1691 Lack of CD44 Variant 6 Expression in Advanced Extrahepatic Bile Duct/Ampullary Carcinoma Yujiro Yokoyama, M.D.1 Eiso Hiyama, M.D.2 Yoshiaki Murakami, M.D.1 Yuichiro Matsuura, M.D.1 Takashi Yokoyama, M.D.2 1 First Department of Surgery, Hiroshima University Faculty of Medicine, School of Medicine, Hiroshima, Japan. 2 Department of General Medicine, Hiroshima University Faculty of Medicine, School of Medicine, Hiroshima, Japan. BACKGROUND. Transmembrane proteins of the CD44 family play roles in cell-cell and cell-matrix interactions, and their aberrant expression has been reported to be associated with the growth and metastasis of various tumors. The authors examined CD44 standard (CD44st) and CD44 variant 6 (CD44v6) expression in extrahepatic bile duct (EHBD)/ampullary carcinoma. METHODS. In 36 EHBD/ampullary carcinomas, immunohistochemical analyses with monoclonal antibodies against the human CD44st protein or CD44v6 protein were performed by the streptavidin-biotin immunoperoxidase method. The relation between expression of the proteins and lymph node metastases or patient outcome was investigated. To verify the lack of CD44v6 mRNA expression in tumors negative for CD44v6 staining, reverse transcriptase-polymerase chain reaction (RT-PCR) and Southern blot hybridization were performed. RESULTS. Immunohistochemical analyses revealed that 13 of 36 carcinomas (36.1%) expressed the CD44v6 protein. Only 2 of 13 CD44v6-expressing primary tumors (15.4%) had regional lymph node metastases, whereas 14 of 23 tumors showing no CD44v6 expression (60.9%) had lymph node metastases (Fisher exact test, P , 0.01). Moreover, a lack of CD44v6 expression was correlated significantly with poor prognosis (generalized Wilcoxon test, P , 0.05). Eleven of 13 CD44v6 positive tumors showed CD44st expression, which also was correlated with prognosis. RT-PCR and Southern blot hybridization revealed a lack of CD44v6 mRNA expression in the tumors that were negative for CD44v6 staining. CONCLUSIONS. The results of the current study suggest that a lack of expression of CD44, especially CD44v6, is correlated with lymph node metastases and poor prognosis, and may be a prognostic factor for patients with EHBD/ampullary carcinoma. Cancer 1999;86:1691–9. © 1999 American Cancer Society. KEYWORDS: CD44, extrahepatic bile duct, ampulla of Vater, cancer, lymph node metastases, prognosis. Presented at the 16th World Congress Collegium Internationale Chirurgiae Digestive, Madrid, Spain, September 16 –19, 1998. The authors thank Dr. T. Kodama for his helpful suggestion and Ms. Fukuba for her excellent technical assistance. Address for reprints: Yujiro Yokoyama, M.D., First Department of Surgery, Hiroshima University Faculty of Medicine School of Medicine, 1-2-3 Kasumi, Minami-ku, Hiroshima, 734-8551, Japan. Received December 31, 1998; revision received April 29, 1999; accepted June 15, 1999. © 1999 American Cancer Society T umor metastases and invasion are the major causes of death in patients with cancer. Because the extrahepatic bile duct (EHBD) has no thick muscular layer, bile duct carcinoma can disseminate easily. Moreover, invasion to large vessels, such as the portal vein or the hepatic artery, occurs frequently, making surgical extirpation difficult. Recently, cell adhesion molecules, such as CD44, E-cadherin, KAI1 (CD82), and integrin, have been reported to play major roles in invasion and metastases of a variety of epithelial and nonepithelial malignant cells. In gastric carcinoma, a decrease of E-cadherin expression was reported to be correlated with poor prognosis of the patients.1 KAI1 protein expression decreased during the progression of prostatic carcinoma.2 Decreased expression of the integrin a-5 subunit was reported to 1692 CANCER November 1, 1999 / Volume 86 / Number 9 be correlated with more malignant phenotypes of human hepatocellular carcinoma.3 CD44, which is one of the well-known cell adhesion molecules, is a polymorphic transmembrane glycoprotein whose gene is located at chromosome 11p and comprises 20 exons that span 50 kilobases.4 CD44 molecules are expressed as several isoforms generated by alternative splicing of exons of the CD44 gene. Recently, correlations between clinicopathologic features of various cancers and CD44 expression have been reported. In solid malignant tumors, high levels of expression of CD44 protein have been reported in advanced breast carcinoma,5 gastric carcinoma,6 multiple myeloma,7 colorectal carcinoma,8 –10 and vulval carcinoma.11 Conversely, CD44 expression has been reported to decrease in high grade and poorly differentiated types of bladder carcinoma,12 prostate carcinoma,13 ovarian carcinoma,14 neuroblastoma,15,16 endometrial carcinoma,17 and oral squamous cell carcinoma.18 Thus, the correlation between CD44 expression and tumor malignancy appears to differ in different types of tumors. Because few detailed studies of CD44 expression in EHBD/ampullary carcinoma have been reported, we analyzed the expression of CD44 variant 6 (CD44v6) and CD44 standard (CD44st) in these malignancies and examined the correlation between the expression of these molecules and the clinicopathologic profiles of the patients. MATERIALS AND METHODS Patients Thirty-six patients with carcinoma of the EHBD/ampulla who underwent surgical resection in the First Department of Surgery, Hiroshima University Hospital, were enrolled in this study. Patient characteristics are shown in Table 1. Clinicopathologic Features There were 28 male patients and 8 female patients, and their mean age at surgery was 63.6 years (range, 41– 83 years). Pathologic classification and clinical staging were performed according to the criteria of the International Union Against Cancer (UICC).19 Among the 36 patients with EHBD/ampullary carcinomas, 25 patients had EHBD carcinoma (8 patients with Grade 1, 16 with Grade 2, and 1 with Grade 3 histopathology), and 11 patients had ampullary carcinoma (6 patients with Grade 1, 4 with Grade 2, and 1 with Grade 3 histopathology). There 6 patients had pT1 tumors, 21 patients had pT2 tumors, 8 patients had pT3 tumors, and 1 patient had a pT4 tumor. Twenty patients had no lymph node metastases, and 16 patients had more than one lymph node metastasis. Six patients had Stage I disease, 11 patients had Stage II disease, 14 patients had Stage III disease, and 5 patients Stage IV disease. Twelve patients died, and 13 were alive and disease free of the patients with EHBD carcinoma. Three patients died, and 8 patients were alive and disease free of the patients with ampullary carcinoma. The mean follow-up time for these patients was 51.7 months (range, 0.8 –143.8 months) after surgical resection. Immunohistochemistry Tissues were fixed in 10% buffered formalin at room temperature and embedded in paraffin for histologic diagnosis. From these tissues, serial sections were cut, mounted on gelatin-dichromate-coated glass slides, deparaffinized, rehydrated, and pretreated with 3% hydrogen peroxide for 15 minutes at room temperature to quench endogenous peroxidase activity. Then, the sections were incubated with 10% normal rabbit serum for 1 hour at room temperature and incubated with the antihuman CD44v6 monoclonal antibody (2F10; R&D Systems, Minneapolis, MN) and antiCD44st monoclonal antibody (F10-44-2; DAKO; Carpinteria, CA) at 4 °C overnight in a moist chamber. After washing in phosphate-buffered saline (PBS), the sections were incubated with biotinylated rabbit antimouse immunoglobulin G (IgG), IgM, and IgA (Nichirei, Tokyo, Japan) for 1 hour at room temperature, followed by rinsing in PBS, and then incubated with peroxidase-conjugated streptavidin (Nichirei). Positive sites were developed with 200 mg/mL of 3,39diaminobenzidine containing 0.04% hydrogen peroxide in PBS. Finally, the sections were counterstained with Mayer hematoxylin and mounted. Negative control sections, which were incubated as described above but without the primary antibody, showed no staining. Samples that showed strong plasma membrane-type staining pattern were interpreted as positive, and samples that showed either no staining or only cytoplasmic staining were interpreted as negative.13 RNA Isolation Total RNA was extracted from two noncancerous EHBD epithelium specimens and two EHBD carcinoma specimens, using the acid-guanidine-phenolchloroform (AGPC) method protocol.20 The purity and concentration of RNA were determined with a spectrophotometer by calculating the A260/A280, ratio and RNA samples were electrophoresed on 1.5% agarose gels to confirm the presence of high-quality 18s and 28s bands. CD44v6 in Bile Duct-Ampullary Carcinoma/Yokoyama et al. 1693 TABLE 1 Patient Features and Immunohistochemical Detection of CD44 Age (yrs)/gender Location Prognosis (months)/status Grade pTNMa Stage CD44v6 CD44st 53/M 60/F 50/M 41/M 57/M 71/M 62/M 72/M 83/M 72/M 58/M 70/Mb 60/Mc 60/F 73/F 67/M 64/F 61/M 58/M 67/F 62/M 63/F 68/M 64/M 77/M 69/M 57/F 72/F 59/M 62/M 66/M 70/M 52/M 71/M 67/M 50/M EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD EHBD Va. Pap. Va. Pap. Va. Pap. Va. Pap. Va. Pap. Va. Pap. Va. Pap. Va. Pap. Va. Pap. Va. Pap. Va. Pap. 143/A 143/A 123/A 121/A 106/A 4/D 86/A 61/A 53/A 14/D 12/D 12/D 8/D 13/D 18/D 12/D 91/A 1/D 21/D 50/A 52/A 45/A 28/D 42/A 32/D 11/D 136/A 114/A 10/D 77/A 58/A 52/A 49/A 17/A 8/D 41/A 2 1 2 1 1 2 1 2 1 1 1 2 2 2 2 2 2 2 2 2 1 3 2 2 2 3 1 1 2 1 2 1 2 1 2 1 pT2pN0pM0 pT2pN0pM0 pT1bpN0pM0 pT3pN0pM0 pT2pN0pM0 pT3pN2pM0 pT1apN0pM0 pT2pN0pM0 pT2pN0pM0 pT3pN0pM0 pT3pN2pM0 pT2pN2pM0 pT2pN2pM0 pT2pN2pM0 pT2pN2pM0 pT2pN1pM0 pT2pN2pM0 pT2pN2pM0 pT2pN0pM0 pT2pN2pM0 pT2pN0pM0 pT2pN1pM0 pT2pN0pM0 pT2pN0pM0 pT2pN0pM0 pT3pN0pM0 pT1pN0pM0 pT2pN1pM0 pT3pN1pM0 pT2pN0pM0 pT3pN1pM0 pT1pN0pM0 pT4pN0pM0 pT1pN0pM0 pT3pN1pM0 pT1pN0pM0 II III I IV-A II IV-A I II II IV-A IV-A III III III III III III III II III II III II II II II I III III II III I IV I III I N N N N P N P P P P N N N N N N P N N N N N N P N N P P N N N P P P N P N N N N P N P N P P N N N N N N P N N N N N N P N N P N N N N P N P N P M: male; F: female; EHBD: extrahepatic bile duct; Va. Pap: Vater papilla; A: alive; D: dead; N: negative; P: positive. a Tumor classification is according to the TMN classification system (Hermanek and Sobin19). b Case 12. c Case 13 (shown in Fig. 2). Reverse Transcriptase and Polymerase Chain Reaction The first strand of cDNA was synthesized from purified total RNA using Superscript II™ (Gibco BRL Life Technologies, Grand Island, NY). The procedure followed the manufacturer’s protocol, which can be summarized as follows: The first-strand cDNA synthesis reaction from 1.0 mg of total RNA was catalyzed by RNase H- reverse transcriptase (RT) using oligo (dT) to hybridize to 3-poly (A) tails. The cDNA was amplified for 30 cycles of polymerase chain reaction (PCR) under the following conditions: 94 °C for 1 minute, 55 °C for 1 minute, and 72 °C for 2 minutes. The sequences of the two primers, P1 and P2, in this PCR were as follows21: primer P1: 59-CAGACCTGCCCAATGCCTTTGATGGACC; primer P2: 59-CAAAGCCAAGGCCAAGAGGGATGCC. These primers were against a 445base pair (bp) portion of the basic CD44 standard gene. The PCR was carried out in a final volume of 50 mL containing the following: 1.0 mL cDNA; 1.5 mM MgCl2, 0.25 mM each of dNTP; 25 mM each of primer (P1 and P2); 25 mM ammonium sulfate buffer, pH 9.5; and 2.5 U Taq polymerase (TaKaRa, Kyoto, Japan). Five microliters of the PCR product were electrophoresed in a 2% agarose gel and stained with ethidium 1694 CANCER November 1, 1999 / Volume 86 / Number 9 bromide. RNA from each sample was processed in parallel without the using the reverse transcriptase (RT) as a control to determine that the product obtained was of cDNA and not a result of contamination by genomic DNA. Moreover, mRNA of a housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (G3PDH), was amplified by RT-PCR in each case as a marker of the integrity of the extracted RNA. Southern Blot Hybridization Twenty microliters of each PCR product were electrophoresed in 1% agarose gels and subjected to Southern blot transfer onto nitrocellulose filters. CD44 exon 11 (v6) specific probe was PCR amplified from genomic DNA derived from normal human peripheral blood cells with the following primers22: primer P3: 59 ACTCAGGCAACTCCTAGTAGTACAACG-39 primer P4: 59-TGTCCCTGTTGTCGAATGGGAGTCTTC-39 The amplified fragment (125 bp) was labeled by a-32P-dTP using the random primer method and then hybridized with the filter. After washing three times, the filter was exposed for 12 hours to Fuji X-ray film (Fuji, Tokyo, Japan) at room temperature. Statistical Analysis Comparisons between CD44v6 expression in tumor tissues and clinicopathologic parameters were performed using the chi-squared test or the Fisher exact test, as appropriate. The Kaplan–Meier survival curves were used to determine survival rates,23 and the differences in survival rates between subgroups of patients were compared using the generalized Wilcoxon test. RESULTS Immunohistochemistry CD44v6 expression was examined in 29 noncancerous EHBD samples but not in the remaining 7 samples, which had lost noncancerous EHBD epithelium due to chronic cholangitis. Two of 29 samples (6.9%), which were degenerated severely and detached from the bile duct wall, showed no immunoreactivity. The remaining 27 noncancerous epithelial tissue samples (93.1%) showed strong plasma membrane-type staining patterns. A similar expression pattern was observed in 13 tumors but not in the remaining 23 tumor samples (63.9%). Figure 1 shows noncancerous EHBD epithelium (Fig. 1A) and a cancerous specimen (Fig. 1B) from the same patient. Noncancerous EHBD samples showed positive staining patterns for CD44st similar to those for CD44v6. All CD44v6 negative samples also were negative for staining by CD44st. Except for 3 samples, all CD44v6 positive samples were stained positively for CD44st. RT-PCR and Southern Blot Hybridization To analyze the mRNA transcription levels in the tissues without expression of CD44v6 as detected by immunostaining, RT-PCR and Southern blot hybridization were performed as follows: The cDNA from the 2 cancerous tissue samples and corresponding noncancerous EHBD epithelium samples were examined to verify the lack of CD44v6 mRNA expression in the CD44v6 negative tumors. The cDNA from 2 noncancerous EHBD epithelium samples yielded prominent, standard-sized CD44 PCR products. The amplified products in these 4 samples varied in size between 445 bp and 1450 bp. Prior to CD44 specific PCR amplification, a housekeeping gene (G3PDH) was amplified successfully using cDNA derived from the purified total RNA in all 4 tissue samples, indicating the integrity of the extracted RNA (data not shown). Negative controls for each of these 4 samples using the cellular RNA were processed in parallel without the RT enzyme to rule out the possibility of genomic DNA contamination. Hybridization with a CD44v6 specific probe yielded a complex pattern of bands in the two noncancerous EHBD epithelium specimens analyzed. However, CD44v6 mRNA was not detected in the 2 EHBD carcinoma specimens (Fig. 2). CD44v6 Status and Clinicopathologic Parameters The correlation between CD44v6 expression in EHBD/ampullary carcinoma and clinicopathologic findings is shown in Table 2. In histopathologic grading, only 4 CD44v6 positive tumors (30.8%) were Grade 2 or Grade 3; conversely, 18 CD44v6 negative tumors (78.3%) were Grade 2 or Grade 3 (P , 0.05). Among the EHBD carcinoma samples, 14 of 18 CD44v6 negative tumors (77.8%) were Grade 2 or Grade 3, and 4 of 7 CD44v6 positive tumors (57.1%) were Grade 1. Among ampullary carcinoma samples, 4 of 5 CD44v6 negative tumors (80.0%) and only 1 CD44v6 positive tumor were Grade 2 or Grade 3, but not significantly. Only 2 of 13 CD44v6 positive tumors (15.4%) had invaded regional lymph nodes, whereas 14 of 23 CD44v6 negative tumors (60.9%) had lymph node metastases (P , 0.01). Among the EHBD carcinoma samples, 10 of 18 CD44v6 negative tumors had invaded regional lymph nodes, whereas 6 of 7 CD44v6 positive tumors lacked lymph node metastases. Among the ampullary carcinoma samples, only 1 CD44v6 positive tumors showed lymph node metastases. Eleven of 13 CD44v6 positive tumors (84.6%) were classified as Stage I or II, whereas CD44v6 in Bile Duct-Ampullary Carcinoma/Yokoyama et al. 1695 FIGURE 1. (A) Noncancerous extrahepatic bile duct epithelium with diffuse, strong, membranous reactivity for CD44 variant 6 (CD44v6) (antihuman CD44v6, clone 2F10 with Mayer hematoxylin counterstain; original magnification, 3200). (B) Tumor cells in extrahepatic bile duct carcinoma showing the absence of CD44v6 protein expression (antihuman CD44v6, clone 2F10 with Mayer hematoxylin counterstain; original magnification, 3200). 14 of 23 CD44v6 negative tumors (60.9%) were classified as Stage III or IV, but the differences between CD44v6 positive tumors and CD44v6 negative tumors were not significant. To evaluate the difference between early tumors, the correlation between CD44v6 expression and clinicopathologic findings was investigated in patients with pT1 or pT2 tumors (Table 3). There was a significant correlation between CD44v6 reactivity and histologic grade, and CD44v6 negative carcinoma, especially CD44v6 negative EHBD carcinoma, tended to have more frequent lymph node metastases, but the increase was not significant. CD44v6 Status and Patient Prognosis Figure 3A shows survival curves for the patients with CD44v6 positive ampullary carcinoma and those with CD44v6 negative carcinoma. In these 2 groups, the 5-year overall survival rates were 100% and 40.0%, respectively. Figure 3B shows survival curves for the patients with CD44v6 positive EHBD carcinoma and 1696 CANCER November 1, 1999 / Volume 86 / Number 9 TABLE 2 Correlation between Expression of CD44v6 and Clinicopathologic Parameters in 36 Patients with Extrahepatic Bile Duct and Ampulla Carcinoma Parameter Histologic grade (1:2 or 3)b EHBD Va. Pap Primary tumor (pT1 or 2:pT3 or 4)b EHBD Va. Pap Regional lymph node metastases (negative:positive) EHBD Va. Pap Stage grouping (I or II:III or IV)b EHBD Va. Pap CD44v6 negative CD44v6 positive P valuea 5:18 4:14 1:4 16:7 15:3 1:4 9:4 4:3 5:1 11:2 6:1 5:1 ,0.05 N.S. N.S. N.S. N.S. N.S. 9:14 8:10 1:4 8:15 6:12 2:3 11:2 6:1 5:1 9:4 5:2 4:2 ,0.01 N.S. N.S. N.S. N.S. N.S. EHBD: extrahepatic bile duct; Va. Pap: Vater papilla; N.S.: not significant. a Fisher exact test. b Histopathologic grading, primary tumor, stage grouping: Tumor classification is according to the TMN classification system (Hermanek and Sobin19). TABLE 3 Correlation between Expression of CD44v6 and Clinicopathologic Parameters in 27 Patients with pT1 or pT2 Carcinoma in Extrahepatic Bile Duct and Ampulla Parameter FIGURE 2. Reverse transcriptase-polymerase chain reaction (RT-PCR) and Southern blot hybridization. CD44 PCR products were hybridized with a CD44 variant 6 (CD44v6) specific probe. The products derived from both noncancerous extrahepatic bile duct epithelium samples, but not those from immunohistochemically CD44v6 negative tumors, were hybridized with CD44v6 specific probes. No.: case number (shown in Table 1); T: extrahepatic bile duct/ampullary carcinoma; N: noncancerous extrahepatic bile duct epithelium; bp: base pairs. those with CD44v6 negative carcinoma. In these 2 groups, the 5-year overall survival rates were 85.7% and 38.9%, respectively. The prognosis of patients with CD44v6 negative carcinoma was significantly poorer compared with that of patients with CD44v6 positive carcinoma (generalized Wilcoxon test: P , 0.05). No patients with CD44v6 positive ampullary carcinoma died. Three of 5 patients with CD44v6 negative ampullary carcinoma died: 2 from peritoneal carcinomatoses and 1 from liver metastases. Only 1 of the 7 patients with CD44v6 positive EHBD carcinoma died of liver metastases, whereas 11 of 18 patients with CD44v6 negative EHBD carcinoma died: 4 from peritoneal carcinomatoses, 3 from liver metastases, and 4 Histologic grade (1:2 or 3)b EHBD Va. Pap Regional lymph node metastases (negative:positive) EHBD Va. Pap Stage grouping (I or II:III or IV)b EHBD Va. Pap CD44v6 negative CD44v6 positive P valuea 3:13 2:13 1:0 8:3 3:3 5:0 ,0.05 N.S. N.S. 7:9 7:8 0:1 7:9 6:9 1:0 9:2 5:1 4:1 9:2 5:1 4:1 N.S. N.S. N.S. N.S. N.S. N.S. EHBD: extrahepatic bile duct; Va. Pap: Vater papilla; N.S.: not significant. a Fisher exact test. b Histopathologic grading, stage grouping: Tumor classification is according to the TMN classification system (Hermanek and Sobin19). from disease progression the details of which were unknown. All patients with pT1 or pT2 ampullary carcinoma are alive disease free. Figure 3C shows survival curves for the 2 groups of patients with pT1 or pT2 EHBD carcinoma. In these 2 groups, the 5-year overall survival rates were 100.0% and 40.0%, respectively. Among the patients with pT1 or pT2 EHBD carcinoma, CD44v6 positive carcinoma also was correlated significantly with favorable prognosis (P , 0.05). CD44v6 in Bile Duct-Ampullary Carcinoma/Yokoyama et al. 1697 FIGURE 3. (A) Overall survival curves of patients suffering from ampullary carcinoma with or without CD44 variant 6 (CD44v6) expression. The survival rate of patients with CD44v6 positive tumors was significantly higher compared with that of patients with CD44v6 negative tumors (P , 0.05). (B) Overall survival curves of patients with extrahepatic bile duct carcinoma with or without CD44v6 expression. The survival rate of patients with CD44v6 positive tumors was significantly higher compared with that of patients with CD44v6 negative tumors (P , 0.05). (C) Survival curves of patients with pT1 or pT2 extrahepatic bile duct carcinoma with or without CD44v6 expression. The survival rate of the patients with CD44v6 positive tumors also was significantly higher compared with that of patients with CD44v6 negative tumors (P , 0.05). CD44st Expression The correlations between CD44st expression and clinicopathologic features were similar to those for CD44v6. The 5-year survival rates of patients with CD44st positive and CD44st negative tumors were 90.0% and 46.2%, respectively, and the difference also was significant (P , 0.05). DISCUSSION Lymph node metastasis is one of the factors that is considered to affect surgical outcome in patients with EHBD/ampullary carcinoma.24 –27 Although many imaging techniques, such as ultrasonography, computed tomography, and magnetic resonance imaging, have been performed to detect lymph node metastases, they are difficult to detect in patients with these tumors before surgery. Therefore, other techniques for predicting the existence of lymph node metastases and patient prognosis would be very useful in the determination of therapeutic regimens for improving patient outcome. Some studies have reported that the lack of certain CD44 isoforms is linked to tumor progression and metastases in patients with bladder carcinoma,12 prostate carcinoma,13 ovarian carcinoma,14 neuroblastoma,15,16 endometrial carcinoma,17 and oral squamous cell carcinoma.18 Fujita et al.17 reported that patients with CD44 negative endometrial carcinoma had a high incidence of lymph-vascular space involvement. They postulated that the decrease in CD44 expression in endometrial carcinoma cells may cause reduced adhesion between cells as well between cells and the basement membrane, resulting in easier detachment from the tissue. We believe that, in patients with EHBD/ampullary 1698 CANCER November 1, 1999 / Volume 86 / Number 9 carcinoma, the lack of CD44 expression may lead to decreased cell-cell adhesion, changes in the cells’ arrangement, and thus to detachment of the cells from the basement membrane, enabling cancer cell migration to both local and distant sites, causing increased regional lymph node metastases. Ashida et al.28 reported that CD44st expression was correlated significantly with the absence of metastases in 47 patients with cholangiocarcinoma. Our study also indicated that the prognosis of patients with CD44st positive tumors tended to be more favorable compared with that of patients with CD44st negative tumors, but the difference was not significant. In the current study, almost all noncancerous EHBD epithelium expressed CD44v6 epitope, as shown by immunohistochemical analysis. We therefore consider that normal bile duct epithelium expresses CD44v6 and epitopes. EHBD epithelial cells stained positively for CD44v6 and in the cell membranes adjacent to the basement membrane and neighboring cells. Therefore, CD44 molecules are supposed to mediate adherence of cells to the basement membrane and to other cells in EHBD epithelium. Conversely, 63.9% of EHBD/ampullary carcinoma samples did not express CD44v6 epitope (Table 1), which suggests that some cancer cells had lost their CD44v6 immunoreactivity during carcinogenesis or disease progression. CD44v6 negative tumors frequently were classified in more advanced stages, mainly as a result of more frequent lymph node metastases, compared with CD44v6 positive tumors (Table 2). We believe that the high metastatic potential of tumors with a lack of CD44v6 expression is likely to cause the short survival period of the patients. One difference in the clinical behavior of EHBD carcinoma and ampullary carcinoma was seen as reported previously.29 In that report, the survival rate of patients with ampullary carcinoma was higher than for those with EHBD carcinoma. To understand the course of progression of these tumors, it may be useful to compare the CD44v6 positive and negative groups at the same stage in the same disease. Figure 3C shows that lack of CD44v6 immunoreactivity may be a predictive factor of poor prognosis for patients with early (pT1 or pT2) EHBD carcinoma. The poor prognosis of patients with ampullary carcinoma also correlated with the lack of CD44v6 expression (P , 0.05). Because all patients with pT1 or pT2 ampullary carcinoma have survived regardless of CD44v6 immunoreactivity, the lack of CD44v6 expression is considered to be a predictive factor of poor prognosis for patients with advanced (pT3 or pT4) ampullary carcinoma. The fact that 6 of 15 deceased patients died of peritoneal carcinomatoses indicates that recurrence easily may occur from cells that detach from primary lesions that lack CD44 or lymph node metastases. We speculate that a lack of CD44v6 leads tumor cells to break out of the normal cell alignment, which causes a worse differentiation grade and easier detachment from the primary lesion and causes tumor cells to invade regional lymph nodes. Even if macroscopic regional lymph adenectomy is performed completely by surgery, tumor cells sometimes remain, and some residual tumor cells disseminate to the peritoneal space and cause peritoneal carcinomatoses, so that the prognosis is poor for these patients. We conclude that a lack of CD44v6 expression in EHBD/ampullary carcinomas is correlated with lymph node metastases and poor patient prognosis. Thus, CD44v6 expression in tumor cells may be a useful prognostic indicator. More aggressive therapy may be necessary for treating patients with carcinoma that lacks CD44v6 expression. REFERENCES 1. Shino Y, Watanabe A, Yamada Y, Tanase M, Yamada T, Matsuda M, et al. Clinicopathologic evaluation of immunohistochemical E-cadherin expression in human gastric carcinomas. Cancer 1995;76:2193–201. 2. Dong JT, Suzuki H, Pin SS, Bova GS, Schalken JA, Isaacs WB, et al. Down-regulation of the KAI1 metastasis suppressor gene during the progression of human prostatic cancer infrequently involves gene mutation or allelic loss. Cancer Res 1996;56:4387–90. 3. Yao M, Zhou XD, Zha XL, Shi DR, Fu J, He JY, et al. Expression of the integrin alpha 5 subunit and its mediated cell adhesion in hepatocellular carcinoma. J Cancer Res Clin Oncol 1997;123:435– 40. 4. Screaton GR, Bell MV, Jackson DG, Cornelis FB, Gerth U, Bell JI. Genomic structure of DNA encoding the lymphocyte homing receptor CD44 reveals at least 12 alternatively spliced exons. Proc Natl Acad Sci USA 1992;89:12160 – 4. 5. Tempfer C, Losch A, Heinzl H, Hausler G, Hanzal E, Kolbl H, et al. Prognostic value of immunohistochemically detected CD44 isoforms CD44v5, CD44v6 and CD44v7-8 in human breast cancer. Eur J Cancer 1996;32a:2023–5. 6. Nishimura S, Chung YS, Yashiro M, Inoue T, Sowa M. CD44H plays an important role in peritoneal dissemination of scirrhous gastric cancer cells. Jpn J Cancer Res 1996;87: 1235– 44. 7. Stauder R, Van Driel M, Schwarzler C, Thaler J, Lokhorst HM, Kreuser ED, et al. Different CD44 splicing patterns define prognostic subgroups in multiple myeloma. Blood 1996;88:3101– 8. 8. Nihei Z, Ichikawa W, Kojima K, Togo S, Miyanaga T, Hirayama R, et al. The positive relationship between the expression of CD44 variant 6 and prognosis in colorectal cancer. Surg Today 1996;26:760 –1. 9. Tanabe KK, Stamenkovic I, Cutler M, Takahashi K. Restoration of CD44H expression in colon carcinomas reduces tumorigenicity. Ann Surg 1995;222:493–503. 10. Kan M, Aki M, Akiyama K, Naruo S, Kanayama H, Kagawa S. High-level expression of the CD44 variant sharing exon v10 in renal cancer. Jpn J Cancer Res 1995;86:847–53. CD44v6 in Bile Duct-Ampullary Carcinoma/Yokoyama et al. 11. Tempfer C, Gitsch G, Haeusler G, Reinthaller A, Koelbl H, Kainz C. Prognostic value of immunohistochemically detected CD44 expression in patients with carcinoma of the vulva. Cancer 1996;78:273–7. 12. Sugino T, Gorham H, Yoshida K, Bolodeoku J, Nargund V, Cranston D, et al. Progressive loss of CD44 gene expression in invasive bladder cancer. Am J Pathol 1996;149:873– 82. 13. Kallakury BV, Yang F, Figge J, Smith KE, Kausik SJ, Tacy NJ, et al. Decreased levels of CD44 protein and mRNA in prostate carcinoma. Correlation with tumor grade and ploidy. Cancer 1996;78:1461–9. 14. Cannistra SA, Kansas GS, Niloff J, DeFranzo B, Kim Y, Ottensmeier C. Binding of ovarian cancer cells to peritoneal mesothelium in vitro is partly mediated by CD44H. Cancer Res 1993;53:3830 – 8. 15. Gross N, Beck D, Beretta C, Jackson D, Perruisseau G. CD44 expression and modulation on human neuroblastoma tumours and cell lines. Eur J Cancer 1995;31a:471–5. 16. Combaret V, Gross N, Lasset C, Frappaz D, Peruisseau G, Philip T, et al. Clinical relevance of CD44 cell-surface expression and N-myc gene amplification in a multicentric analysis of 121 pediatric neuroblastomas. J Clin Oncol 1996; 14:25–34. 17. Fujita N, Yaegashi N, Ide Y, Sato S, Nakamura M, Ishiwata I, et al. Expression of CD44 in normal human versus tumor endometrial tissues: possible implication of reduced expression of CD44 in lymph-vascular space involvement of cancer cells. Cancer Res 1994;54:3922– 8. 18. Kunishi M, Kayada Y, Yoshiga K. Down-regulated expression of CD44 variant 6 in oral squamous cell carcinomas and its relationship to regional lymph node metastasis. Int J Oral Maxillofac Surg 1997;26:280 –3. 19. Hermanek P, Sobin L. International Union Against Cancer TNM classification of malignant tumours. 5th ed. Berlin: Springer-Verlag, 1997. 1699 20. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 1987;162:156 –9. 21. Stamenkovic I, Amiot M, Pesando JM, Seed B. A lymphocyte molecule implicated in lymph node homing is a member of the cartilage link protein family. Cell 1989;56:1057– 62. 22. Hofmann M, Rudy W, Zoller M, Tolg C, Ponta H, Herrlich P, et al. CD44 splice variants confer metastatic behavior in rats: homologous sequences are expressed in human tumor cell lines. Cancer Res 1991;51:5292–7. 23. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457– 81. 24. Kurosaki I, Tsukada K, Hatakeyama K, Muto T. The mode of lymphatic spread in carcinoma of the bile duct. Am J Surg 1996;172:239 – 43. 25. Kayahara M, Nagakawa T, Ohta T, Kitagawa H, Miyazaki I. Surgical strategy for carcinoma of the papilla of Vater on the basis of lymphatic spread and mode of recurrence. Surgery 1997;121:611–7. 26. Pichlmayr R, Weimann A, Klempnauer J, Oldhafer KJ, Maschek H, Tusch G, et al. Surgical treatment in proximal bile duct cancer. A single-center experience. Ann Surg 1996;224: 628 –38. 27. Talamini MA, Moesinger RC, Pitt HA, Sohn TA, Hruban RH, Lillemoe KD, et al. Adenocarcinoma of the ampulla of Vater. A 28-year experience. Ann Surg 1997;225:590 – 600. 28. Ashida K, Terada T, Kitamura Y, Kaibara N. Expression of E-cadherin, alpha-catenin, beta-catenin, and CD44 (standard and variant isoforms) in human cholangiocarcinoma: an immunohistochemical study. Hepatology 1998;27:974 – 82. 29. Yeo CJ, Cameron JL, Sohn TA, Lillemoe KD, Pitt HA, Talamini MA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s. Ann Surg 1997;226:248 – 60.