Intraductal papillary-mucinous tumor (IPMT) of the pancreas is a relatively new tumor classification that has been the topic of an increasing number of reports. IPMT is associated with the massive dilatation of the pancreatic duct and its branches by the copious production of mucin. IPMT was first described by Ohhashi et al.1 in 1982 as a “mucous-secreting pancreatic cancer.” This entity has since been reported not only in Japan,2, 3, 4 but also in Europe and the United States5, 6, 7, 8 under a variety of names. The term IPMT has been adopted by the World Health Organization's International Histological Classification of Tumors.9 IPMT can be further classified as a benign adenoma, a moderate dysplasia (borderline) or a carcinoma (noninvasive or invasive), but reliably distinguishing one type from another is difficult to perform preoperatively, and all IPMTs are suspected of having malignant potential.6, 7, 8 Compared to ductal adenocarcinoma (DC) of the pancreas, IPMT clearly has a more favorable natural history characterized by a much longer survival period and a much higher cure rate.5, 6, 7, 8
Inasmuch as various oncogenes are now known to be involved in the pathogenesis of cancer and of pancreatic carcinomas in particular, the lower incidence of malignancy and the reduced aggressiveness of IPMT compared to DC may be the result of a different spectrum of genetic changes. K-ras mutation, a type of mutation that is frequently detected in pancreatic carcinomas,10, 11, 12, 13, 14, 15, 16 may be an important event in the neoplastic process of IPMT.12, 17, 18 K-ras mutations have been recently shown to occur at a relatively early stage of multistep carcinogenesis in pancreas lesions. These mutations have also been used as a clonal marker in myelodysplastic syndromes.19
IPMT, like DC, arises from the epithelial lining of pancreatic ducts. Histologically, however, IPMT lesions differ from DC lesions by exhibiting varying degrees of hyperplasia and neoplasia at different tumor loci.5 Whether the histologic variation observed in IPMT lesions corresponds with a variation in genotype has not been clarified. To investigate this point, DNA must be selectively extracted from the lesions in question and analyzed. Using microdissection, DNA samples are not contaminated with DNA from normal tissue or other histologically different tissues. Few studies have analyzed K-ras mutations in several different IPMT lesions (IPMT has a large number of foci than DC) from the same individual. We obtained specimens using a microdissection method and analyzed the distribution of K-ras mutations in a large number of different IPMT lesions. We then compared our results to those obtained for DC lesions. In addition, we also compared the clinical features of IPMT and DC patients. Our observations provide useful information on the role of K-ras mutations in tumorigenesis.
MATERIAL AND METHODS
Twenty surgically resected IPMT specimens and 7 DC specimens were obtained from 18 men and 9 women, ranging in age from 52–80 years (mean, 65 years). All patients were seen at Keio University Hospital (Tokyo, Japan) between 1988 and 1999. All IPMT lesions were classified according to the World Health Organization's International Histological Classification of Tumors.9 Using the criteria outlined by this classification system, the IPMT lesions were divided into IPMT-adenomas (benign adenoma and borderline moderate dysplasia) and IPMT-carcinomas (malignant). The IPMT-carcinoma group included both in situ and infiltrating carcinomas. Follow-up examinations were scheduled for all patients: 3 clinical stagings were scheduled during the first 2 years after diagnosis, 2 stagings were scheduled between the third and fifth years, and one staging was scheduled at the beginning of the sixth year. Paraclinical investigations consisted of computed tomography and ultrasonography examination and a serologic test for tumor markers (CEA, CA19-9). The mean observation period was 45 months (range, 2–118 months). All patients were followed to death and living status.
The resected specimens were immediately fixed in 10% buffered formalin. After 1 or 2 weeks, the fixed specimens were serially sectioned (5 mm) and embedded in paraffin using routine methods. The paraffin sections were then stained with hematoxylin and eosin. After the slides from each individual patient were reviewed, the main tumor, peritumoral lesions and separated lesions were designated for microdissection. Pathologically distinct lesions located at least 1 cm away from the main tumor were defined as separated lesions; lesions within the vicinity of the main tumor and that did not satisfy the requirements of separated lesions were defined as peritumoral lesions.
DNA was extracted from the paraffin sections according to a previously described procedure with minor modifications.20, 21, 22 Serial sections (4- or 10-μm thick) were made from the paraffin-embedded tissue blocks and placed on glass slides. The 4-μm sections were stained with hematoxylin and eosin, but the 10-μm sections were stained with hematoxylin and eosin after microdissection. Using comparative microscopic observations of the hematoxylin and eosin-stained sections for orientation, tiny fractions of the epithelial lesions (10 to 1,000 cells) were dissected from the 10-μm section using a micromanipulator (Zeiss, Oberkochen, Germany). These microdissection samples were transferred to 5 μl of proteinase K digestion buffer (20 μg of proteinase K/mL in 10 mM Tris/HCl and 1 mM EDTA [pH 8]) in PCR vials followed by inactivation of the proteinase K. Digestion was performed for 10 min at 55°C to demask the DNA followed by the inactivation of proteinase K (15 min, 96°C).
Enriched PCR by Bst-N1 digestion
The DNA was amplified by PCR according to a previously described method with minor modifications.21, 23 Amplifications were performed using a thermal cycler (Perkin-Elmer Corp., Branchburg, NJ) and a kit (Takara Corp, Tokyo, Japan) according to the following protocol: 30 cycles at 94°C for 1 min, 55°C for 1 min and 72°C for 30 sec followed by an additional 7 min at 72°C. The following synthetic oligonucleotides were used as primers: primer A, 5′-ACTGAATATAAACTTGTGGTAGTTGGACCT-3′; primer B, 5′-TCAAAGAA TGGTCCTGGACC-3′; primer C, 5′-TAATATGTCGACTAAAACAAGATTTACCTC-3′. The underlined bases represent mismatches from the normal K-ras DNA sequence. The first PCR was performed using primers A and B, generating a 157-bp fragment containing 2 Bst-N1 restriction sites in the wild-type K-ras allele; if a mutation was present at the first or 2nd position in codon 12, the fragment only contained one Bst-N1 site. Next, 10 μl of the PCR product was digested with 20 units of Bst-N1 (New England Biolabs, Beverly, MA) in a 25-μl mixture containing 10 mM of DTT and 5 μg of 0.1% BSA. The reaction was continued for 16 hr at 60°C. After enzyme inactivation at 96°C, 5 μl of the product was reamplified using primers A and C for 40 cycles of the protocol used for the first PCR, generating a 135-bp fragment. For each amplification, DNA from Panc1 diluted with normal DNA at a ratio of 1:250 and DNA from MKN1 carrying the wild type K-ras sequence were amplified as positive and negative controls.
Non-RI SSCP analysis of K-ras gene mutations
After confirming the reliability of the 2-step PCR by electrophoresis through an 8.0% acrylamide gel, producing a 135-bp fragment, the K-ras gene mutations were detected using a non-RI single-strand conformation polymorphism (SSCP) analysis, as described previously with minor modifications.21, 24 After denaturation at 85°C for 10 min, 2 μl of the PCR product was mixed with 10 μl of a loading solution containing 90% deionized formamide, 20 mM EDTA and 0.05% bromphenol blue and xylene cyanol. The loading solution (10 μl) was then applied to a 15% polyacrylamide gel (Atto Corp., Tokyo, Japan). Electrophoresis was performed at a constant voltage of 200 V produced using an ECPS 3000/150 power supply (Pharmacia LKB Biotechnology, Tokyo, Japan) in a continuous buffer system consisting of 25 mM Tris and 192 mM glycine. During the electrophoresis, the buffer temperature was adjusted to 18°C using a cooling pipe with continuously circulating, cool (14–16°C) water. The running time was 4 hr. The gel was then stained with silver using a kit purchased from Daiichi Pure Chemical Company, Ltd. (Tokyo, Japan) according to the manufacturer's instructions.
DNA sequencing was performed in the samples that did not exhibit the normal pattern of electrophoretic mobility (n = 154). Sequencing reactions were performed using an ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems), and the products were analyzed using an ABI Prism 310 Genetic Analyzer.
The mean values were compared using a Student's t-test. Statistical comparisons between groups were made using a chi-square test or a Fisher exact test. Postoperative survival rates were calculated using the Kaplan-Meier method, and the differences between curves were measured using the log-rank test. Statistical significance was defined as p < 0.05.
A total of 266 duct lesions from 20 patients with IPMT and 88 duct lesions from 7 patients with DC were identified. Ductal lesions were classified by a pathologist according to the criteria of the World Health Organization's International Histological Classification of Tumours.9 There were 168 lesions (mean 8.4 lesions/case) from IPMT patients (n = 20 cases) and 55 lesions (mean 7.8 lesions/case) from DC patients (n = 7 cases) found in the main tumor. There were 26 lesions from IPMT patients (n = 7 cases) and 16 lesions from DC patients (n = 5 cases) designated as peritumoral lesions. There were 43 lesions from IPMT patients (n = 10 cases) and 6 lesions from DC patients (n = 2 cases) designated as separated lesions (Fig. 1, Table I).
Table I. Genotypes of Micro-Dissected Lesions According to Tumor Site
At least one K-ras mutation was detected in the main tumors of 80.0% (16/20) of IPMT and 100% of DC patients, as shown in Table II (not a significant difference). At least one K-ras mutation was detected in 83.3% (5/6) of the IPMT-carcinoma and in 78.6% (11/14) of the IPMT-adenoma patients.
Table II. Number and Variety of K-ras Mutations Found in Main Tumors of IPMT and DC Patients
More than 2 kinds of K-ras mutation were detected in the main tumors of 43.8% (7/16) of IPMT patients with multiple K-ras mutations in their main tumor and in 0% (0/7) of the DC patients (IPMT vs. DC, p < 0.05). More than 2 kinds of K-ras mutation were detected in the main tumors of 60% (3/5) of IPMT-carcinoma patients with multiple K-ras mutations in their main tumor and in 36.4% (4/11) of the IPMT-adenoma patients (IPMT-carcinoma vs. DC, p < 0.05).
The following transitions were observed in the main tumors of DC patients: GGT to GAT (4/7 cases, 57.1%) and GGT to GTT (3/7 cases, 42.9%). In the IPMT patients, the following single K-ras mutations were found: GGT to GAT (4/9 cases, 44.4%), GGT to GTT (3/9 cases, 33.3%), GGT to TGT (1/9 cases, 11.1%) and GGT to GCT (1/9 cases, 11.1%). In IPMT patients with multiple K-ras mutations, the following transitions were detected: GGT to GAT (5/7 cases, 71.4%), GGT to GTT (4/7 cases, 57.1%), GGT to AGT (2/7 cases, 28.6%) and GGT to CGT (2/7 cases, 28.6%).
K-ras mutations in peritumoral lesions were observed in 66.7% (4/6) of IPMT patients who had one or more K-ras mutations in their main tumor and in 80% (4/5) of DC patients with K-ras mutations in main tumor. Identical mutation patterns between main tumor and peritumoral lesions were observed in 100% (4/4) of IPMT patients with K-ras mutations in both main tumor and peritumoral lesions and in 50% (2/4) of DC patients.
K-ras mutations in separated lesions were observed in 62.5% (5/8) of IPMT patients who had one or more K-ras mutations in main tumor and in 50% (1/2) of DC patients with K-ras mutations in main tumor. Identical mutation patterns between main tumor and separated lesions in at least one lesion were observed in 100% (5/5) of IPMT patients with K-ras mutations in both main tumor and separated lesions and in 0% (0/1) of the DC patient. Different mutation patterns between main tumor and separated lesions were observed in 40% (2/5) of IPMT patients with K-ras mutations in both main tumor and separated lesions and in 100% (1/1) of the DC patient.
The number of lesions with a K-ras mutation in patients with different stages of neoplasia showed the following: normal epithelium 6.9% (2/29), hyperplasia (flat and papillary hyperplasia) 29.8% (28/94), adenoma (benign and borderline adenoma) 31.4% (37/118), IPMT-carcinoma 48% (12/25) and DC-carcinoma 47.3% (26/55). The number of lesions with a K-ras mutation increases as the histologic grade becomes more severe.
The clinicopathologic characteristics of the patients are shown in Table III. The average age of the 6 patients with IPMT-carcinoma was 72 years (range, 66–80 years), while that of the 14 patients with IPMT-adenoma was 64 years (range, 52–73 years) and that of the 7 patients with DC was 63 years (range, 52–78 years). The IPMT-carcinoma group was older than the DC group and significantly older than the IPMT-adenoma group (p < 0.05).
Table III. Clinicopathological Characteristics of Patients According to Histological Findings
The 5-year survival rate after resection for the IPMT group was 94.7% (IPMT-carcinoma group, 83.3%; IPMT-adenoma, 100%), which was significantly better than that of the DC group (p = 0.0021; Fig. 2). Survival rate of IPMT-carcinoma patients with one type K-ras mutation in primary tumor was better than that with more than 2 types K-ras mutation in primary tumor (Fig. 3).
The progression of human pancreatic carcinoma is strongly associated with the presence of K-ras mutations, with mutations found in 70–100% of lesions.10, 11, 12, 13, 14, 15, 16 Thus, K-ras mutation may be an important event in the neoplastic process. Recently, K-ras mutations have been shown to occur at a relatively early stage of carcinogenesis in pancreas.7, 16, 22 In our study, at least one K-ras mutation was found in 80% of IPMT patients and 100% of DC patients (no significant difference). The reported percentage of IPMT patients with a K-ras mutation ranges from 0–86%5, 17, 18, 25, 26, 27, 28 but was 81 and 86% in the two most recent studies.17, 18 The most probable suggested reason for this large discrepancy is that the constituent cells of IPMTs are often heterogeneous; in the more recent studies, multiple specimens were examined in each subject.17 Our results show that the frequency at which K-ras mutations were observed was not different between IPMT and DC patients.
In our present study, only one kind of mutation was observed in the main tumor of each DC subject, whereas more than 2 types of K-ras mutation were observed in 43.8% of IPMT patients (p < 0.05). In particular, more than 2 types of K-ras mutation were observed in 60% of IPMT-carcinoma patients; this incidence was higher than that of IPMT-adenoma patients, but the difference was not significant. These results support the heterogeneous nature of IPMT cells and suggest that the degree of heterogeneity may progress as the tumor advances. Lyons et al.19 concluded that ras mutations in myelodysplastic syndromes were heterogeneous and suggested that they may occur at an early stage of carcinogenesis. Lyons et al. further suggested that ras mutations might be useful as a clonal marker for studying myeloid malignancies. Similarly, the presence of different K-ras mutations in the same pancreatic carcinoma increases the likelihood that the tumor arose from multiple foci.15 Previous studies have reported multiple kinds of K-ras mutations in 3–13% of pancreatic carcinomas.12, 15, 16, 29 Motojima et al.15 reported that the detection of different mutations in the same tumor suggests that pancreatic carcinomas may be multicentric. While Fujii et al.30 reported that the allelic heterogeneity of IPMT might, in part, be due to the slow growth rate of these neoplasms, we believe that the increasing heterogeneity is intrinsic to the carcinogenesis of pancreatic cancer.
The types of K-ras mutations found in the main tumors of the IPMT patients in our study were not particularly different from those in the DC patients. Our analyses showed that all the mutations occurred at codon 12 with the exception of 2 cases where mutations occurred at codon 13 (GGC to GGG). GGT to GAT and GGT to GTT transitions at codon 12 were common in both IPMT and DC patients. These results are compatible with those of previous reports.15, 22
To study the genetic and clonal associations between the main tumor and other pancreatic ductal lesions, we investigated the presence and type of K-ras mutations in main tumor, peritumoral lesions and separated lesions. K-ras mutations were frequently observed not only in main tumor, but also in the peritumoral and separated lesions of both IPMT and DC patients. Without K-ras mutation in main tumor of IPMT patients, the mutation was not observed in peritumoral or separated lesions. In each IPMT patient with one or more peritumoral lesions, at least one identical mutation was observed in both the main tumor and the peritumoral lesions. In DC patients with peritumoral lesions, the same mutation in both locations was observed only 50% of the time. With regard to separated lesions, at least one identical mutation was observed in both the main tumor and the separated lesion in all of the IPMT patients with one or more separated lesions. However, different mutations were also observed in the separated lesions of 40%(2/5) of the IPMT patients with separated lesions. Moskaluk et al.31 reported that K-ras mutations often occur within pancreatic ducts and pancreatic ducts in the parenchyma surrounding K-ras positive tumors may be vulnerable to K-ras mutations. However, the number of lesions with a K-ras mutation increased as the degree of abnormality increased from normal epithelium to hyperplasia, to adenoma and to carcinoma in IPMT patients. These results support the hyperplasia–adenoma–carcinoma sequence in the progression of IPMT. Z'graggen et al.18 reported the same results and concluded that K-ras mutation was an important event in carcinogenesis for most IPMT lesions. Our results also support this conclusion, even though K-ras mutations occur at a relatively early stage in the carcinogenesis of the pancreatic lesions and are liable to occur within the pancreatic ducts of the parenchyma surrounding a K-ras positive tumor.
Clinically, the average ages of the IPMT-adenoma, IPMT-carcinoma and DC patients were 64, 72 and 63 years, respectively. Patients with IPMT-carcinoma tended to be older than patients with DC and were significantly older than patients with IPMT-adenoma. IPMT has mainly been reported in elderly men between the ages of 60 and 70 years. This observation suggests that IPMT may slowly evolve from adenoma to carcinoma according to a multistage process. Patients with IPMT have been previously reported to have a good clinical prognosis after resection because of the slow progression and low malignancy rate associated with this tumor.6, 7, 8 In our study, the 5-year survival rate for IPMT patients who underwent a resection was 94.7% (IPMT-carcinoma, 83.3%; IPMT-adenoma, 100%). This survival rate is significantly better than that for DC. Azar et al.7 reported that the overall actuarial 3-year survival rate of IPMT patients was 79%. Thus, patients with IPMT have a favorable long-term survival rate after adequate resection, unlike patients with DC.
In follow-up examinations for patients who have undergone operations, it is useful to distinguish whether or not the tumor is slow growing. A genetic heterogeneous primary tumor, which has more than 2 types of K-ras mutation in main tumor, may indicate a slow-growing tumor. In fact, the survival rate of IPMT-carcinoma patients with a genetic heterogeneous primary tumor was better than that of IPMT-carcinoma patients with a genetic homogenous primary tumor, which has one type K-ras mutation in main tumor.
In conclusion, IPMTs not only exhibit a clinically characteristic slow growth rate, but their main tumor is remarkably genetic heterogeneity. IPMT is thought to occur according to a hyperplasia – adenoma – carcinoma sequence, because of at least one identical mutation between the main tumor and the peritumoral lesion and the stepwise increase of K-ras mutations from normal epithelium to hyperplasia, to adenoma and to carcinoma. K-ras mutation appears to be an important event in IPMT carcinogenesis and occurs at a relative early stage, as in DC carcinogenesis.
The authors thank Mrs. Y. Nakamura, Mr. T. Muramatsu and M. Nakano for their technical assistance.