Iacobuzio-Donahue, Christine A. M.D., Ph.D.; Wilentz, Robb E. M.D.; Argani, Pedram M.D.; Yeo, Charles J. M.D.; Cameron, John L. M.D.; Kern, Scott E. M.D.; Hruban, Ralph H. M.D.
Mucinous cystic neoplasms (MCNs) are a distinct form of pancreatic neoplasm characterized by the presence of neoplastic cysts not connected to the native pancreatic duct system which contain a proliferation of mucinous epithelium surrounded by ovarian-like stroma. These tumors can be placed into one of four categories based on the appearance of the epithelial component after complete histologic examination[colon] invasive mucinous cystadenocarcinoma, mucinous cystic neoplasm with carcinoma in situ, borderline mucinous cystic neoplasm, and mucinous cystadenoma. 1[ndash]3 Until recently, the clinical behavior of MCNs has been a source of debate, as some authors reported that histologically benign neoplasms had the capacity to recur and metastasize. 5,6 Most of these histologically benign MCNs which metastasize are, in fact, undersampled or incompletely resected invasive mucinous cystic carcinomas. For example, Wilentz et al. followed 61 patients with completely resected and entirely examined MCNs, and found that only those MCNs with an associated invasive carcinoma recurred or metastasized. 22,28
If unresected, it is thought that MCNs can progress from mucinous cystadenoma, to borderline mucinous cystic neoplasm, to mucinous cystic neoplasm with carcinoma in situ, to invasive mucinous cystadenocarcinoma. 6,15,21 Thus, a neoplastic progression of MCNs is postulated, similar to that described for adenomas leading to carcinomas in the colon or for pancreatic intraepithelial neoplasia (PanIN) leading to infiltrating duct carcinomas in the pancreas. 11,12,15,18,23[ndash]26 Indeed, several groups have shown a higher frequency of various genetic alterations in more advanced MCNs. 4,12,17 For example, activating point mutations of the K-ras gene have been shown to occur with greater frequency in MCNs with severe epithelial dysplasia or histologic evidence of invasion, compared with noninvasive MCNs with lesser degrees of epithelial dysplasia. 4,17 Similar findings have been reported for the p53 tumor-suppressor gene, 12,17 with an increased frequency of p53 immunolabeling detectable in advanced MCNs (carcinoma in situ or invasive carcinomas), but not in early MCNs associated with no or borderline epithelial changes.
Unlike K-ras and p53, which have been shown to undergo genetic alterations in MCNs, the role of DPC4 (MADH4, SMAD4) tumor-suppressor gene alterations in the development of MCNs has not been examined. DPC4 is a tumor-suppressor gene on chromosome 18q that mediates the downstream effects of the TGF-beta superfamily signaling, resulting in growth inhibition and apoptosis. 7[ndash]10 Inactivation of the DPC4 tumor-suppressor gene is relatively specific for pancreatic adenocarcinoma, with genetic inactivation detectable in approximately 55[percnt] of conventional pancreatic ductal adenocarcinomas. 14,19,20 In contrast, inactivation of the DPC4 gene is rare in intraductal papillary mucinous neoplasms (IPMNs) of the pancreas, with up to 97[percnt] of the invasive carcinomas associated with IPMNs showing strong and diffuse Dpc4 protein expression. 16 The differences in frequency of DPC4 gene inactivation between these two forms of pancreatic neoplasm may account for the significantly different clinical outcomes observed for these two neoplasms.
Recently, Wilentz et al. showed that immunohistochemical labeling for the Dpc4 protein is an extremely sensitive and specific marker for DPC4 genetic alterations, 27 thus providing a simple test for DPC4 gene status in archival material. Therefore, we analyzed Dpc4 protein expression in a spectrum of MCNs. By determining the patterns of Dpc4 expression in this large set of MCNs representing the neoplastic spectrum of this tumor, we sought to determine the role of DPC4 gene inactivation in the genetic progression of this distinctive pancreatic neoplasm.
MATERIALS AND METHODS
Pancreaticoduodenectomy (Whipple resections) and distal pancreatectomy specimens from 36 patients with MCNs were collected from the files of The Johns Hopkins Hospital from 1987 to 1998. Clinical and pathologic data were obtained from the patients' medical records and The Johns Hopkins Hospital Surgical Pathology files. Clinical and pathologic variables included age, gender, tumor size, tumor location, grade of epithelial dysplasia, presence or absence of an invasive carcinoma, and patient survival.
Grading of Dysplasia
The criteria for inclusion of a tumor as a mucinous cystic neoplasm were an intrapancreatic cystic neoplasm containing a proliferation of mucinous lining epithelium not connected to a native pancreatic duct and surrounded by an ovarian-like or dense hyalinized stroma. Each neoplasm was then further categorized according to accepted pathologic criteria 15,21,25 by three of the authors (C.I.D., R.E.W., and R.H.H.) with agreement in all cases. Briefly, mucinous cystadenomas showed a lining mucinous epithelium with no significant cellular or architectural atypia. Those categorized as borderline mucinous cystic neoplasms showed moderate cellular or architectural atypia, including nuclear enlargement, irregularity, and hyperchromatism, loss of nuclear polarity, nucleolar prominence, and papillary complexity. In situ carcinomas showed severe cytologic atypia, cribriform or bridging structures without fibrovascular cores and/or mitoses. Invasive mucinous cystadenomas were mucinous cystic neoplasms with an associated invasive carcinoma within the stroma of the neoplasm, adjacent pancreatic parenchyma, or peripancreatic soft tissues.
The hematoxylin and eosin-stained slides from each of the cases were screened by light microscopy for sections containing an MCN and adjacent normal pancreas. Unstained 5-[mgr]m sections were then cut from the paraffin block selected for each case and deparaffinized by routine techniques. Next, slides were treated with 1[times] sodium citrate buffer (diluted from 10[times] heat-induced epitope retrieval buffer; Ventana-Bio Tek Solutions, Tucson, AZ, USA) before steaming for 20 minutes at 80[deg]C. Slides were then cooled 5 minutes before incubating with a 1[colon]100 dilution of monoclonal antibody to Dpc4 protein (clone B8; Santa Cruz, CA, USA, 1[colon]100 dilution) using the Bio Tek-Mate 1000 automated stainer (Ventana-Bio Tek Solutions). Finally, anti-Dpc4 antibody was detected by adding secondary antibody followed by avidin-biotin complex and 3,3[acute]-diaminobenzidine chromagen. Sections were counterstained with hematoxylin.
Each immunolabeling procedure included an infiltrating pancreatic ductal adenocarcinoma previously shown to contain at least one wild-type DPC4 allele as a positive control and a pancreatic ductal adenocarcinoma with a known homozygous deletion of DPC4 as a negative control. 13,14
Immunohistochemical labeling of Dpc4 was evaluated by three of the authors (C.I.D., R.H.H., and R.E.W.), with agreement in all cases examined. The immunolabeling pattern of each case was scored as positive, weakly positive, or negative. Neoplasms scored as positive showed strong, diffuse cytoplasmic labeling of the neoplastic epithelium, with scattered immunoreactive nuclei also identified. Neoplasms scored as weakly positive showed only faint cytoplasmic labeling of neoplastic epithelium, with scattered immunoreactive nuclei present also. Neoplasms scored as negative showed no detectable cytoplasmic or nuclear Dpc4 protein. Positive and weakly positive cases were combined into a [ldquo]positive[rdquo] category for statistical analysis. The extent of immunolabeling was also categorized into diffuse (if the entire neoplasm labeled) or focal (if only focal expression was noted). In each case, the ovarian-like stroma was also scored as positive or negative for Dpc4 protein expression.
Normal pancreatic ducts, islets of Langerhan's, pancreatic acini, lymphocytes, and stromal fibroblasts, all which had moderate expression of the DPC4 gene product, served as a positive internal control in each of the sections.
Summary data are expressed as mean [plusmn] standard deviation unless otherwise indicated. When comparing differences between distributions, the Fisher exact test for sample sizes less than n [equals] 5 was used. Means were compared with a two-tailed t test or the nonparametric Mann-Whitney rank sum test. The latter was used to compare distributions when the assumption of normality was not valid. Probability values of 0.05 or less were considered significant.
Clinical and pathologic characteristics of the 36 patients with an MCN were collected and analyzed. The mean age of all patients was 51.8 [plusmn] 17.6 years, representing 29 women and 7 men. Twenty-five MCNs were located in the tail of the pancreas (69[percnt]) and the remaining 11 (31[percnt]) were located in the pancreatic head. The mean size of the 35 tumors with available size data was 5.9 [plusmn] 3.1 cm.
Of the 36 MCNs collected, 23 (64[percnt]) were classified as a mucinous cystadenoma, 1 (3[percnt]) as a borderline mucinous cystic neoplasm, 5 (14[percnt]) as a mucinous cystic neoplasm with carcinoma in situ, and 7 (19[percnt]) as an invasive mucinous cystadenocarcinoma. Three of the invasive carcinomas were classified as a mucinous (colloid) carcinoma, and four were classified as tubular (ductal) carcinomas. Lymph node metastases were present in 2 (29[percnt]) of the 7 invasive mucinous cystadenocarcinomas at the time of surgery.
Follow-up data was available for all seven patients with an invasive MCN, and for 27 of the 29 patients with a noninvasive MCN with a mean/median follow up for all cases of 49.9/39.5 months. Among the seven patients with an invasive MCN, 3 (43[percnt]) had expired of their disease, compared with only 1 of the 27 (4[percnt]) patients with noninvasive MCNs (p [equals] 0.002). This latter patient with a noninvasive MCN died of other causes.
Dpc4 Immunohistochemical Staining of MCNs
Dpc4 immunohistochemical labeling was detected in the mucinous epithelium of 30 of 36 (83[percnt]) MCNs analyzed. In all cases with positive Dpc4 expression, epithelial labeling was detected as strong, diffuse cytoplasmic labeling with occasional nuclear labeling also observed. In the remaining six MCNs, no detectable labeling of Dpc4 protein was observed in the mucinous epithelium. However, in all 36 cases (100[percnt]), the adjacent stroma expressed Dpc4 protein irrespective of the presence or absence of epithelial labeling for Dpc4.
Dpc4 labeling of the epithelial components of MCNs was then evaluated with respect to tumor morphology. All 23 mucinous cystadenomas (100[percnt]), the 1 borderline mucinous cystic neoplasm (100[percnt]), and all 5 mucinous cystic neoplasms with carcinoma in situ (100[percnt]) showed strong positive labeling for Dpc4 protein (Fig. 1A, B). In contrast, the invasive component of only 1 of 7 (14[percnt]) invasive mucinous cystadenocarcinomas showed strong positive labeling, with the invasive components of the remaining 6 invasive tumors negative for Dpc4 protein (Fig. 1C). The one invasive cystadenocarcinoma with positive labeling was a tubular (ductal) carcinoma. A comparison of the frequency of Dpc4 protein expression in the noninvasive MCNs (29 of 29) versus the invasive MCNs (1 of 7) was significantly different (p [equals] 0.0001).
MCNs are thought to progress from mucinous cystadenomas, to borderline mucinous cystic neoplasms, to mucinous cystic neoplasms with carcinoma in situ, to invasive mucinous cystadenocarcinomas. The regions of the invasive MCNs that showed less advanced histopathologic features were therefore examined. In all seven invasive MCNs, regions were identified showing mucinous epithelium of the borderline or carcinoma in situ type. Interestingly, in the six cases in which the invasive carcinoma did not label, areas of noninvasive mucinous epithelium with carcinoma in situ were found which showed positive labeling for Dpc4 protein.
Relationship of Dpc4 Immunohistochemical Staining to Clinicopathologic Data
Clinical and pathologic data on the 36 MCNs were collected and analyzed with respect to Dpc4 expression. Because of the high frequency of expression of Dpc4 in our set of MCNs (30 of 36; 83[percnt]), no significant differences were found with respect to age, gender, tumor location, or patient survival. However, a significant difference with respect to tumor size was found for those tumors which expressed Dpc4 (mean tumor size, 5.4 [plusmn] 2.6 cm) compared with those that did not (mean tumor size, 8.3 [plusmn] 4.3 cm; p [equals] 0.03).
Mucinous cystic neoplasms are a distinctive type of pancreatic neoplasm characterized by a cystic neoplasm not connected to the native pancreatic duct system and containing a proliferation of mucinous lining epithelium surrounded by a dense, often ovarian-like stroma. It is thought that MCNs can undergo a neoplastic progression leading to invasive carcinoma, similar to that described in the development of colon carcinomas and pancreatic ductal adenocarcinomas. 11,12,15,18,23[ndash]26 Activating point mutations in the K-ras oncogene have been reported in 46[percnt] of MCNs, 4,17 and accumulation of the p53 gene product in up to 80[percnt] of tumors has also been reported. 12,17 In both examples, the frequency of K-ras gene mutation or p53 immunolabeling was greater in more advanced MCNs, in keeping with the model of genetic progression of these tumors.
Recently, a sensitive and specific immunohistochemical stain for the DPC4 gene product has been developed. 27 We therefore examined the expression of Dpc4 protein in 36 MCNs representing the range of histologic types of these neoplasms. Strong and diffuse expression of Dpc4 protein was detected in 30 of the 36 cases (83[percnt]). When stratified into the four histopathologic types (mucinous cystadenoma, borderline mucinous cystic neoplasm, mucinous cystic neoplasm with carcinoma in situ, and invasive mucinous cystadenocarcinoma), all 29 of the 29 (100[percnt]) noninvasive MCNs labeled for Dpc4 protein, compared with the invasive MCNs of which only 1 of 7 (14[percnt]) labeled for Dpc4 protein. This difference in expression was highly significant (p [equals] 0.0001). Loss of Dpc4 protein expression was also found to correlate significantly with tumor size in our cohort of cases. Those MCNs showing loss of Dpc4 expression were larger in size than those retaining expression of Dpc4 (mean tumor size, 8.3 [plusmn] 4.3 cm vs 5.4 [plusmn] 2.6 cm; p [equals] 0.03). This difference, however, most likely reflects the size differences between noninvasive and invasive mucinous cystic neoplasms.
Areas of neoplastic epithelium showing carcinoma in situ were identified within the six invasive MCNs noted for their loss of Dpc4 labeling. These areas were interesting in that they showed foci of carcinoma in situ with strong labeling of Dpc4, suggesting the loss of Dpc4 expression occurs late in neoplastic progression and possibly at the transition from noninvasive carcinoma in situ to invasive MCN. Similar findings have been shown for the tumor-suppressor gene p53 in MCNs, in which increased expression of the p53 gene product has also been demonstrated in advanced tumors. However, when analyzed by histologic grade, p53 immunolabeling was found to be detectable in both MCNs with either carcinoma in situ or invasion. 12,17 Therefore, our data indicate that the genetic inactivation of Dpc4 occurs late in the neoplastic progression of MCNs and possibly following genetic alterations of p53. Furthermore, the high frequency of loss of Dpc4 expression in invasive MCNs compared with noninvasive MCNs suggests a relationship between DPC4 genetic inactivation and the development of invasiveness in these tumors.
Inactivation of the DPC4 tumor-suppressor gene has also been shown in pancreatic adenocarcinomas, with genetic inactivation detectable in 0[percnt] of PanINs 1 and 2, in 30[percnt] of PanIN3s, and in 55[percnt] of infiltrating conventional pancreatic ductal adenocarcinomas. 14,19,20 In contrast, inactivation of DPC4 is rare in IPMNs of the pancreas, with up to 97[percnt] of the invasive carcinomas associated with IPMNs showing strong and diffuse Dpc4 protein expression. 16 Based on these observations, we have postulated that the differences in frequency of DPC4 gene inactivation between these two forms of pancreatic neoplasm may be related to the significantly different clinical outcomes observed for these two neoplasms. That invasive MCNs show loss of Dpc4 immunolabeling, but are biologically less aggressive than conventional ductal adenocarcinomas, suggests that additional alterations may occur in ductal adenocarcinomas in conjunction with DPC4 genetic inactivation, resulting in the more aggressive biologic phenotype observed for these neoplasms.
In summary, we report evidence that DPC4 gene inactivation, detected as loss of Dpc4 protein expression, is a frequent event in invasive MCNs compared with noninvasive MCNs. Loss of Dpc4 expression in invasive MCNs may play a role in the progression of these neoplasms from noninvasive MCNs with carcinoma in situ to invasive MCN. Further studies are warranted to determine additional genetic events that may be important in the development and genetic progression of these tumors, and especially to determine the physiological changes that are directly attributable to DPC4 gene inactivation in their malignant progression.
1. Albores-Saavedra J, Angeles-Angeles A, Nadji M, et al. Mucinous cystadenocarcinoma of the pancreas. Morphologic and immunocytochemical observations. Am J Surg Pathol 1987; 11[colon]11[ndash]20.
2. Albores-Saavedra J, Gould EW, Angeles-Angeles A, et al. Cystic tumors of the pancreas. Pathol Annu 1990; 25[colon]19[ndash]50.
3. Albores-Saavedra J, Nadji M, Henson DE, et al. Entero-endocrine cell differentiation in carcinomas of the gallbladder and mucinous cystadenocarcinomas of the pancreas. Pathol Res Pract 1998; 183[colon]169[ndash]75.
4. Bartsch D, Bastian D, Barth P, et al. K-ras oncogene mutations indicate malignancy in cystic tumors of the pancreas. Ann Surg 1998; 228[colon]79[ndash]86.
5. Compagno J, Oertel JE. Microcystic adenomas of the pancreas (glycogen-rich cystadenomas). A clinicopathologic study of 34 cases. Am J Clin Pathol 1978; 69[colon]289[ndash]98.
6. Compagno J, Oertel JE. Mucinous cystic neoplasms of the pancreas with overt and latent malignancy (cystadenocarcinoma and cystadenoma). A clinicopathologic study of 41 cases. Am J Clin Pathol 1978; 69[colon]573[ndash]80.
7. Dai JL, Schutte M, Bansal RK, et al. TGF[bgr] responsiveness in DPC4/SMAD4-null cancer cells. Mol Carcinog 1999; 26[colon]37[ndash]43.
8. Dai JL, Turnacioglu KK, Schutte M, et al. Dpc4 transcriptional activation and dysfunction in cancer cells. Cancer Res 1998; 58[colon]4592[ndash]7.
9. Dai JL, Ansal RK, Kern SE. G1 cell cycle arrest and apoptosis induction by nuclear Smad4/Dpc4[colon] phenotypes reversed by a tumorigenic mutation. Proc Natl Acad Sci USA 1999; 96[colon]1427[ndash]32.
10. de Winter JP, Roelen BA, ten Dijke P, et al. DPC4 (SMAD4) mediates transforming growth factor-beta1 (TGF-beta1) induced growth inhibition and transcriptional response in breast tumour cells. Oncogene 1997; 14[colon]1891[ndash]9.
11. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61[colon]759[ndash]67.
12. Flejou JF, Boulange B, Bernades P, et al. p53 protein expression and DNA ploidy in cystic tumors of the pancreas. Pancreas 1996; 13[colon]247[ndash]52.
13. Hahn SA, Hoque ATMS, Moskaluk CA, et al. Homozygous deletion map at 18q21.1 in pancreatic cancer. Cancer Res 1996; 56[colon]490[ndash]4.
14. Hahn SA, Schutte M, Hoque ATMS, et al. DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 1996; 271[colon]350[ndash]3.
15. Hruban RH, Wilentz RE. Pancreas. In[colon] Weidner N, Cote RJ, Suster S, et al., eds. Modern Surgical Pathology. Philadelphia, PA[colon] WB Saunders, 2000.
16. Iacobuzio-Donahue CA, Klimstra D, Adsay NV, et al. DPC-4 protein is expressed in virtually all human intraductal papillary mucinous neoplasms of the pancreas[colon] comparison with conventional ductal carcinomas. Am J Pathol 2000. In press.
17. Jimenez RE, Warshaw AL, Z'graggen K, et al. Sequential accumulation of k-ras mutations and p53 overexpression in the progression of pancreatic mucinous cystic neoplasms to malignancy. Ann Surg 1999; 230[colon]501[ndash]11.
18. Klimstra D, Longnecker DS. K-ras mutations in pancreatic ductal proliferative lesions. Am J Pathol 1994; 145[colon]1547[ndash]50.
19. Moskaluk CA, Hruban RH, Schutte M, et al. Genomic sequencing of DPC4 in the analysis of familial pancreatic carcinoma. Diagn Mol Pathol 1997; 6[colon]85[ndash]90.
20. Schutte M, Hruban RH, Hedrick L, et al. DPC4 gene in various tumor types. Cancer Res 1996; 56[colon]2527[ndash]30.
21. Wilentz RE, Albores-Saavedra J, Hruban RH. The exocrine pancreas. In[colon] Albores-Saavedra J, Henson DE, eds. The Pathology of Incipient Neoplasia, 3rd ed. New York, NY[colon] Oxford, 2000.
22. Wilentz RE, Albores-Saavedra J, Zahurak M, et al. Pathologic examination accurately predicts prognosis in mucinous cystic neoplasms of the pancreas. Am J Surg Pathol 1999; 23[colon]1320[ndash]7.
23. Wilentz RE, Chung CH, Sturm PDJ, et al. K-ras mutations in the duodenal fluid of patients with pancreatic carcinoma. Cancer 1998; 82[colon]96[ndash]103.
24. Wilentz RE, Geradts J, Maynard R, et al. Inactivation of the p16 (INK4A) tumor-suppressor gene in pancreatic duct lesions[colon] loss of intranuclear expression. Cancer Res 1998; 58[colon]4740[ndash]4.
25. Wilentz RE, Hruban RH. Pathology of cancer of the pancreas. Surg Oncol Clin N Am 1998; 7[colon]43[ndash]65.
26. Wilentz RE, Iacobuzio-Donahue CA, Argani P, et al. Inactivation of the DPC4 tumor-suppressor gene in pancreatic intraepithelial neoplasia (PanIN)[colon] evidence that DPC4 inactivation occurs late in neoplastic progression. Cancer Res 2000; 60[colon]2002[ndash]6.
27. Wilentz RE, Su GH, Dai JL, et al. Immunohistochemical labeling for Dpc4 mirrors genetic status in pancreatic[colon] a new marker of DPC4 inactivation. Am J Pathol 2000; 156[colon]37[ndash]43.
28. Zamboni G, Castelli P, Pea M, et al. Mucinous cystic tumor of the pancreas recurring after 11 years as cystadenocarcinoma with foci of choriocarcinoma and osteoclast-like giant cell tumor. Surg Pathol 1994; 5[colon]253[ndash]62.
[copy] 2000 Lippincott Williams [amp] Wilkins, Inc.