Predictive Markers and Cancer Prevention
Expression of cyclooxygenase-2 in the subserosal layer correlates with postsurgical prognosis of pathological tumor stage 2 carcinoma of the gallbladder †
Article first published online: 14 JAN 2002
Copyright © 2002 Wiley-Liss, Inc.
International Journal of Cancer
Volume 98, Issue 3, pages 427–434, 20 March 2002
How to Cite
Kawamoto, T., Shoda, J., Asano, T., Ueda, T., Furukawa, M., Koike, N., Tanaka, N., Todoroki, T. and Miwa, M. (2002), Expression of cyclooxygenase-2 in the subserosal layer correlates with postsurgical prognosis of pathological tumor stage 2 carcinoma of the gallbladder . Int. J. Cancer, 98: 427–434. doi: 10.1002/ijc.10222
Parts of this work have been published in abstract form (Gastroenterology 1999;116:G0060; Gastroenterology 2000;118:A451).
- Issue published online: 28 FEB 2002
- Article first published online: 14 JAN 2002
- Manuscript Accepted: 2 NOV 2001
- Manuscript Revised: 11 OCT 2001
- Manuscript Received: 21 AUG 2001
- Ministry of Health and Welfare, Japan
- Ministry of Education. Grant Number: 09670509
- University of Tsukuba Research Projects, Japan
- distant metastasis;
- postsurgical survival;
- tumor biology
Postsurgical recurrence at distant sites frequently occurs in pathological tumor stage 2 (pT2) carcinoma of the gallbladder even though the carcinoma is limited to the gallbladder wall. Little is known, however, about the molecular events leading to its development and progression. A large body of evidence suggests that cyclooxygenase-2 (COX-2) is up-regulated in carcinoma tissues and plays roles in promoting cell-proliferation, growth and metastasis of carcinoma cells. In the present study, immunohistochemistry was performed to determine the expression levels of COX-2 in the subserosal layer of 33 cases of pT2 gallbladder carcinoma in which curative resections had been performed and to determine the correlations of the expression levels of COX-2 with mode of recurrence and postsurgical survival. Immunostaining of COX-2 in the epithelia was recognized in more than 80% of normal epithelia, noncancerous pathological lesions of the gallbladder except for intestinal metaplasia and pT1–4 carcinoma specimens. Intense staining was observed in large percentages of hyperplastic lesions (65%), pT2 carcinoma specimens (76%) and pT3 and pT4 carcinoma specimens (64%) compared to the percentages of normal epithelia and other pathological lesions (0–25%). Intense staining was also observed in the adjacent stroma in pT2 carcinoma specimens (33%) and in those in pT3 and pT4 carcinoma specimens (43%) but only in small percentages of the stroma adjacent to normal epithelia and pathological lesions (0–8%). In situ hybridization confirmed the existence of COX-2 mRNA in both the cancerous epithelia and adjacent stroma of pT2-pT4 carcinomas. In 33 cases of pT2 carcinoma, distant recurrence, i.e., liver metastasis, was seen in 3 of 9 cases of pT2 carcinoma (33%, P<0.05) with intense stromal staining in the subserosal layer and in 1 of 24 cases (4%) without intense staining, whereas no significant correlation was found between parameters of pathological malignancies (histological grade, lymphatic permeation, venous permeation and lymph node metastasis) and the intensity of stromal staining in the subserosal layer. The postsurgical survival outcome was significantly poorer in the former than in the latter (p = 0.010). In pT2 gallbladder carcinoma, upregulation of COX-2 in the stroma adjacent to the cancerous epithelia in the subserosal layer correlates with the aggressiveness of the disease, such as the tendency to form distant recurrences. This phenotype may serve as a unique biological feature associated with the malignant behavior of pT2 gallbladder carcinoma. © 2002 Wiley-Liss, Inc.
Gallbladder carcinoma has always been associated with a dismal overall prognosis.1–5 The 5-year survival rate after surgery has recently been reported to be between 5–13%.1–5 The clinical course of gallbladder carcinoma has been thought to depend on the depth of tumor invasion.6–8 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 and the 5-year postsurgical survival rate in cases of pT2 carcinoma, about 50–80%, is intermediate between that of pT1 carcinoma and that of pT3 and pT4 carcinomas.9–11 Neither mode of recurrence nor postsurgical prognosis of pT2 gallbladder carcinoma has been found to correlate with any parameters of histopathological malignancies such as lymphatic permeation or venous permeation in the subserosal layer.12 This might be attributed to the variety of ways the tumor progresses. Factors affecting the progression of less-advanced pT2 carcinoma have not been fully elucidated.
An association of the carcinoma with cholelithiasis13 or an anomalous arrangement of the pancreatobiliary duct14, 15 has suggested that long-term inflammation may modulate tumorigenesis or progression of the carcinoma. Supporting this notion, in situ lesions of dysplasia are frequently found in the epithelia adjacent to the gallbladder carcinoma associated with gallstones.16 In addition, inflammatory changes are often observed in the noncancerous epithelia adjacent to the advanced carcinoma.16 Because a number of studies have shown that arachidonate17 and prostaglandin (PG) E2 levels18, 19 are increased in human carcinoma tissues, interest has been focused on the expression of inflammatory enzymes on the arachidonate cascade in carcinoma of the gallbladder.
Cyclooxygenase (COX) is a rate-limiting enzyme in PG synthesis.20 A second isoform of COX, COX-2, is induced by mitogens,21 cytokines,21 and growth factors,22 and produces PG-involved inflammation23 and cell growth.24 Recent studies25–28 have shown that COX-2 mRNA expression is markedly elevated in most tissues of human colorectal carcinoma and have suggested a putative role of COX-2 in tumorigenesis, growth and progression of the carcinoma. Notifying a biological function of COX-2 in carcinoma tissues, overexpression of COX-2 in carcinoma cells has been found to be associated with biochemical changes, including activation of membrane metalloproteinase and PG synthesis,29 which in turn cause phenotypic change of increased invasiveness of the carcinoma cells.29
In this retrospective analysis, the immunohistochemical expression of COX-2 was studied in formalin-fixed, paraffin-embedded surgical specimens from patients with gallbladder carcinomas of different depths of invasion (pT1-pT4) and the results were compared. Carcinoma cells in the deepest sites of the invasion have been thought to have a greater capability to invade and metastasize than do carcinoma cells in other regions.30, 31 Therefore, in the 33 cases of pT2 carcinoma, correlations of the expression level of COX-2 at the deepest invading sites in the subserosal layer of pT2 carcinoma, as a predictor for the invasive/metastatic potential, with clinicopathological findings, mode of recurrence and postsurgical survival were investigated. Increased COX-2 expression in the stroma adjacent to the cancerous epithelia in the subserosal layer was found to be correlated with aggressiveness of the disease, such as the tendency to form distant recurrences. As for epithelial-stromal interactions, upregulation of COX-2 in the stroma and its biological effect on the carcinoma cells may contribute to tumor growth and progression of gallbladder carcinoma.
MATERIAL AND METHODS
Specimens from 52 patients (10 males and 42 females) with gallbladder carcinoma (5 patients with pT1, 33 patients with pT2, 7 patients with pT3 and 7 patients with pT4 carcinoma) were included in the present study. All cases of pT1 and pT2 carcinomas had been curatively resected with a free surgical margin. The mean age of the patients was 66 ± 2 years (mean ± SEM) (range, 35–86 years). The patients were diagnosed as having gallbladder carcinoma and underwent operations between October 1976 and October 1999 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 (TNM) classification of the American Joint Committee on Cancer (AJCC):32 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 1 adjacent organ, or both (extension of 2 cm or less into the liver); and pT4, a tumor that has extended more than 2 cm into the liver or into 2 or more adjacent organs. The resection procedures are shown in Table I. For pT1 carcinoma, simple cholecystectomy was performed in 4 of the 5 patients and cholecystectomy combined with bile duct resection was performed in the other patient. For pT2 carcinoma, simple cholecystectomy was performed in 19 of the 33 patients, cholecystectomy combined with bile duct resection was performed in 8 patients, cholecystectomy with combined bile duct resection and hepatic resection was performed in 3 patients and cholecystectomy combined with pancreatoduodenectomy together with bile duct resection and hepatic resection was performed in 3 patients. For pT3 and pT4 carcinomas, simple cholecystectomy was performed in 4 of the 14 patients, cholecystectomy combined with bile duct resection was performed in 1 patient, cholecystectomy with combined bile duct resection and hepatic resection was performed in 7 patients and cholecystectomy combined with pancreatoduodenectomy together with bile duct resection and hepatic resection was performed in 2 patients. Histological examination revealed that all cases of pT2 carcinoma, in which the tumor had invaded perimuscular connective tissues with no extension beyond the serosa, had neither hepatic infiltration of carcinoma nor invasion into the hepatoduodenal ligament.
|pT1 (n = 5)||pT2 (n = 33)||pT3 · pT4 (n = 14)|
|Bile duct resection||1||14||10||4||10|
Follow-up periods until April 2000 ranged from 2.2–183.8 months. Of the 33 patients, 19 were alive as of April 2000, 11 had died from postsurgical recurrences (Table II) 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 analyses were performed using Stat View and its survival tools (Abacus Concepts, Berkeley, CA) for Macintosh.
The 33 patients with pT2 carcinoma were divided into 2 groups based on the intensity of stromal staining of COX-2 at the deepest invading sites in the subserosal layer: 1 group of cases in which the intensity of COX-2 staining in the stromal cells adjacent to cancerous epithelia was graded as G0 or G1 (Group A) and the other group of cases in which the intensity in the stromal cells was graded as G2 (Group B).
2Distant organ includes the liver in 1 patient.
3Distant organs included the liver in 4 patients, lung in 1, brain in 1 and bones in 1.
4p < 0.05, significantly different between A and B.
In addition, normal gallbladder specimens and specimens with chronic inflammatory changes were obtained at surgery from 10 patients who had undergone hepatectomy because of metastatic liver carcinoma and from 26 patients with gallbladder stones who had undergone cholecystectomy, respectively.
Immunohistochemical expression of cyclooxygenase-2 in gallbladder carcinoma
Immunostaining of COX-2 was performed by the avidin-biotin complex technique using a Vectastain Elite ABC kit (Vector, Burlingame, CA), as described previously.28, 33 Formalin-fixed, paraffin-embedded specimens were serially sectioned at a thickness of 4 μm, placed onto microscopic slides and then deparaffinized and rehydrated. After blocking endogenous peroxidase activity in 0.3% hydrogen peroxide/methanol for 30 min, antigen retrieval was performed by microwave treatment in 0.1 M citrate buffer (pH 6.0). Then non-specific binding sites were blocked with 0.3% normal goat serum diluted to 1:66.7 in PBS for 20 min. The primary antibody raised against human COX-234 (IBL18515: Immuno-Biological Laboratories, Gumma, Japan) was used at a dilution of 1:100. It was applied to tissue sections and incubated in a humidified chamber at room temperature for 60 min. After washing the sections with PBS for 10 min, biotinylated goat antirabbit IgG (Vector) was applied to the tissue sections and incubated at room temperature for 10 min. After washing for 15 min, a streptavidin peroxidase reagent was applied and incubated at room temperature for 10 min. Finally, the reaction product was visualized using developing color by incubating the slides in a solution of 0.3% hydrogen peroxide, diaminobenzidine tetrahydrochloride (DAB) and PBS. A negative control was made using bovine serum albumin instead of the antibody against COX-2. Counter staining was done with hematoxylin. Specificity was determined by preabsorption of the anti-COX-2 antibody with the COX-2 synthetic polypeptide, which was used as an immunogen (17 amino acids, position 251–267: TVKDTQAEMIYPPQVPE) for generation of the antibody,34 before staining. Immunostaining with normal rabbit serum and anti-COX-2 antibody absorbed with the synthetic COX-2 polypeptide was completely negative.
Evaluation of the 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 and adjacent stroma cells in each section was counted and the immunohistochemical expression of COX-2 in gallbladder carcinoma was evaluated in terms of the intensity and positive rate of the immunostaining in both the cancerous epithelia and the adjacent stroma cells. The intensity was defined by comparing the intensity in smooth muscles or vascular endothelia as internal built-in controls, as described previously.33 Figure 1 shows pictures representing the intensities of immunostaining of COX-2 in gallbladder tissues. The intensity was graded as 0, 1, or 2: Grade 0 (G0), an intensity less than that in internal controls (smooth muscles) or negative staining; Grade 1 (G1), an intensity in cancerous epithelia or adjacent stroma cells of gallbladder carcinoma similar to that in internal controls; Grade 2 (G2), an intensity greater than that in internal controls. When more than 5% of the total number of cancerous epithelia or adjacent stroma cells in each section was scored as Grade 1 or 2, the section was judged as being positive for COX-2 staining and then the positive rate (expressed as percentage) was calculated by counting the epithelia or stroma cells expressing COX-2.
RNA isolation and complementary DNA synthesis
Total RNA was isolated from the gallbladder carcinoma specimens, which had been frozen in liquid nitrogen and stored at −80°C, using TriZol reagent by the modified method described by Chomczynski and Sacchi.35 First-strand complementary deoxyribonucleic acids (cDNAs) were synthesized from total RNA with Moloney murine leukemia virus reverse transcriptase by the random primer method.
Reverse transcriptase-polymerase chain reaction
Semiquantitative reverse transcriptase-polymerase chain reaction (RT-PCR) was performed using a DNA Thermal Cycler (model PJ 2000; Applied Biosystems, Inc., Foster City, CA). PCR was subjected to each cycle (G3PDH, 20; COX-2, 30) at 94°C for 1 min, at 55°C for 2 min and at 72°C for 2 min. Aliquots of the reaction mixture were electrophoresed on a 2% agarose gel. PCR primers were designed from cDNA sequences for human COX-236 and then synthesized using an Applied Biosystems DNA synthesizer (model 392; Applied Biosystems., Inc., Foster City, CA) as follows: G3PDH, sense 5′-GAACGGGAAGCTCACTGGCATGGC-3′, antisense 5′-TGAGGTCCACCACCCTGTTGCTG-3′; COX-2, sense 5′-AAGCCTTCTCTAACCTCTCC-3′, antisense 5′-TAAGCACATCGCATACTCTG-3′. A plasmid vector into which the objective coding region of human COX-236 had been inserted was used as a positive control. In each experiment, RT-PCR was done in triplicate. In the quantitative assessment, the amounts of fluorescence intensity were measured using a FluorImager (Molecular Dynamics, Sunnyvale, CA). The data were expressed relative to the amount of G3PDH mRNA present in each specimen and then averaged. The semiquantitative measurement of RT-PCR was assessed using various amounts of template RNA obtained from the liver tissue specimens. The density of the band of the PCR product stained by ethidium bromide was observed to increase proportionately with template RNA for each PCR reaction. Linearity between template RNA amount and PCR product was obtained when cDNA products reverse-transcribed using total RNA of <100 ng were amplified at 30 cycles.
Synthesis of cRNA probe of cyclooxygenase-2
A riboprobe was synthesized from a plasmid vector into which the objective coding region of COX-236 had been inserted. Briefly, pBluescript KS containing the coding region of human COX-2 was prepared in the laboratory of Dr. T. Tanabe (National Cardiovascular Center Research Institute, Osaka, Japan). The inserted coding region was a 414-bp fragment of human COX-2 (from 7,634 to 8,047 of D28235, GenBank/EMBL Data Bank36). The plasmid was linearized and an antisense RNA probe was transcribed with T3 RNA polymerase (for human COX-2) in the presence of cytidine 5′-(α-thio) triphosphate [35S] to a specific activity of 1.0 × 109 cpm/μg.
In situ hybridization of cyclooxygenase-2 in gallbladder carcinoma
Tissues of gallbladder carcinoma were immediately frozen in liquid nitrogen and stored at −80°C. Frozen sections of 8 μm in thickness were cut on a cryostat and thaw-mounted onto poly-L-lysine-coated slides. The sections were dried at room temperature and fixed with 4% formaldehyde in phosphate-buffered saline (PBS) for 9 min, rinsed in PBS twice and acetylated with 0.25% acetic anhydride in 0.1 M triethanolamin/0.9% NaCl for 9 min at room temperature. After dehydration, the sections were air-dried and stored at −80°C until use.
In situ hybridization was performed as described previously.37 Hybridization was carried out in a buffer containing 50% formamide, 2× SSC, 10 mM Tris-Cl, pH 7.5, 1 × Denhardt solution, 10% dextran sulfate, 0.2% SDS, 100 mM DTT, 500 μg/mL sheared single-stranded salmon sperma DNA and 250 μg/mL yeast tRNA. A riboprobe preheated at 80°C for 3 min in 1 M DTT was added to the hybridization buffer at 7 × 104 cpm/μL. The hybridization solution was applied to the sections, which were then covered with a coverslip and sealed by rubber cement. After incubation at 57°C for 5 hr, the slides were immersed in 2× SCC to remove the coverslips, washed for 1 hr by warming in 2× SCC, 10 mM Tris-Cl, pH 7.5 and 1 mM EDTA and then washed again in 0.1× SSC at 60°C for 1 hr. After dehydration in an ascending ethanol series, the slides were air-dried and exposed to Amersham b-max film for 2 weeks at room temperature or dipped in NTB-2 emulsion (Eastman Kodak, NY) diluted 1:1 with distilled water. After exposure for 3–6 weeks at 4°C, the dipped slides were developed in COPINAL (FUJIFILM, Tokyo, Japan) diluted 1:2 in distilled water, fixed and counterstained with hematoxylin-eosin.
Values are given as means ± SEM. A 2-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 logrank test. A p-value of <0.05 was defined as statistically significant.
Immunohistochemical expression of cyclooxygenase-2 in gallbladder carcinoma
In specimens of normal gallbladders and noncancerous pathological lesions of the gallbladder (hyperplasia, pseudopyloric gland metaplasia and dysplasia), immunostaining of COX-2 was observed mostly in the epithelia and smooth muscles (Figs. 1a, 2a–c). In specimens of pT2–4 gallbladder carcinomas, however, intense staining (G2 intensity) was observed in both the cancerous epithelia and the stroma adjacent to the epithelia (Fig. 2d). The COX-2 stainings in the stroma included fibroblasts (arrowheads in Fig. 2d), vascular endothelial cells and inflammatory mononuclear cells.
The immunohistochemical expressions of COX-2 were studied in normal gallbladders, different types of noncancerous pathological lesions of the gallbladder and pT1–pT4 gallbladder carcinomas (Table III). In the epithelia, immunostaining of COX-2 was observed in more than 80% of normal epithelia, noncancerous pathological lesions of the gallbladder except for intestinal metaplasia and pT1–4 carcinoma specimens (Table III). No significant differences were found between the expression rates of COX-2 in the normal epithelia, noncancerous pathological lesions (except for intestinal metaplasia) and cancerous epithelia. COX-2 staining of G2 intensity was observed in large percentages of hyperplastic lesions (65%), pT2 carcinoma specimens (76%) and pT3 and pT4 carcinoma specimens (64%) compared to the percentages of normal epithelia (0%) and other pathological lesions, pseudopyloric gland metaplasia (20%), intestinal metaplasia (0%) and dysplasia (25%). The differences between these percentages were statistically significant (Table III). COX-2 staining of G1 intensity was observed in large percentages of normal epithelia (80%) and pathological lesions, pseudopyloric gland metaplasia (67%) and dysplasia (75%). Intestinal metaplastic lesions were mostly negative for epithelial staining. In the stroma, COX-2 staining of G2 intensity was observed in 33% of pT2 carcinoma specimens and in 43% of pT3 and pT4 carcinoma specimens. The expression rate of stromal G2 intensity was significantly higher in pT2 carcinoma and pT3 and pT4 carcinoma specimens than in normal epithelia (0%) and in pathological lesions, hyperplasia (5%), pseudopyloric gland metaplasia (0%) and dysplasia (8%) (Table III). COX-2 staining of G1 intensity was observed in the stroma of pathological lesions, hyperplasia (15%), pseudopyloric gland metaplasia (20%), dysplasia (17%) and in the stroma of pT1 carcinoma (20%), pT2 carcinoma (42%) and pT3 and pT4 carcinoma specimens (43%). Intestinal metaplastic lesions as well as normal epithelia were negative for stromal staining.
The staining of COX-2 in the cancerous epithelia appeared to be independent of the depth of the invasion of gallbladder carcinomas. Parallel to the depth of invasion, however, the carcinoma tissues showed an intense staining of COX-2 in the stroma adjacent to the cancerous epithelia. This observation was also consistent with the data of COX-2 in immunoblot analysis.
Tissue mRNA levels of cyclooxygenase-2 in gallbladder carcinoma
The tissue steady-state mRNA level of COX-2 was determined by reverse transcriptase-polymerase chain reaction (RT-PCR). Figure 3 shows the PCR-assisted amplification of COX-2 in the tissues of gallbladder carcinomas of different depths of invasion. The mRNA levels of COX-2 were significantly higher in 4 specimens of pT1 carcinoma (79 ± 6% of G3PDH, mean ± SEM, p < 0.05), 12 specimens of pT2 carcinoma (100 ± 6%, p < 0.01) and 6 specimens of pT3 and pT4 carcinomas (157 ± 14%, p < 0.01) compared to the level in 10 specimens of normal gallbladders (48 ± 2%). In terms of the depth of invasion, the mRNA level was significantly higher in pT3 and pT4 carcinomas than in pT1 or pT2 carcinoma (p < 0.01).
In situ hybridization of mRNA of cyclooxygenase-2 in gallbladder carcinoma
Paired analysis of immunohistochemistry and in situ hybridization of COX-2 was performed in tissue specimens of pT2 gallbladder carcinoma (Fig. 4a–d). COX-2 mRNA was expressed widely in the epithelia (Fig. 4c) and focally in the adjacent stroma (Fig. 4d), as well as in the smooth muscles (data not shown), in the specimens of pT2 gallbladder carcinoma tissues. The extent and distribution of COX-2 mRNA resembled those of COX-2 protein in individual cases (Fig. 4a,b).
Relationship between pathological malignancies and intense stromal staining of cyclooxygenase-2 in patients with pT2 carcinoma
The 33 patients with pT2 carcinoma were divided into 2 groups based on the intensity of stromal staining of COX-2 at the deepest invading sites in the subserosal layer: 1 group of cases in which the intensity of COX-2 staining in the stromal cells adjacent to cancerous epithelia was graded as G0 or G1 at the deepest invading sites (Group A) and the other group of cases in which the intensity in the stromal cells was graded as G2 (Group B). A comparison of Groups A and B 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 difference in the parameters of pathological malignancies between the 2 groups (Table IV).
|n||Histological grade||Lymphatic permeation||Venous permeation||Lymph node metastasis||Involved surgical margin|
|pT3 · pT4||14||7||7||2||12||5||9||4||10||6||8|
Relationship between mode of recurrence in patients with pT2 carcinoma and intense stromal staining of cyclooxygenase-2 in the specimens
The postsurgical recurrent modes in patients with pT2 carcinoma in Groups A and B were compared. Of the 22 patients in Group A, 1 (5%) had metastasis in a distant organ (liver) and 7 (32%) had peritoneal dissemination (Table II). Of the 11 patients in group B, 6 (55%) had metastasis in distant organs (liver in 4, lung in 1, brain in 1 and bones in 1) and 3 (27%) had peritoneal dissemination (Table II). Interestingly, it should be noted that in pT2 carcinoma, postsurgical metastasis in distant organs was found to be more frequent in the patients in Group B than in those in Group A. The difference was statistically significant (p < 0.05). Intense stromal staining of COX-2 at the deepest invading sites may therefore be an important predictor of postsurgical metastasis at distant organs in pT2 carcinoma.
Relationship between postsurgical survival of patients with pT2 carcinoma and intense stromal staining of cyclooxygenase-2 in the specimens
For the 33 patients with pT2 gallbladder carcinoma, a comparison of Groups A and B was made with special reference to postsurgical survival and the results of the analysis showed that the survival rate of the 11 patients in Group B was significantly lower than that of the 22 patients in Group A (p = 0.010; Fig. 5a). The 5-year survival rates were estimated to be 83% for the patients in Group A and 22% for the patients in Group B. To exclude the influence of lymph node metastasis on postsurgical survival, a comparison was made between the 16 patients in Group A without metastasis and the corresponding 7 patients in Group B. The analysis revealed that the survival rate of the patients in Group B was significantly lower than that of the patients in Group A (p = 0.024; Fig. 5b). The 5-year survival rates were estimated to be 83% for the patients in Group A and 0% for the patients in Group B.
The clinical outcome of patients with gallbladder carcinomas depends on the depth of invasion.10 Despite a theoretical advantage for gallbladder carcinoma not involving the serosa, however, the postsurgical prognosis of pT2 carcinoma is not necessarily favorable.9–11 This may be because approximately half of the patients have malignant infiltration into the lymphatic, venous and perineural spaces and that the frequency of lymph node metastasis is 50%.9, 10 Among the 33 patients who had undergone surgery with curative intent for less-advanced pT2 gallbladder carcinoma, the 11 patients with intense stromal staining of COX-2 at the deepest invading sites in the subserosal layer had a poorer postsurgical survival rate, because of metastasis at distant organs (Table II), than did the 22 patients without (Fig. 5a). This was also true for the 22 cases of pT2 carcinoma without lymph node metastasis (Fig. 5b). Intense stromal staining of COX-2 may be involved in metastatic processes of the carcinoma cells into distant organs through means other than lymphatic permeation or lymph node metastasis. Thus, the results indicate that gallbladder carcinoma cells in the deepest invading sites showing the intense stromal staining have a strong potential for metastasis and that micrometastasis may already have occurred at the time of surgery if the stromal staining was intense at the deepest invading sites. The depth of invasion may affect a biological behavior of the carcinoma cells through an upregulation of COX-2 expression in the stroma.
In the present study, the immunohistochemical expression of COX-2 at the deepest invading sites correlated with neither histological grades nor parameters of histopathological malignancies in the 33 cases of pT2 gallbladder carcinoma. Because of the poor association, the intense stromal staining of COX-2 at the deepest invading sites may be considered to be an independent prognostic marker for pT2 carcinoma. This phenotype could serve as a unique biological feature associated with the malignant behavior of pT2 carcinoma.
Confirmed by the results of the present study, the expression levels of COX-2 in gallbladder carcinoma were increased parallel to the depth of invasion; in pT3 and pT4 carcinomas of the gallbladder, a substantial increase in COX-2 mRNA and protein levels was observed compared to the levels in pT1 and pT2 carcinomas. By in situ hybridization and immunohistochemistry, increased expression of COX-2 mRNA and protein was observed in the stroma cells (fibroblasts, vascular endothelial cells and inflammatory mononuclear cells) adjacent to the cancerous epithelia of the advanced pT3 and pT4 carcinomas. COX-2 may be induced by unknown factors (such as cytokines or growth factors) in stromal cells located adjacent to the cancerous epithelia of advanced carcinoma. Therefore, the main sources of COX-2 in the tissues of pT3 and pT4 carcinomas may be not only the cancerous epithelia but also the adjacent stroma. Supporting these observations, immunostaining of COX-2 has been observed in the stromal cells in tissues of colonic adenoma39 and colorectal carcinoma.27, 28
Interest should be focused on the biological effects of COX-2 on tumor growth and progression of gallbladder carcinoma. As indicated in several reports,39–42 PGE2 produced by COX-2-expressing carcinoma cells and stromal cells may play an important role in tumor growth and progression. This is because PGE2 may stimulate carcinoma cell proliferation,43 inhibit apoptosis in carcinoma cells,43 promote immunosuppression in carcinoma tissues by preventing activation of inflammatory cells,28 and induce growth factors important for the progression of carcinomas.42 COX-2-derived PGE2 may play an important role in the formation and maintenance of the stroma and vessel structure in carcinoma tissues, because PGE2 stimulates mitogenesis in fibroblasts41 and induces angiogenesis.44, 45 A markedly increased production of hepatocyte growth factor (HGF) in COX-2-expressing human fibroblasts via a PG-mediated pathway46 is most interesting in terms of epithelial-stromal interactions.
Therapeutic options for gallbladder carcinoma are still limited. There has been no evidence available that extended surgery for pT2 gallbladder carcinoma, including hepatectomy or pancreatoduodenectomy, improves postsurgical prognosis of the disease. Moreover, although extended cholecystectomy combined with common bile duct resection and complete regional lymphadenectomy becomes a standard surgical procedure for treatment of pT2 carcinoma,47 the 5-year survival rate was from 50–80%.9–11 Thus in pT2 carcinoma, where the primary tumor can be removed at a less-advanced point in time but where metastases can arise despite a theoretical advantage for carcinoma not involving the serosa, patients may benefit from an antimetastatic chemoprevention strategy. Some adjuvant chemotherapies during the postsurgical period would be necessary in cases of pT2 carcinoma. It will be important to determine on the basis of the results of our study whether COX-2 expression is an independent prognostic factor in pT2 carcinoma through the study of a large number of the cases and also to elucidate on the observation of a decreased metastatic potential of carcinoma cells through treatment with a COX-2 inhibitor29 if selective inhibitors of COX-2 are useful against the development or progression of the disease, especially for the cases showing intense stromal staining of COX-2.
In summary, the results of the present study suggest that in cases of pT2 gallbladder carcinoma upregulation of COX-2 in the stroma adjacent to the cancerous epithelia in the subserosal layer correlates with the aggressiveness of the disease, such as the tendency to form distant recurrences. The mechanistic basis for the upregulation of COX-2 and intense stromal expression at the deepest invasive sites in pT2 carcinoma with poor prognosis remains to be further elucidated.
- 12Clinical study on gallbladder cancer invading into subserosal layer with special reference to relationship between extent of cancer and prognosis. J Jpn Biliary Assoc 1993;7: 133–8., , , et al.
- 16Malignant tumors of the gallbladder. In: HartmannWH, editor. Tumors of the gallbladder and extrahepatic bile ducts. Washington DC: Armed Forces Institute of Pathology, 1986. 28–123., .
- 32AJCC Cancer Staging Manual. 5th ed. Philadelphia, Pennsylvania: Lippincott-Raven, 1997., , , et al.
- 44Cyclooxygenase gene expression in inflammation and angiogenesis. Ann NY Acad Sci 1993;686: 197–204., , , et al.
- 45Role of prostaglandin E1 and copper in angiogenesis. J Natl Cancer Inst 1982;6: 475–82., , .