The incidence and mortality of cholangiocarcinoma are increasing despite improvements in the diagnostic method. Since the sensitivity of brushing cytology for cholangiocarcinoma is not satisfactory, a novel diagnostic marker needs to be established. A recent report has suggested upregulation of the calcium-binding protein S100P in cholangiocarcinoma. The expression status of S100P in normal bile duct and cholangiocarcinoma tissues was assessed by immunohistochemistry. The expression levels of S100P mRNA in the brushing cytology samples during endoscopic retrograde cholangiopancreatography (ERCP) from benign biliary strictures and cholangiocarcinoma were assessed by real-time reverse transcription–polymerase chain reaction (RT-PCR). The sensitivity and specificity of each diagnostic strategy was compared. S100P was frequently expressed in the cholangiocarcinoma tissues, but not in the normal bile duct. The brushing cytology samples from the cholangiocarcinoma cases revealed higher expression levels of S100P compared with the benign biliary strictures. The relative expression level of S100P could determine the cholangiocarcinoma at higher sensitivity than classical cytology, and the combination of the S100P expression level and cytology yielded a sensitivity of 90.0%, with a specificity of 92.0%. Calcium-binding protein S100P is a novel marker of cholangiocarcinoma. Detecting the S100P expression levels in brushing cytology samples has a diagnostic value, which will be helpful for better diagnosis of cholangiocarcinoma. (Cancer Sci 2011; 102: 150–156)
Cholangiocarcinoma has a high mortality rate due to the difficulty of early diagnosis and limitations in therapeutic options. The incidence of cholangiocarcinoma, especially intrahepatic cholangiocarcinoma, is increasing.(1) Despite the improvement in diagnostic methods, mortality from cholangiocarcinoma is still increasing.(2) The 5-year survival of patients with cholangiocarcinoma ranges 20–43%, indicating the poor prognosis of this malignancy.(3) Surgical resection is the only curative therapy, but most patients present at an advanced clinical stage.
One of the clinical problems in the diagnosis of cholangiocarcinoma is the difficulty in obtaining histological evidence of malignancy. According to a previous report, the sensitivity of diagnostic methods such as brushing cytology during endoscopic retrograde cholangiopancreatography (ERCP) is around 60%.(4,5) Conventional diagnostic methods such as computed tomography, magnetic resonance imaging, positron emission tomography and endoscopic ultrasound are performed,(6) but accurate diagnosis of cholangiocarcinoma using a single method is still challenging, with the result that these diagnostic methods have become complementary investigations.(7) Recently, endoscopic ultrasonography-guided fine-needle aspiration (EUS-FNA) biopsy has been applied for the diagnosis of brushing cytology-negative cholangiocarcinoma,(8–10) but the potential risk for a tumor seeding still remains.(3) Based on these clinical circumstances, a novel diagnostic marker for cholangiocarcinoma that can be obtained by a less invasive procedure needs to be established.
Recent advances in the biochemical field have enabled the comprehensive analysis of gene expression profiles in cancer cells from various organs. Obama et al.(11) conducted a comprehensive genome-wide analysis in human cholangiocarcinoma, and the calcium-binding protein S100P was identified as an upregulated gene. S100P is a calcium-binding protein that belongs to the S100 protein family, which consists of at least 13 members.(12) S100P is reported to stimulate cell proliferation and cell survival in NIH3T3 cells via the receptor for advanced glycation end products (RAGE) signaling pathway.(13) Furthermore, S100P is reported to be highly expressed in neoplastic lesions of the pancreas,(14) and its expression contributes to tumor progression.(15) A previous report has suggested the diagnostic value of S100P mRNA in pancreatic juice for the early detection of pancreatic cancer.(14)
In the present study, we assessed expression of S100P in surgically resected cholangiocarcinoma specimens, and found frequent expression of S100P in cholangiocarcinoma tissues. We further assessed the expression levels of S100P mRNA in microdissected cholangiocarcinoma and normal bile duct epithelium, and confirmed higher expression of S100P mRNA in the cholangiocarcinoma epithelium than in the normal bile duct. Comparison of the S100P mRNA expression levels in the brushing cytology samples from cholangiocarcinoma and benign biliary strictures revealed improved sensitivity compared with classical cytology. These findings will contribute to developing a better diagnostic method for cholangiocarcinoma.
Materials and Methods
Tissues. Cholangiocarcinoma tissues were obtained from patients who underwent surgical resections for the tumors from January 2000 to December 2007 in our institute. The tissues collected at the time of surgery were fixed in 10% paraformaldehyde overnight and embedded in paraffin wax for the immunohistochemistry. Forty one cholangiocarcinoma tissues were used for the immunohistochemistry. Written informed consent was obtained from every patient before surgery. The forceps biopsy samples were collected from patients who underwent diagnostic ERCP for biliary stricture from April 2007 to March 2010. We performed ERCP in 60 cholangiocarcinoma patients and 25 benign biliary stricture patients for the diagnosis. The forceps biopsy of the biliary stricture was performed in 50 patients (41 cholangiocarcinoma patients and nine benign biliary stricture patients). These biopsy samples were also used for the immunohistochemistry. Written informed consent was obtained from every patient before ERCP.
Immunohistochemistry. Expression of the S100P protein in human cholangiocarcinoma tissues was investigated by immunohistochemistry. The diagnosis of cholangiocarcinoma was pathologically confirmed in each case. The tissue sections were deparaffinized and antigens were retrieved by boiling the sections in Target Retrieval Solution (Dako, Carpinteria, CA, USA) in a microwave oven. The sections were then incubated in methanol with 0.3% hydrogen peroxide for 30 min in order to block the endogenous peroxidase activity. Thereafter, the histofine kit (Nichirei, Tokyo, Japan) for S100P (610306, BD Transduction Laboratories, San Jose, CA, USA) was used. Visualization of the immunoreaction was carried out in 0.06 mM 3,3′-diaminobenzidine tetrahydrochloride (Wako, Osaka, Japan) containing 2 mM hydrogen peroxide in phosphate-buffered saline (PBS) for several minutes at room temperature. For the negative control, the immunostaining processes were performed by replacing the primary antibody with PBS. The negative control sections showed no specific immunoreactivity. The degree of immunostaining was evaluated as follows: negative, <30% positive cells were found; positive, more than 30% positive cells were found.
Microdissection of cholangiocarcinoma tissue and RNA extrac-tion. Cholangiocarcinoma tissues were obtained from patients who underwent surgical operations for tumors from January 2008 until March 2010. The tissues collected at the time of surgery were immediately embedded in Tissue-Tek O.C.T. compound medium (Sakura, Tokyo, Japan), frozen in liquid nitrogen and stored at −80°C. The frozen tissues were cut into 8-μm thick sections using cryostat (Jung CM3000; Leica, Nussloch, Germany), and six cholangiocarcinoma tissues or five normal tissues were subjected to LASER-captured microdissection using a Leica CIR MIC system (Leica microsystems, Wetzkar, Germany). Sample collection and usage were performed under written informed consent obtained from each patient before surgery. Total cellular RNA was extracted from each sample using the mirVan miRNA Isolation Kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s protocol.
Brushing cytology samples and RNA extraction. The brushing cytology samples were collected from patients who underwent diagnostic ERCP for the biliary stricture from April 2007 to March 2010. We performed ERCP in 60 cholangiocarcinoma patients and 25 benign biliary stricture patients for the diagnosis. The brushing cytology of the biliary stricture was performed in every patient using Rapid Exchange Cytology Brushes (Boston Scientific Corporation, Tokyo, Japan), and part of the brushing cytology samples were subjected to RNA extraction for the S100P mRNA quantification. Total cellular RNA was extracted from each sample using the mirVana miRNA Isolation Kit (Applied Biosystems) according to the manufacturer’s protocol. All diagnoses of cholangiocarcinomas were confirmed finally by brushing cytology, endoscopic forceps biopsy or histological examination after surgical resection (only in operable cases). Cytology results of class IV or V, or histological results demonstrating cholangiocarcinoma were considered to be positive. Cases with negative cytology and histology in the preoperative examinations were subjected to surgery based on the existence of a tumor confirmed by other modalities of examination, such as EUS or computed tomography. The diagnosis of a benign biliary stricture was made by negative cytology and histology results, or other modalities that showed findings concordant with non-malignant disease and non-progressive disease for at least 3 months. Sample collection and usage were done under written informed consent prior to the ERCP.
Real-time RT-PCR. Real-time RT-PCR analyses were performed using a QuantiTect SYBR Green RT-PCR Kit (QIAGEN, Valencia, CA, USA) according to the manufacturer’s protocol. The primer sequences for Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and S100P were described previously.(14,16) The annealing temperatures for these primer sets were 60°C. The specificity of each PCR reaction was confirmed by melting curve analysis.(17) The copy number of S100P or GAPDH in each 4 ng of total RNA sample was calculated using LightCycler software (Roche diagnostics, Basel, Switzerland), and the relative S100P expression in each sample was calculated.
Statistical analysis. The correlation between the S100P immunohistochemistry results and the clinical features was assessed by the Chi-squared test. The correlation between the clinical features and final diagnoses was assessed by the Chi-squared test. The difference between two groups was analyzed by Mann–Whitney U test. A P value of <0.05 was regarded as statistically significant. The median relative expression level of S100P mRNA is displayed with the range.
Ethical committee approval. The experimental design was approved by the ethics committee of Tohoku University Hospital, article no. 2008-506.
Detection of calcium-binding protein S100P in the cholangio-carcinoma tissues. We examined calcium-binding protein S100P expression in cholangiocarcinoma tissues by immunohistochemistry using a specific antibody and examined the association of its expression with the clinicopathological features. Expression of S100P was found in 31 of 41 (75.6%) cholangiocarcinoma tissues, while no staining was seen in normal bile duct epithelium (Fig. 1). There was no association between S100P expression and the patients’ age, gender, stage, histological classification and lymph node metastasis. The results of the immunohistochemistry are summarized in Table 1.
Table 1. Correlation between the clinicopathological findings and S100P expression
*Analyzed by χ2 test. P < 0.05 was considered to be statistically significant.
Lymph node metastasis
To confirm that S100P is expressed in cholangiocarcinoma cells, we examined the expression level of S100P mRNA in the microdissected cholangiocarcinoma or normal bile duct epithelium. As shown in Figure 2A, cholangiocarcinoma or normal bile duct epithelium was collected using LASER-captured microdissection. The relative expression level of S100P in each sample is summarized in Figure 2B, and cholangiocarcinoma revealed significantly higher expression levels of S100P than normal bile duct epithelium.
The S100P mRNA expression level in the brushing cytology sample from cholangiocarcinoma was higher than that in the benign biliary stricture. Since the cholangiocarcinoma cells were considered to be a source of S100P, we next tried to detect S100P mRNA in the brushing cytology samples. The clinical features of patients who received the brushing cytology during the ERCP are summarized in Table 2. Patients with cholangiocarcinoma tended to have intrahepatic lesions. This result might be explained by the fact that benign biliary stricture patients contained chronic pancreatitis and autoimmune pancreatitis patients. The diagnoses of patients with benign biliary strictures are summarized in Table 3. As shown in Figure 3A, the brushing cytology samples from the cholangiocarcinoma revealed significantly higher expression levels of S100P mRNA than those from benign biliary strictures.
Table 2. Differences in the clinicopathological features between cholangiocarcinomas and benign bile duct strictures
*Analyzed by χ2 test. P < 0.05 was considered to be statistically significant.
Location of the stricture
Table 3. The diagnoses of 25 benign biliary strictures
Primary sclerosing cholangitis
Measurement of the S100P mRNA expression level in the brushing cytology sample improved the sensitivity of the cholangiocarcinoma diagnosis. Based on these findings, we tried to compare the levels of S100P mRNA expression in the brushing cytology sample with the diagnosis of cholangiocarcinoma. By the receiver operating characteristic (ROC) curve analysis, we determined the cut-off value of the relative S100P mRNA expression level for cancer diagnosis as 0.041 (Fig. 3B). On this setting, the diagnosis of cholangiocarcinoma by the S100P mRNA measurement in the brushing cytology sample reached a sensitivity of 83.3%, with a specificity of 92.0% (Table 4). The classical cytology showed a sensitivity of 56.7%, with a specificity of 100%, which was a comparable result with previous reports.(4,5) Furthermore, by regarding the brushing cytology sample with a positive cytology or more than 0.041 of the relative S100P mRNA expression as cholangiocarcinoma, the sensitivity reached 90.0%.
Table 4. The sensitivity, specificity and accuracy of each diagnostic strategy
S100P + cytology
Immunohistochemistry of S100P in the forceps biopsy sample improves the sensitivity of the cholangiocarcinoma diagnosis. Because the sensitivity of brushing cytology is reported to be rather low compared with forceps biopsy,(4,18) we assessed the sensitivity and specificity of the conventional forceps biopsy from the biliary stricture for the diagnosis of cholangiocarcinoma. The forceps biopsy showed a sensitivity of 46.3%, with a specificity of 100%. The immunohistochemistry of S100P in these samples showed a sensitivity of 58.5%, with a specificity of 100%. Typical immunohistochemistry of S100P in a biopsy sample is shown in Figure 4. By regarding the forceps biopsy sample with positive histology or positive immunohistochemistry of S100P as cholangiocarcinoma, the sensitivity reached 68.3%. These results are summarized in Table 5.
Table 5. The sensitivity, specificity and accuracy of each diagnostic strategy
S100P IHC + histology
Characteristics of cytology-positive/S100P-positive cases and cytology-negative/S100P-positive cases of cholangio carcinoma. A typical cholangiocarcinoma case with a positive cytology result and positive S100P expression level is shown in Figure 5. A 66-year-old man, who was referred to our department for further examination of a dilatation of the intrahepatic bile duct (Fig. 5A), was found to have bile duct stenosis at the right hepatic duct (Fig. 5B). Brushing cytology was performed (Fig. 5C), and the sample contained a sufficient number of cells representing the typical features of adenosquamous carcinoma (Fig. 5D). The relative expression level of S100P was 0.392, above the cut-off value.
The other case of cholangiocarcinoma showed a negative cytology result, but this case also had similar dilatation of the intrahepatic bile duct and bile duct stenosis at the right hepatic duct (Fig. 6A,B). Brushing cytology was also performed in this case (Fig. 6C). Since the brushing cytology sample contained only a small amount of dyskaryotic cells (Fig. 6D), this case was diagnosed as class III. The relative expression level of S100P was 0.177, above the cut-off value. Both cases underwent surgical operations, and the final diagnoses were confirmed by histological diagnosis (Figs 5E,6E) as cholangiocarcinomas.
In the present study, the calcium-binding protein S100P was identified as a novel cholangiocarcinoma marker. Despite improvements in other diagnostic methods such as positron-emission tomography and magnetic resonance imaging, diagnosing cholangiocarcinoma still remains clinically difficult. Anaccurately obtained biopsy sample from the biliary stricture is required for a precise diagnosis, but the endoscopic delivery of biopsy forceps is sometimes problematic. A recent report suggested that the combination of immunohistochemistry of S100P, von Hippel-Lindau gene product and IMP3 are useful markers of cholangiocarcinoma,(19) supporting our findings. The brushing cytology technique for RNA using Rapid Exchange Cytology Brushes (Boston Scientific Corporation) can be performed during ERCP. Previously, immunostaining of p53 or Ki-67 in a forceps biopsy specimen was reported to improve the sensitivity of the diagnosis for cholangiocarcinoma when combined with the histology.(20) The sensitivity improved from 53% up to 75%, but the specificity was reduced. Similarly, detection of telomerase activity in the bile duct biopsy specimens improved the sensitivity of the bile duct cancer diagnosis when combined with p53 overexpression, but in a relatively small sample size (16 cases).(21) Another approach by detecting aspartyl beta-hydroxylase and homeobox B7 mRNA in the brushing cytology sample also improved the sensitivity of brushing cytology up to 82%.(22) Compared with these methods, our method yielded comparable sensitivity and maintained favorable specificity.
Of note, the measurement of the S100P expression level has an advantage in cases whose brushing cytology specimen contains only a small amount of atypical cells, which can lead to inconclusive cytology results. Since the sensitivity of RT-PCR is high, the existence of even a very small number of atypical cells might be detected by this method. In addition, the S100P expression level could be measured within 1 day, which also benefits earlier clinical decision making. Furthermore, obstructive jaundice is a typical clinical manifestation of cholangiocarcinoma,(23) and immediate biliary drainage is required in cases of cholangitis.(24) In these situations, placement of a plastic biliary stent or endoscopic naso-biliary drainage tube is widely performed, but chronic inflammation due to the plastic stent will cause artificial thickening of the bile duct epithelium.(25) Even in these cases, brushing cytology can be performed, and the measurement of S100P expression can also be carried out.
Based on our observation, 10 out of 60 (16.7%) cases of cholangiocarcinoma could not be diagnosed by the S100P expression level, but the diagnoses of four cases were determined by the cytology results. Similarly, 26 out of 60 (43.3%) cases of cholangiocarcinoma could not be diagnosed by cytology, but 20 cases could be determined by the S100P expression level. According to these results, the combined diagnostic strategy would be optimal for clinical application. The sensitivity of forceps biopsy in this study was not high enough to render unnecessary measurement of the S100P expression. In addition, forceps biopsy was not possible in every case of biliary stricture (possible in 50 cases out of a total 85 cases). Although the combination of S100P immunohistochemistry with conventional histology improved the diagnostic sensitivity, the difficulty in obtaining sufficient amounts of tissue by forceps biopsy still remains. In cases in which sufficient amounts of tissue could not be obtained, S100P immunohistochemistry showed as poor sensitivity, the same as conventional cytology.
This study has limitations since the design is a single-center based, retrospective study. Further confirmation of the usefulness of the S100P expression level in brushing cytology specimens is required.
This work was supported in part by Grant-in-aid 19590745 and 19890015 from the Ministry of Education, Science, Sports and Culture in Japan.
The authors declare no financial support or relationships that may pose a conflict of interest.