SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

It is difficult to distinguish pancreatic ductal adenocarcinoma (PDAC) from chronic pancreatitis (CP) when stricture is present in the pancreatic duct. Endoscopic brushing cytology is a convenient method for investigating strictures in the pancreatic duct, however, the diagnostic sensitivity of this method for PDAC is reported to be low (40–70%). Recently, we revealed that MSX2 is frequently expressed in PDAC cells but not in normal cultured pancreatic duct or stellate cells. Thus, we analyzed MSX2 expression levels in brushing samples to examine whether this would differentiate PDAC from CP. Cytologic brushing specimens were obtained from pancreatic duct strictures during endoscopic retrograde cholangiopancreaticography in 82 patients. The brushing fluid was subjected to cytological diagnosis and RNA extraction. The expression level of MSX2 was evaluated by one-step real-time RT-PCR. MSX2 expression levels were significantly higher in PDAC than in CP (P = 0.0000007), and the expression level was associated with positive cytology (P = 0.013). The sensitivity, specificity, and diagnostic accuracy for PDAC of cytology and MSX2 expression in ductal strictures were: 47.4%, 100%, and 63.4%, and 73.7%, 84.0%, and 79.3%, respectively. The sensitivity and accuracy of MSX2 expression levels for diagnosis were much higher than those of cytology. This suggests that the evaluation of MSX2 levels in endoscopic retrograde cholangiopancreaticography brushing samples would be useful for distinguishing PDAC from CP. (Cancer Sci 2011; 102: 157–161)

Pancreatic ductal adenocarcinoma is one of the most malignant gastrointestinal tumors. Once pancreatic cancer is clinically evident, it progresses rapidly to develop metastatic lesions, frequently by the time of diagnosis. Furthermore, these tumors are usually resistant to conventional chemotherapy and radiation therapy. Therefore, accurate methods for detection of early PDAC are required to improve its poor prognosis. The discrimination of early pancreatic carcinoma from CP is sometimes very difficult because both diseases show similar stricture in the pancreatic duct. Thus, differential diagnosis for PDAC and CP is of critical importance to avoid delaying treatment of early pancreatic cancer. Endoscopic ultrasound-guided fine needle aspiration biopsy is an effective technique in the diagnosis and staging PDAC, and accuracy for the diagnosis of malignancy is high with sensitivity ranging from approximately 80% to 90%.(1–3) However, the availability of this procedure is restricted because the necessary equipment is expensive and the technique itself is difficult. Endoscopic pancreatic duct brushing is a convenient method for strictures in the main pancreatic duct or in the second branch. However, the diagnostic sensitivity of this method for PDAC is reported to be low (40–70%)(4–7) compared with that of EUS-FNAB.

A member of the homeobox genes (Hox genes) family, MSX2 is present in a variety of sites, including premigratory cranial neural crest, tooth, retina and lens, apical ectodermal ridge, and mammary gland.(8–11) In the development of these organs, the expression patterns of this gene suggest its active involvement in epithelial–mesenchymal interactions.(12,13) Enhanced levels of transcripts for MSX2 have been shown in a variety of carcinoma cell lines of epithelial origin compared to their corresponding normal tissues.(14) Recently, we have shown that MSX2 was frequently expressed in pancreatic carcinoma cell lines and PDAC tissues but not in benign cultured cells, including PSC, which are largely responsible for fibrogenesis of CP(15) or normal human pancreatic tissues,(16,17) suggesting that detection of MSX2 could be a useful marker for differentiating PDAC from CP. Therefore, we investigated the mRNA expression level of MSX2 in 82 endoscopic brushing samples from stricture of the pancreatic duct to examine whether MSX2 expression would distinguish malignant from benign pancreatic disease and could improve the diagnostic yield of brush cytology.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Pancreatic tissue samples and microdissection.  The pancreatic tissues collected at the time of surgery were embedded immediately in Tissue-Tek OCT 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 a cryostat (Jung CM3000; Leica Biosystems, Nussloch, Germany) then fixed with cold methanol and stained with hematoxylin. Histologically cancerous ductal cells and non-cancerous cells (mixed with inflammatory duct cells and stromal cells) from the sections were dissected using a Leica CIR MIC system (Leica Microsystems, Wetzlar, Germany) according to the manufacturer’s protocols. These microdissected samples (eight cancerous and nine non-cancerous regions) were subjected to RNA extraction. RNA was prepared using a mirVana miRNA isolation kit that can isolate total cellular RNA (Applied Biosystems, Foster City, CA, USA).

Patients.  Brushing cytology was attempted in 95 patients who were suspected of pancreatic carcinoma due to strictures of the pancreatic duct during ERCP at Tohoku University Hospital (Sendai, Japan) between April 2007 and March 2010. Prior to ERCP, these patients were examined by various imaging methods, such as abdominal ultrasonography, EUS, computed tomography, MRI, or PET, and at least one of these modalities indicated suspicious lesion of tumor in the pancreas. Brushing specimens were obtained from pancreatic strictures during ERCP from 82 patients. Characteristics of all patients are summarized in Table 1. Final diagnoses were PDAC (n = 57) and CP (n = 25). Final histological confirmation for PDAC was done by EUS-FNA (n = 21), biopsy from invaded lesion (n = 1), and operation (n = 35). Informed consent was obtained from all patients, and the study was approved by the ethics committee at Tohoku University.

Table 1.   Clinical characteristics of patients with pancreatic ductal adenocaricnoma (PDAC) or chronic pancreatitis (CP) who received brushing cytology
 PDACCPP-value*
  1. *P-values obtained using the χ2-test. †Location in the pancreas where brushing was carried out.

Age
 <60 years14120.0360
 ≥60 years4313 
Gender
 Male25200.0025
 Female325 
Location†
 Head39200.5000
 Body164 
 Tail21 
Total5725 

Sampling and RNA extraction.  The samples were collected only when ductal stricture was found during ERCP using Rapid Exchange cytology brushes (Boston Scientific, Natick, MA, USA) with 0.025–0.035-inch guide wire. The stricture was brushed with several to-and-fro motions and the brush was placed in a vial of saline solution and RNA extraction buffer (Qiagen, Hilden, Germany) for cytologic diagnosis and RNA extraction, respectively.

The cells in saline solution were centrifuged and the precipitate was stained by a standard Papanicolaou technique. The evaluations of cytological examination were classified according to the definition of Goodale et al.(18)

Total RNA was extracted using a mirVana miRNA isolation kit that can isolate total cellular RNA (Applied Biosystems). One-step quantitative real-time RT-PCR was carried out on each sample by adding 4 ng total RNA with a QuantiTect SYBR Green RT-PCR kit (Qiagen) using LightCycler (Roche Diagnostics, Basel, Switzerland). RNA concentration was determined with an ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The primer pairs used were: MSX2 forward, 5′CCGCCTCGGTCAAGTCGGAAAAT3′ and reverse, 5′TGGAGAGGTACTGTTTCTGACGG3′; and GAPDH forward, 5′GGCGTCTTCACCACCATGGAG3′ and reverse 5′AAGTTGTCATGGATGACCTTGGC3′. All reactions were carried out according to the manufacturer’s protocol. Plasmids including the coding region of MSX2 or GAPDH were used as standards and the copy number of MSX2 in each sample was normalized to the respective GAPDH copy number. The specificity of each PCR reaction was confirmed by melting curve analyses.

Statistical analysis.  Differences in clinical parameters between PDAC and CP patients were analysed by the χ2-test. Correlation between results of cytology or MSX2 expression levels and location of stricture were validated by the χ2-test. The association of MSX2 expression levels and location of stricture was evaluated by the χ2-test. Differences in mRNA expression of MSX2 between PDAC and CP were analysed by the Mann–Whitney U-test. P-values <0.05 were considered statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Detection of MSX2 mRNA expression in pancreatic cancer tissues.  The MSX2 mRNA expression levels in human pancreatic tissues were analyzed to assess whether this gene expression can discriminate PDAC from non-cancerous lesions. As shown in Figure 1(A), cancerous duct cells were carefully dissected and MSX2 mRNA expression was compared to non-cancerous regions using the real-time RT-PCR technique. MSX2 was significantly expressed in cancer cells compared to non-cancer cells (= 0.0052; Fig. 1B), indicating the validation of this gene expression level would be useful to distinguish PDAC cells from non-cancer cells.

image

Figure 1.  Expression of MSX2 mRNA in microdissected lesions was detected by quantitative real-time RT-PCR. RNA extraction was carried out from the microdissected lesions. Cancer lesions stained with hematoxylin (left in A) were selectively dissected (right in B). Seventeen lesions were microdissected and RNAs were extracted. RNAs were subjected to one-step quantitative real-time RT-PCR and the target gene was normalized to GAPDH expression. MSX2 mRNA expression was significantly higher in cancerous lesions than non-cancerous lesions (P = 0.0052, Mann–Whitney U-test). MSX2/GAPDH represents MSX2 copy number/GAPDH copy number/μL.

Download figure to PowerPoint

MSX2 mRNA expression levels significantly higher in brushing specimens from PDAC compared to those from CP.  In 13 of 95 patients (13.7%), cytological brushing could not be done because the guide wire was not able to pass the ductal stricture. These 13 patients were finally diagnosed as PDAC by EUS-FNA or surgery. The clinicopathological characteristics of patients who received the cytology during ERCP are summarized in Table 1. Patients with PDAC were older than those with CP, and there were more male patients with CP than PDAC. In the remaining 82 patients, endoscopic brushing was successfully carried out and satisfactory specimens were obtained from all cases. No major complications, including acute pancreatitis, occurred after brushing cytology in the current study. The sensitivity of routine brush cytology for PDAC was 47.4% (27/57) with 100% specificity and 63.4% diagnostic accuracy. No significant association of tumor location with positive rate of cytology in the tumor location was found (Table 2).

Table 2.   Association of tumor site with cytological examination or MSX2 expression level in pancreatic cancer
 LocationP-value*
HeadBodyTail
  1. *P-values obtained using the χ2-test.

Cytology (−)20820.39
Cytology (+)1980 
MSX2 (−)8520.039
MSX2 (+)31110 

MSX2 mRNAs in brushing samples were successfully detected and quantified by normalization to the respective GAPDH copy number (Fig. 2A). The mean expression level of MSX2 mRNA was 0.012 ± 0.0024 (MSX2/GAPDH copy number/μL, mean ± standard error) and 0.0026 ± 0.0004 in PDAC and CP, respectively (Fig. 2B). The expression levels of MSX2 mRNA was significantly higher in PDAC samples compared to those of CP (= 0.0000007).

image

Figure 2.  Expression of MSX2 mRNA in brush samples from patients with pancreatic ductal adenocarcinoma (PDAC) and chronic pancreatitis (CP). (A) Representative real-time RT-PCR curve showed significant difference in MSX2 expression levels between PDAC and CP. (B) Expression of MSX2 was significantly higher in specimens from PDAC than in those from CP (P < 0.001, Mann–Whitney U-test).

Download figure to PowerPoint

MSX2 mRNA expression level useful to distinguish PDAC from CP.  In order to apply this expression value for diagnosis of pancreatic carcinoma, we then analyzed the receiver operating characteristic curve and defined the cut-off value as 0.00413 (Fig. 3). The MSX2 expression level was judged positive when it was equal to or higher than the cut-off value. Using this cut-off value, sensitivity, specificity for malignancy, and accuracy for diagnosis was 73.7%, 84.0%, and 79.3%, respectively. The positive rate of the MSX2 expression level was significantly high in head (79.5%, 31/39) and body (68.8%, 11/16) compared to tail (0%, 0/2) (P = 0.039; Table 2). Significant correlation was observed between positive cytology and MSX2 expression level (P = 0.013; Table 3). In 27 cases with positive cytology, 24 cases (88.9%) had a higher MSX2 expression level than the cut-off value. In contrast, four of 25 cases of CP (16%) showed positive MSX2 expression. In these four cases with MSX2-positive CP, two cases clearly showed definitely higher MSX2 expression levels, but the other two cases were only slightly higher than the cut-off value, as shown in Figure 2(B).

image

Figure 3.  Receiver operating characteristic curve analysis for the cut-off point of the MSX2 mRNA expression level in a brushing cytology sample. The area under the curve is 0.846. Arrow indicates the cut-off value of 0.00413, which gives the highest sensitivity and specificity.

Download figure to PowerPoint

Table 3.   Correlation between cytological examination and MSX2 expression in pancreatic cancer
 CytologyP-value*
NegativePositive
  1. *P-values obtained using the χ2-test.

MSX2 negative1230.013
MSX2 positive1824 

There were 18 (60%) cases showing positive MSX2 expression among 30 cases of negative cytology. Figure 4 shows a representative case with negative cytology but MSX2 positive expression. This patient was admitted to our hospital for evaluation of MPD dilation. Computer tomography depicted the dilation of MPD but could not identify tumor (Fig. 4A). However, EUS showed the mass as a suspicious hypoechoic area (Fig. 4B). Then ERCP was done and duct brusing was carried in the stricture (Fig. 4C). The result of cytological examination was suspicious for malignancy but could not definitively diagnose (Fig. 4D). Pancreaticoduodenectomy was carried out and a tumor with diameter 15 mm was detected in the pancreatic head (Fig. 4E). Final diagnosis by histological examination was pancreatic adenocarcinoma (Fig. 4F). In this case, the MSX2 expression level (0.00587) was higher than the cut-off value, suggesting that the MSX2 level could detect carcinoma cells in which a confident cytological diagnosis could not be obtained.

image

Figure 4.  Representative case showing negative cytology but positive MSX2 expression in brushing cells. The patient admitted to our hospital for examination due to dilation of the main pancreatic duct (MPD). (A) Computed tomography showed diffuse dilation of MPD but it could not detect a mass lesion. (B) Endoscopic ultrasonography depicted a mass-like region (arrows). (C) Endoscopic retrograde cholangiopancreaticography was done and brushing cytology was carried out from the MPD stricture (arrow head). (D) The brushing specimens showed atypical cells but could not be diagnosed as cancer cells. (E) Macroscopic finding of resected pancreatic head. Tumor with 15 mm diameter was found in the pancreatic head. Scale line = 10 mm. (F) Microphotograph of resected tissue (stained with H&E) indicated well to moderately differentiated adenocarcinoma (original magnification, ×200).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

The differential diagnosis between PDAC and CP is generally difficult because both appear to be solid masses, like lesions with fibrotic changes, with stricture in the pancreatic duct. Although EUS-FNAB is an effective technique in the diagnosis of pancreatic mass, lesions presenting as strictures without a demonstrable mass should be investigated by other methods such as cytologic examination of pancreatic juice or pancreatic duct brushing. In addition, expensive equipment and difficulties with the technique restrict its availability. However, brushing cytology is a convenient method for pancreatic ductal stricture,(7) although its sensitivity is lower than that of EUS-FNAB.(1–7) To improve the sensitivity of brushing cytology for PDAC, we investigated MSX2 expression in brushing samples. In our previous study, MSX2 was expressed in PDAC cells but not in normal pancreatic duct cells or normal cultured cells, including PSC,(16,17) suggesting that the detection of MSX2 expression would be useful to discriminate PDAC from CP. As expected, we clearly revealed in the current study that: (i) expression of MSX2 mRNA was significantly higher in microdissected PDAC cells than non-cancerous cells; (ii) significantly high expression levels of MSX2 mRNA were detected in brushing specimens from PDAC compared to those from CP; and (iii) diagnostic sensitivity and accuracy by analyzing MSX2 expression levels were much higher than those by cytological examination.

Many genetic alterations have been shown in PDAC, the most common genetic alterations of PDAC include oncogene (K-ras), and tumor suppressor gene mutations (p16, p53, SMAD4).(19) Of these, K-ras mutation occurs at very high frequency.(20) However, K-ras muations are detected in CP as well as PDACs,(21) therefore, it is difficult to use the K-ras mutation as a tool to differentiate PDAC from CP. For example, K-ras mutation was analyzed in brush samples with 34 cases of PDAC and 11 with CP and was frequently detected in both cancer (87%) and pancreatitis (40%).(4) However, Parsi et al. measured DNA concentrations of methylated cyclin D2, NPTX2, and TFPI2 promoter in biliary or pancreatic duct brush samples to improve the diagnostic accuracy of brushing cytology. They revealed that nine of 66 (13.6%) disease controls, as well as 30 of 41 (73.2%) patients with PDAC, had methylation of one or more genes in their endoscopic brush samples.(22) This sensitivity or false positive rate for diagnosis of PDAC is similar to our current results. In addition, the diagnostic sensitivity or accuracy by our evaluation of MSX2 expression compares favorably to other markers such as telomerase(23) or MUC1(24) that have been reported to differentiate benign from malignant pancreatic diseases. In this context, the measurement of MSX2 expression levels in the brushing samples would be a good tool to help distinguish PDAC from benign pancreatic duct, in addition to other molecular analyses as above.

MSX2 is suggested to be a downstream target for the ras signaling pathway(25) and is expressed in carcinoma of epithelial origin,(14) suggesting the association of this gene with the development of pancreatic carcinoma, as pancreatic carcinoma harbors frequent mutation of the K-ras gene.(19,20) Accordingly, we have shown that PDAC tissues frequently expressed MSX2 protein and this correlated with less differentiation of carcinoma cells and venous infiltration. Also, forced expression of MSX2 in pancreatic carcinoma cells enhanced the aggressive phenotype through epithelial to mesenchymal transition.(26) This evidence suggested that MSX2 is particularly expressed in invading carcinoma cells. Thus, a high level of MSX2 expression was likely to be frequently detected in ductal strictures that were caused by invading PDAC cells.

In the current study, only four of 25 cases (16%) of CP showed positive MSX2 expression. Chronic pancreatitis is generally known as a risk factor for PDAC and contains premalignant ductal changes, such as PanIN, that show K-ras mutation or other molecular alterations.(27) Thus, molecular markers might tend to show relatively high false positive rates in CP.(4,22) This is likely to be favorable to predict the occurrence of PDAC in CP, but not to avoid unnecessary procedures. In contrast, MSX2 mRNA expression was shown to be below the detection level in premalignant lesions of the pancreas such as adenoma or hyperplastic lesions of intraductal papillary mucinous neoplasm.(17) This supports the fact that MSX2 analysis resulted in low false positive rates in CP. Therefore, repeated evaluation of MSX2 levels in CP patients might help in the early detection for occurrence of PDAC with high specificity, as MSX2 is only expressed in carcinoma cells. Alternatively, CP patients with positive MSX2 expression might have small malignant lesions that could not be detected by standard imaging. Thus, in such cases, careful follow-up by imaging examination would be necessary.

Cytology is the gold standard to diagnose PDAC preoperatively. However, inadequate sampling or a low rate of neoplastic cells in the specimens would prevent accurate diagnosis.(7) Thus, additional methods such as molecular marker analysis might help to encourage efforts to repeat sampling to obtain a cytological diagnosis, or to select EUS-FNAB if a mass is present. As shown in Figure 4, the validation of MSX2 expression levels was useful to diagnose PDAC when cytology alone could not confirm the presence of cancer cells, indicating that MSX2 analysis itself can be a good molecular marker and assist cytological examination. However, this validation could not provide positive results in PDAC located in the tail, and neither could cytological examination, suggesting the technical difficulty in sampling from the pancreatic tail. In addition, 13 of 95 cases (13.7%) failed to gain the brushing specimen for cytological examination and MSX2 analysis because of severe ductal stricture. To increase the sensitivity or diagnostic accuracy for PDAC by MSX2 analysis, further examination for ductal stricture in the tail and improvement of equipment for endoscopic brushing would be required.

There were three PDAC cases that showed positive cytology but negative MSX2 expression. The reason why the presence of carcinoma cells in the brushing specimens did not provide positive MSX2 expression levels is not clear. However, the lack of carcinoma cells in the RNA extraction buffer might partly explain this discrepancy, because all of the carcinoma cells in the brushing specimens were placed in saline solution for cytology. Alternatively, these PDAC cells might express very low levels of MSX2 mRNA. The combination of MSX2 analysis with cytology or clinical diagnosis by imaging examination would be important to avoid false negative diagnoses.

In conclusion, MSX2 was significantly expressed in brush samples of PDAC compared to those of CP. The sensitivity for malignancy and accuracy for diagnosis were much higher than cytology, suggesting that the evaluation of MSX2 expression levels is a good technique for differentiating PDAC from CP when stricture is present in the pancreatic duct.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

This work was supported in part by grants-in-aid (No. 21590870 and No. 20390202) from the Ministry of Education, Science, Sports and Culture, Japan.

Abbreviations
CP

chronic pancreatitis

ERCP

endoscopic retrograde cholangiopancreaticography

EUS

endoscopic ultrasonography

EUS-FNAB

EUS-guided fine needle aspiration biopsy

MPD

main pancreatic duct

PDAC

pancreatic ductal adenocarcinoma

PSC

pancreatic stellate cells

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References