserosal layer is correlated with aggressiveness of the disease. This phenotype may serve as a unique biological feature associated with the malignant behavior.
Gallbladder carcinoma has always been associated with a dismal overall prognosis (1, 2, 3, 4, 5) . This is because the disease is usually detected at an advanced stage. The clinical course of gallbladder carcinoma has been thought to depend on the depth of tumor invasion (6, 7, 8) . Despite a theoretical advantage for gallbladder carcinoma (a tumor invading the perimuscular connective tissues but not extending beyond the serosa or into the liver), the prognosis of pathological tumor stage 2 (pT2) carcinoma is not necessarily favorable. The rate of 5-year postsurgical survival in cases of pT2 carcinoma, about 50–80%, is intermediate between that of pT1 carcinoma and that of pT3 and pT4 carcinomas (9, 10, 11) . It has been found that the parameters of histopathological malignancies such as lymphatic permeation and venous permeation in the subserosal layer are not correlated with either mode of recurrence or postsurgical prognosis of pT2 carcinoma (12) . These findings are attributed to the variety of ways in which pT2 carcinoma progresses and to the fact that prognostic factors affecting the progression of less-advanced lesions such as pT2 carcinoma have not been fully elucidated.
Mucin glycoproteins (mucins) expressed on carcinoma cells representing a metastatic phenotype are qualitatively and quantitatively different from those of the nonmetastatic phenotype (13, 14, 15, 16) . Alterations in the glycosylation of carcinoma-associated mucins (e.g., MUC1) may modulate the biological behavior of carcinoma cells and allow the cells to disseminate, invade, and survive at distant organ sites (17) . These alterations include aberrant glycosylation, resulting in the formation of disaccharides such as sialosyl-Tn antigen (18, 19, 20, 21) , and incomplete glycosylation, resulting in an accumulation of mucin core-region oligosaccharide structures such as Tn (19 , 22, 23, 24, 25) . These carbohydrate antigens on the cell surface are useful for monitoring tumor status in subjects with malignant diseases (19 , 25, 26, 27) .
In biosynthesis of carbohydrate structures, the initial glycosylation of mucin-type O-linked proteins is catalyzed by one of the UDP-N-acetyl-α-d-galactosamine-polypeptide N-acetylgalactosaminyltransferases (GalNAc-T family enzymes; Refs. 28, 29, 30 ). Of the GalNAc-T isozymes, the expression of GalNAc-T3 is highly tissue-specific for organs containing secretory epithelial glands (29, 30, 31, 32) , and this enzyme is expressed at high levels in human adenocarcinoma cell lines (33) . Moreover, MUC1 is glycosylated on the threonine of GVTS motif in tandem repeat (34) at which only GalNAc-T1 and T3 can give GalNAc (28) .
On the other hand, GalNAc-T family enzymes are localized in the Golgi apparatus (35) and are used as markers for Golgi apparatus elements. Notably, their subcellular localization has been shown to change under conditions of malignant transformation of cells (36) because of alterations in the organization of cytoskeletal elements functioning in maintenance of the organization of the Golgi apparatus (37) . Therefore, the expression level of GalNAc-T3 may not only reflect a reorganization of Golgi apparatus elements in association with either malignant transformation of epithelial cells or the differentiation status of adenocarcinoma cells but also may cause modification of carbohydrate structures of carcinoma-associated mucins, which, in turn, influences the biological behavior of carcinoma cells through phenotypic changes.
In this retrospective analysis, the immunohistochemical expression of GalNAc-T3 was studied in formalin-fixed, paraffin-embedded surgical specimens from patients with gallbladder carcinomas of different depths of invasion (pT1-pT4). It is thought that carcinoma cells in the deepest sites of invasion have a greater capability to invade and metastasize than do carcinoma cells in other regions (38 , 39) . Therefore, in 34 cases of pT2 carcinoma, correlations of the immunohistochemical localization types of GalNAc-T3 at the deepest invading sites in the subserosal layer of pT2 carcinoma, as a predictor of invasive/metastatic potential, with the clinicopathological findings, mode of recurrence, and postsurgical survival were investigated.
MATERIALS AND METHODS
Specimens from 79 patients (34 males and 45 females) with gallbladder carcinoma (5 with pT1, 34 with pT2, 9 with pT3, and 31 with pT4 carcinomas) were included in this study. All of the pT1 and pT2 carcinomas were curatively resected with a free surgical margin. The mean age of the patients was 66 years (range, 40–89 years). The patients were diagnosed as having gallbladder carcinoma and underwent operations between October 1976 and February 2002 in the Hospital of the University of Tsukuba School of Medicine and in the Surgical Department of Nagasaki Central National Hospital. Gallbladder carcinoma was diagnosed on the basis of histological findings and classified according to the Tumor-Node-Metastasis classification of the American Joint Committee on Cancer (40) , pT1, a tumor confined to mucosa or muscle coat; pT2, a tumor that has invaded the perimuscular connective tissues with no extension beyond the serosa or into the liver; pT3, a tumor that has perforated the serosa (visceral peritoneum) or has directly invaded one adjacent organ, or both (extension of ≤2 cm into the liver); and pT4, a tumor that has extended >2 cm into the liver and/or into two or more adjacent organs. The surgical procedures are shown in Table 1⇓ . For pT2 carcinoma, simple cholecystectomy was performed in 16 of the 34 patients, cholecystectomy combined with bile duct resection was performed in 8 patients, cholecystectomy with combined bile duct resection and hepatic resection was performed in 5 patients, cholecystectomy combined with bile duct resection and pancreatoduodenectomy was performed in 1 patient, cholecystectomy combined with pancreatoduodenectomy together with bile duct resection and hepatic resection was performed in 2 patients, and cholecystectomy combined with hepatic resection was performed in 2 patients. Histological examination revealed that all cases of pT2 carcinoma had neither hepatic infiltration of carcinoma nor invasion into the hepatoduodenal ligament.
Follow-up periods until February 2002 ranged from 2.2 to 178.3 months. Of the 34 patients, 20 were alive as of February 2002, 11 had died from distant metastasis (Table 2)⇓ , and 3 had died from some other disease (cerebral infarction in 2 and pancreatic carcinoma in 1). The latter 3 patients were treated as lost cases. Survival curves were assessed by the Kaplan-Meier method. Statistical analysis was performed using Stat View and its survival tools (Abacus Concepts, Berkeley, CA) for Macintosh.
For pT3 and pT4 carcinomas, 21 cases who had synchronous metastasis in the liver and/or peritoneum at the time of surgery (Any N M1) and 19 patients who had neither synchronous distant metastasis nor lymph node metastasis (N0 M0) were selected to compare the expression level of GalNAc-T3 in the two groups.
In addition, gallbladder specimens obtained at surgery from 6 subjects with gallbladder stones who had undergone cholecystectomy were used as intact gallbladders.
Antibody Raised Against GalNAc-T3.
A polyclonal antibody (pAb) raised against human GalNAc-T3 was prepared in the laboratory of Taiho Pharmaceutical Company, Ltd. (Saitama, Japan) by multiple immunizations of a New Zealand White rabbit using synthetic peptides as described previously (33) . A monoclonal antibody raised against human GalNAc-T3 (41) , which was kindly supplied from Dr. Henrik Clausen (Copenhagen University, Copenhagen, Denmark), was also available. By comparing the reactivity, both the pAb and monoclonal antibody could be used in the immunoblot analysis. However, in the immunohistochemistry, the pAb gave clear immunostainings, whereas the monoclonal antibody gave only trace or weak stainings.
Cell Lines and Culture Conditions.
Mz-ChA-1 and Mz-ChA-2, gallbladder carcinoma cell lines (42) , and Sk-ChA-1, a bile duct carcinoma cell line (42) , were obtained from Dr. Alexander Knuth (Johaness-Gutenberg University, Mainz, Germany). TGBC-1-TKB, TGBC-2-TKB, and TGBC-44-TKB, gallbladder carcinoma cell lines (43) , were obtained from Dr. Takeshi Todoroki (University of Tsukuba, Ibaraki, Japan). KMBC, a bile duct carcinoma cell line, and KMC-1 and KMCH-1, cholangiocellular carcinoma cell lines (44 , 45) , were obtained from Dr. Masamich Kojiro (Kurume University School of Medicine, Kurume, Japan). The cells were maintained in DMEM containing 10% heat-inactivated FCS (Hyclone Laboratories, Inc., Logan, UT) in a humidified atmosphere with 5% carbon dioxide at 37°C.
Immunoblot Analysis of GalNAc-T3 in Gallbladder Carcinoma and Cultured Biliary Carcinoma Cells.
Immunoblot analysis of GalNAc-T3 was performed using the lysates of either frozen tissue specimens or cultured cells as described previously (33) . Briefly, the membranes were incubated overnight in 2% BSA in PBS at 4°C and then mixed with the pAb of GalNAc-T3 diluted in 2% BSA (dilution, 1:5000) for 2 h. Each membrane was probed again with a monoclonal antibody of β-actin (Sigma-Aldrich Co., St. Louis, MO) diluted in 2% BSA (dilution, 1:1000). After washing with PBS, the membranes were incubated with goat antirabbit or antimouse IgG labeled with horseradish peroxidase (Zymed Laboratories Inc., San Francisco, CA) for 40 min. The membranes were then washed with PBS and treated with ECL (Amersham, Buckinghamshire, United Kingdom) to visualize the antibodies that had bound.
Immunostaining of GalNAc-T3 in Gallbladder Carcinoma.
Immunostaining of GalNAc-T3 was performed using the pAb of GalNAc-T3. Gallbladder carcinoma tissues that had been preserved in 10% formalin and then embedded in paraffin were serially sectioned at 2 μm in thickness, mounted on silane-coated slides, and deparaffinized. The slides were immersed for 20 min in 0.3% hydrogen peroxide in methanol to deplete endogenous peroxidase. After washing with PBS, the slides were incubated with a protein blocking agent for 5 min at room temperature in a humidity chamber. The slides were then stained by the indirect immunoperoxidase method using the pAb (dilution, 1:3000). A negative control was made using BSA instead of the pAb.
Evaluation of sections was performed by a single pathologist who was blinded to the clinical characteristics and pathological grade of response. The total number of cancerous epithelia in each section was evaluated. The immunohistochemical localization of GalNAc-T3 was classified into granular and diffuse types based on the predominant subcellular distribution: granular type was GalNAc-T3 showing granular staining and being restricted predominantly in the supranuclear area of the cancerous epithelia; and diffuse type was GalNAc-T3 showing no granular staining and being found in the cytoplasm of the cancerous epithelia. Fig. 1⇓ shows a representative immunohistochemical localization of GalNAc-T3 in cancerous epithelia in the subserosal layer of pT2 carcinoma. The localization of GalNAc-T3 was judged to be granular or diffuse type when >50% of the total number of cancerous epithelia in each section showed granular- or diffuse-type subcellular distribution, and the localization was examined in both the mucosal or proper muscle layers (surface site) and the subserosal layer (invading site).
Steady-State mRNA Level of GalNAc-T3.
Total RNA was isolated from the gallbladder carcinoma specimens (from 3 with pT1, 6 with pT2, 2 with pT3, and 13 with pT4 carcinomas) and the gallbladder specimens (from 6 gallstone subjects) using TRIzol reagent. Competitive reverse transcription-PCR was performed using a DNA Thermal Cycler (Gene Amp PCR System 2400; Perkin-Elmer, Inc., Wellesley, MA). PCR was subjected to each cycle (GalNAc-T3, 45; GAPDH, 35) at 94°C for 30 s, at 58°C for 30 s, and at 72°C for 1 min. PCR primers were designed from cDNA sequences for human GalNAc-T3 (33) and then synthesized (Proligo LLC, Inc., Boulder, CO) as follows: GAPDH, sense 5′-GACCCCTTCATTGACCTCAA CTAC-3′, antisense 5′-CAGTGATGGCATGGACTGTGGT-3′; competitor, 5′-CAGTGATGGCATGGACTGTGGTTCACACCCATGACGAAC-3′; GalNAc-T3, sense 5′-TGGTTGGCTAGAACCTCTGTTG G-3′, antisense, 5′-GGT CTGGCACATACACCTCTGG-3′, competitor, 5′-TGGTTGGCTAGAACCTCTGTTGGCTGGGAGTCG CTTCCT-3′. In each experiment, PCR was done in triplicate. For quantitative assessment, the amounts of fluorescence intensity were measured using FluorImager (Molecular Dynamics, Sunnyvale, CA). The data were expressed relative to the amount of GAPDH mRNA present in each specimen and then averaged.
Values are given as means ± SE. A two-sided χ2 test was used for comparison of clinicopathological data between groups. The survival of patients was recorded every month, and patient survival was analyzed by the method of Kaplan-Meier. Differences in the survival of patients in subgroups were analyzed by the log-rank test. A P < 0.05 was defined as statistically significant.
Immunoblot Analysis of GalNAc-T3 in Gallbladder Carcinoma and Cultured Biliary Carcinoma Cells.
GalNAc-T3 protein was included in the lysates of gallbladder carcinoma tissues and cultured biliary carcinoma cells. GalNAc-T3 expression was confirmed in all of the specimens of gallbladder carcinoma and intact gallbladders (Fig. 2A)⇓ . Protein levels of GalNAc-T3 were increased in the tissue specimens of well- or moderately differentiated pT1-pT4 gallbladder carcinomas compared with the levels in the specimens of adjacent noncancerous gallbladders and intact gallbladders (Fig. 2A)⇓ . The protein levels appeared to increase in parallel with the depth of invasion. However, the protein level was found to be very low in a case of poorly differentiated pT4 carcinoma.
As shown in Fig. 2B⇓ , the protein levels were higher in the lysates of well- or moderately differentiated carcinoma cell lines (Mz-ChA-1, KMBC, and KMC-1) than in those of poorly differentiated carcinoma cell lines (Mz-ChA-2, TGBC-1-TKB, TGBC-2-TKB, SK-ChA-1, and KMCH-1). However, the protein level was very high in the lysate of TGBC-44-TKB, a poorly differentiated adenocarcinoma cell line. These results imply that the GalNAc-T3 protein level is increased in gallbladder carcinomas and is related to the differentiation status of cancerous epithelia and to that of cultured adenocarcinoma cells.
Steady-State mRNA Level of GalNAc-T3 in Gallbladder Carcinoma and Cultured Biliary Carcinoma Cells.
To determine whether the high level expression of GalNAc-T3 protein in gallbladder carcinoma tissues and in well- or moderately differentiated adenocarcinoma cells was caused by an increased level of the gene transcription, the steady-state mRNA levels of GalNAc-T3 in carcinomas of different depths of invasion (3 with pT1, 6 with pT2, 2 with pT3, and 13 with pT4 carcinomas), their adjacent noncancerous tissues and intact gallbladder tissues, and cultured biliary carcinoma cells were measured by competitive reverse transcription-PCR (Fig. 3)⇓ .
The mRNA level of GalNAc-T3 was significantly higher in the tissue specimens of gallbladder carcinomas (0.11 ± 0.02% of G3PDH mRNA, mean ± SE; P < 0.05) than in the adjacent noncancerous tissues (0.05 ± 0.01%) and in the specimens of intact gallbladders (0.03 ± 0.01%; Fig. 3-A⇓ ). However, the mRNA levels were not significantly different in the specimens of pT1-pT2 carcinomas (0.11 ± 0.03%) and the specimens of pT3-pT4 carcinomas (0.11 ± 0.03%; Fig. 3B⇓ ). With respect to the histological grade, the mRNA level tended to be higher in the specimens of papillary adenocarcinoma (0.13 ± 0.04%) than in the specimens of tubular adenocarcinoma (0.104 ± 0.031%) and in the specimens of poorly differentiated adenocarcinoma (0.09 ± 0.04%; Fig. 3C⇓ ). However, the differences were not statistically significant.
In the cultured biliary carcinoma cells, the mRNA levels of GalNAc-T3 were not significantly different in the well- or moderately differentiated adenocarcinoma cell lines (data not shown). The mRNA levels were not correlated with the protein levels. The results indicate that the high expression levels of GalNAc-T3 in the gallbladder carcinomas and the well- or moderately differentiated adenocarcinoma cell lines may not be simply caused by increased transcription levels.
Immunohistochemical Localization of GalNAc-T3 in Gallbladder Carcinoma.
The immunohistochemical localization of GalNAc-T3 in pT1-pT4 gallbladder carcinomas and that in noncancerous pathological lesions of the gallbladders with gallbladder stones were studied (Table 3)⇓ . Reflecting the increased protein levels detected by the immunoblot analysis (Fig. 2)⇓ , the immunostaining of GalNAc-T3 was found to be more intense in cancerous epithelia than in noncancerous pathological lesions and in normal epithelia of the gallbladders (Fig. 4)⇓ . The localization was heterogeneous in gallbladder carcinomas with granular or diffuse type of subcellular distribution, in contrast with the localization in intact gallbladders, which was solely granular type (Fig. 4)⇓ . The in situ noncancerous pathological lesions in the gallbladder, i.e., hyperplasia, pseudopyloric gland metaplasia, and dysplasia, were solely granular type (Fig. 4)⇓ . However, no diffuse type of subcellular distribution was observed in either normal epithelia or noncancerous pathological lesions of the gallbladders.
The localization of GalNAc-T3 appeared to be heterogeneous in carcinomas of different depths of invasion (Table 3)⇓ , although it was found that the immunohistochemical localization varied from pT1 carcinoma in the early stage to pT3 and pT4 carcinomas in the advanced stage. In pT1 carcinoma, the localization was granular type in all of the cases. In pT2 carcinoma, the localization was granular type in 82% and diffuse type in 18% of the cases at the noninvading surface sites. At the deepest invading sites in the subserosal layer (Fig. 5)⇓ , the localization was granular type in 50% and diffuse type in 50% of the cases. In pT3 and pT4 carcinomas, the localization was granular type in 53% and diffuse type in 47% of the cases at the noninvading surface sites, and at the deepest invading sites, the localization was granular type in 23% and diffuse type in 77% of the cases. Notably, the proportion of diffuse-type localization of GalNAc-T3 at the deepest invading sites was increased significantly in pT2 carcinoma and in pT3 and pT4 carcinomas compared with the proportion in noncancerous pathological lesions, pT1 carcinoma, and their corresponding surface sites of pT2-pT4 carcinomas. In contrast, the proportion of granular-type localization of GalNAc-T3 was significantly decreased at the deepest invading sites. It should be noted that the proportion of cancerous epithelia showing diffuse-type localization was increased at the deepest invading sites in parallel with the depth of invasion.
Relationship between Pathological Malignancies and Localization Type of GalNAc-T3 in Patients with pT2 Carcinoma.
The 34 patients with pT2 carcinoma were divided into two groups based on the localization type of GalNAc-T3 at the deepest invading sites in the subserosal layer. A comparison of granular- and diffuse-type groups was made with special reference to pathological malignancies, i.e., histological grade, lymphatic permeation, venous permeation, and lymph node metastasis. The results revealed no significant differences in the parameters of pathological malignancies between the two groups (Table 4)⇓ . A comparison was also made for pT3 and pT4 carcinoma patients with synchronous metastasis in the liver and/or peritoneum and those without metastasis. The results revealed no significant differences in the parameters of pathological malignancies between the two groups (Table 4)⇓ .
Relationship between Mode of Recurrence in Patients with pT2 Carcinoma and Localization Type of GalNAc-T3 in the Specimens.
The postsurgical recurrent mode in patients with pT2 carcinoma in the granular- and diffuse-type groups were compared. Of the 17 patients in the diffuse-type group, 5 had peritoneal dissemination, 6 had metastasis in distant organs, and 3 had lymph node metastasis (Table 2)⇓ . In contrast, of the 17 patients in the granular-type group, 4 had peritoneal dissemination, 1 had metastasis in distant organs, and none had lymph node metastasis. Interestingly, it should be noted that in pT2 carcinoma, postsurgical recurrence in distant organs was found to be more frequent in patients in the diffuse-type group (65%) than in those in the granular-type group (23%). The difference was statistically significant (P < 0.05). Therefore, the diffuse-type localization of GalNAc-T3 at the deepest invading sites may be an important biological predictor of postsurgical recurrence of pT2 carcinoma.
Relationship between Synchronous Metastasis in the Liver and/or Peritoneum of Patients with pT3 or pT4 Carcinoma and Localization Type of GalNAc-T3 in the Specimens.
The localization types of GalNAc-T3 at the deepest invading sites were compared in pT3 and pT4 carcinoma patients with synchronous metastasis in the liver and/or peritoneum at the time of surgery and those without metastasis. Of the 21 patients who had synchronous distant metastasis in the liver and/or peritoneum (Any N M1), the localization of GalNAc-T3 at the deepest invading sites was granular type in 1 (5%) and diffuse type in 20 (95%; Table 5⇓ ). On the other hand, of the 19 patients who had neither synchronous distant metastasis nor lymph node metastasis (N0 M0), the localization was granular type in 8 (42%) and diffuse type in 11 (58%). The proportion of diffuse-type localization at the deepest invading sites was significantly higher in patients with synchronous metastasis than in those without metastasis (P < 0.05).
Relationship between Postsurgical Survival of Patients with pT2 Carcinoma and Localization Type of GalNAc-T3 in the Specimens.
The overall postsurgical survival rate in the 34 patients with pT2 gallbladder carcinoma was poor in contrast with the quite favorable prognosis of patients with pT1 carcinoma (Table 4)⇓ . Similarly, a comparison of the granular- and diffuse-type groups with special reference to postsurgical survival showed that the survival rate of patients in the diffuse-type group was significantly lower than that of patients in the granular-type group (P = 0.031; Fig. 6⇓ ).
The protein and mRNA levels of GalNAc-T3 were increased in the tissue specimens of well- or moderately differentiated gallbladder carcinomas in parallel with the depth of invasion compared with the levels in the specimens of adjacent noncancerous gallbladders and intact gallbladders (Fig. 2A⇓ and Fig. 3⇓ ). In the immunohistochemistry, the localization of GalNAc-T3 was granular type in all of the noncancerous epithelia of the gallbladders (Table 3)⇓ , which suggests that the granular-type localization is the normal subcellular distribution of GalNAc-T3. In contrast, the proportion of diffuse-type localization was increased in the cancerous epithelia in advanced cases or in those in the deepest invading sites (Table 3)⇓ . Among the 34 patients who had undergone surgery with curative intent for less-advanced pT2 gallbladder carcinoma, those showing diffuse-type localization of GalNAc-T3 at the deepest invading sites in the subserosal layer had a poorer postsurgical survival rate (Fig. 5)⇓ , due to peritoneal dissemination or metastasis in distant organs, than those showing granular-type localization (Table 2)⇓ . The results indicate that gallbladder carcinoma cells in the deepest invading sites showing diffuse-type localization of GalNAc-T3 have a strong potential for metastasis and that micrometastasis may already have occurred at the time of surgery if the localization is diffuse type at the deepest invading sites in the subserosal layer. This hypothesis may be supported by the observation that diffuse-type localization of GalNAc-T3 predominates at the deepest invading sites of pT3 and pT4 carcinomas with synchronous metastasis in the liver and/or peritoneum (Table 5)⇓ . Thus, alterations in the subcellular localization may be closely associated with tumor biological aspects of gallbladder carcinoma, e.g., aggressiveness to form distant organ metastasis. These results are consistent with those of a recent study on the immunohistochemical expression of GalNAc-T3 in colorectal carcinomas (46) .
The prognosis of pT2 carcinoma is not necessarily favorable despite a theoretical advantage for the carcinoma not invading the perimuscular connective tissues and not extending beyond the serosa or into the liver, This may be because approximately half of the patients had malignant infiltration into the lymphatic, venous, and perineural spaces, and the frequency of lymph node metastasis was 50% (9 , 10) . In this study, the diffuse-type localization of GalNAc-T3 at the deepest invading sites correlated with neither histological grade nor parameters of clinicopathological malignancies in 34 cases of pT2 carcinoma. Due to the poor association of the expression level with clinicopathological findings of pT2 carcinoma, the localization type of GalNAc-T3 at the deepest invading sites may be considered to be an independent prognostic marker for pT2 carcinoma.
Alterations in the subcellular localization of GalNAc-T3 may be simply associated with the status of tumor differentiation and caused by loss of functional differentiation of the carcinoma cells, that is, the failure to establish or maintain polar expression of normal epithelial localization of this enzyme. The diffuse-type localization of GalNAc-T3 in gallbladder carcinoma indicates a reorganization of the Golgi apparatus elements in the carcinoma cells as reported previously (36) . Therefore, the processing and targeting pathway of GalNAc-T3 may be defective in the cells. However, in cases of pT2 gallbladder carcinoma, there was no strong correlation between histological grade (tumor differentiation), i.e., papillary, tubular, or poorly differentiated adenocarcinoma, and expression level of GalNAc-T3 at the deepest invading sites (Table 4)⇓ , indicating that diffuse-type localization cannot be explained solely by differences in histological grade of pT2 carcinoma. Instead, the immunohistochemical expression level of GalNAc-T3 may be a useful biological tool to scale functional differentiation of carcinoma cells of pT2 gallbladder carcinoma.
Malignant transformation of glandular epithelia is accompanied by alterations in the biochemical and biological characteristics of mucins. The alterations in mucin molecules include not only changes in mucin gene expression levels but also aberrant glycosylation of mucin core polypeptides (47 , 48) . Carcinoma-associated alterations in the expression levels of GalNAc-T3 and its enzyme activity may be related to aberrant glycosylation of mucins. O-glycosylation is regulated in part by differential expression of GalNAc-T in epithelial cells (28 , 31 , 32 , 49, 50, 51) . Also, sialylation of a greater portion of carbohydrates of mucins by sialyltransferases is increased in the process of tumor progression (52 , 53) . If the level of GalNAc-T3 activity in gallbladder carcinomas is high in relation to the increased protein levels (Fig. 2)⇓ , it is likely that the pronounced O-glycosylation of mucins leads to increased production of sialylated mucin(s) in the cells. The findings in our recent study (54) suggest that the antagonizing effect of sialylated MUC1 mucin(s) on E-cadherin is potentiated at the deepest invading sites of advanced gallbladder carcinoma.
In summary, the results of this study suggest that the expression of GalNAc-T3 at protein and mRNA levels is up-regulated in gallbladder carcinomas of different depths of invasion and that the immunohistochemical expression level of GalNAc-T3 in the subserosal layer of cases of pT2 gallbladder carcinoma is correlated with aggressiveness of the disease, such as the tendency to form distant recurrences, and with postsurgical survival. This phenotype may serve as a unique biological feature associated with the malignant behavior and may help to identify patients in need of closer follow-up as well as more aggressive treatment. From clinical perspectives, it would be important to determine whether GalNAc-T3 expression is an independent prognostic factor in pT2 carcinoma through the study of a large number of cases.
Grant support: Grants-in-aid for Research on Intrahepatic Calculi from the Ministry of Health and Welfare, Japan, a grant-in-aid from the Ministry of Education (No. 09670509), and a grant-in-aid from The University of Tsukuba Research Projects, Japan.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Junichi Shoda, Department of Gastroenterology, Institute of Clinical Medicine, University of Tsukuba, 1-1-1 Tennodai, Tsukuba-shi, Ibaraki 305-8575, Japan. Phone: 81-29-853-3124; Fax: 81-29-853-3124; E-mail:
- Received July 11, 2003.
- Revision received November 14, 2003.
- Accepted November 18, 2003.