Main

p63 is a member of the p53 family of molecules. It has a high degree of homology with p53, but its function seems to be substantially different. Although considered a regulator of cell cycle and apoptosis like p53, p63 does not appear to be as essential as p53 itself for the cellular machinery.1, 2, 3, 4, 5 For instance, p63-null mice do not develop tumors.4, 6 The role of p63 in tumor progression or suppression seems to be far more limited.2, 4, 5, 7, 8, 9, 10, 11, 12 While the role of p63 in cellular pathways is not well understood, it has been suggested that it is the molecular switch for initiation of the epithelial stratification program.13

Not only does the role, but the regulation of p63 also appear to be complex. Often, there is no correlation between the amplification of p63 gene and the expression levels of p63 protein.3 This is partly attributed to the various post-translational modifications that this molecule undergoes.4 Furthermore, there are various isoforms of p63,13, 14, 15 the expression of which varies significantly from tissue to tissue.15 p63 is expressed both as a multiple alternatively spliced C-terminal isoform and as N-terminally deleted, dominant-negative protein with reciprocal functional regulation.4

The isoform of p63 that is referred to as DeltaN (DNp63), which lacks the transactivation domain, is detectable immunohistochemically in the ‘stem’ cell types (basal cells and myoepithelial cells)14, 15, 16, 17, 18, 19, 20, 21, 22 but is typically lacking in the differentiated epithelial cells except for squamous epithelium.18

DNp63 has been found to have two major applications in diagnostic pathology.

(I) As a marker of ‘stem’ (basal or myoepithelial) cells. Typically, the normal ducts of prostate and breast, unlike the invasive carcinoma glands of these organs, have a continuous peripheral layer of basal/myoepithelial cells. This specific expression of DNp63 has rendered it a highly reliable marker in the distinction of invasive carcinoma from benign ducts in these exocrine organs.14, 15, 16, 17, 18, 19, 20, 21

(II) As a marker of squamous differentiation. Another application of DNp63 in surgical pathology has been identified as its specificity for squamous differentiation. As the key molecule of the epithelial stratification program,13 DNp63 expression is highly specific for squamous differentiation in neoplasia, and is usually not detected in nonsquamous carcinomas such as adenocarcinoma, small cell carcinoma or mesothelioma and others.3, 9, 21, 23, 24, 25, 26 This is valuable in the differential diagnosis of poorly differentiated malignant tumors of undetermined origin.

The experience with p63 expression in the pancreas is limited. Ito et al27 have noted p63 gene amplification in >60% of the ductal adenocarcinomas of the pancreas; however, they have not found any correlation with clinicopathologic findings. Immunohistochemical expression of DNp63, on the other hand, as applied in other exocrine organs, has not been systematically tested in the normal pancreas or in the pathologic conditions of this organ.

Materials and methods

Immunohistochemical expression of DNp63 was tested in 105 cases. A total of 25 examples of non-neoplastic pancreatic tissue, 25 examples of pancreatic intraepithelial neoplasia (PanIN) of various grades, 50 cases of pancreatic adenocarcinoma of ductal type and five examples of adenosquamous/squamous carcinoma were identified from the files of the authors’ institutions (Harper University Hospital, Karmanos Cancer Institute and Wayne State University; VA Medical Center; Providence Hospital). The pathology material, including the surgical pathology reports, routine formalin-fixed, paraffin-embedded and H&E-stained sections of all cases were reviewed.

Sections of the non-neoplastic pancreata included various types of non-neoplastic processes such as squamous/transitional metaplasia (five cases; Figure 1), which can manifest as multilayered epithelium mimicking high-grade PanIN (Figure 2), and incidental microcysts (five cases; Figures 3 and 4), seen amidst normal pancreatic lobules.

Figure 1
figure 1

Squamous metaplasia in pancreatic ducts.

Full size image
Figure 2
figure 2

Squamous metaplasia manifesting as multilayered epithelium with atypia that mimics high-grade PanIN.

Full size image
Figure 3
figure 3

Centroacinar microcyst lined by attenuated cells containing amorphous muco-secretory plugs.

Full size image
Figure 4
figure 4

Centroacinar microcyst with squamoid epithelium.

Full size image

In all cases, immunohistochemical stains for p63 (Santa Cruz Biotechnology, CA, USA; dilution 1:100; treatment Steam&EDTA) were performed. Routine streptavidin–biotin peroxidase method was used. Negative and positive controls (prostate) were included with each batch of slides tested.

Results

No DNp63 expression was noted in normal pancreatic ducts (0/25). All PanINs were negative (0/25; Figure 5). On the other hand, all five foci of squamous/transitional metaplasia showed strong and uniform nuclear positivity for this marker (5/5; Figure 6). Expression was detected in all cells of squamous nature. DNp63 labeling was also seen in five incidental microcysts (5/5). Typically these cysts had a basal layer of ovoid cells with minimal/no cytoplasm, which showed nuclear labeling with p63. There was a separate luminal layer composed of one or more cells, which were nonmucinous and flattened with transparent cytoplasm, and this was negative for p63. Pale, acidophilic concentrations, typical of acinar secretions, were present in the lumen of these cysts (Figure 7). However, the lining cells did not show morphologic features of acinar cells (devoid of any intracytoplasmic granules and no prominent nucleoli present).

Figure 5
figure 5

No expression of DNp63 was noted in PanINs.

Full size image
Figure 6
figure 6

All foci of squamous metaplasia showed strong nuclear positivity for DNp63.

Full size image
Figure 7
figure 7

DNp63 expression was also seen in all centroacinar microcysts.

Full size image

Among 50 invasive ductal carcinomas, DNp63 expression was detected only in areas of squamous differentiation (two cases, Figure 8). Ordinary tubular and poorly differentiated nonsquamous components of these two cases were completely negative. DNp63 was completely absent in ordinary ductal areas (0/48). In addition, the squamous components of all of adenosquamous/squamous carcinoma cases were also strongly positive (5/5).

Figure 8
figure 8

Among invasive carcinoma of the pancreas, DNp63 expression was detected only in areas of squamous differentiation.

Full size image

The results are shown in Table 1.

Table 1 DNp63 expression in pancreatic lesions
Full size table

Discussion

DNp63 is valuable in distinguishing intraductal squamous metaplasia from intraductal neoplasia, because it is consistently expressed in squamous/transitional metaplasia in pancreatic ducts, but typically absent in the PanIN. Transitional/squamous metaplasia typically manifests as multilayered epithelium with some degrees of disorganization, and may therefore mimic high-grade PanINs. PanIN diagnosis is important. Whereas low-grade PanINs are common incidental findings, high-grade PanINs are typically associated with ductal adenocarcinoma.28 Furthermore, high-grade PanINs may have potential for progressing into invasive carcinoma and therefore should duly be reported in the surgical pathology report in a patient with no other neoplasia.29, 30 Therefore, the distinction of PanINs from squamous metaplasia is important, not only for researchers but also for diagnosticians. At the same time, this differential expression of DNp63 raises the interesting question of whether the progression of PanIN can be turned off by induction of p63. As a proapoptotic molecule,15 a homologue of p53, and as the molecule that triggers epithelial stratification,13 p63 is likely to be responsible for squamous transformation of the ductal epithelial cells. Therefore, it is conceivable that activation (or induction) of p63 may divert the intraductal neoplastic cells towards a more mature, metaplastic process.

As in other organs,3, 9, 23, 24, 25 DNp63 is a reliable marker of squamous differentiation also in pancreatic carcinoma. Among invasive carcinomas of the pancreas, DNp63 immuno-expression is largely limited to the areas of squamous differentiation. Ordinary ductal adenocarcinomas, including the poorly differentiated components, are typically negative for this marker. DNp63 may be useful in distinguishing poorly differentiated pancreatic carcinomas from poorly differentiated squamous carcinomas of especially pulmonary origin. Occasionally, this differentiation is important in determining the primary source of metastatic carcinomas, both in the liver and in the pancreas itself. Pulmonary and pancreatic carcinomas are two of the most common primaries that metastasize to the liver,31 and they are also the most common sources of ‘carcinoma of unknown origin’.32, 33 Furthermore, lung cancer is the most common tumor to metastasize to the pancreas.34 Therefore, on occasion this differentiation becomes an important issue and p63 can be successfully utilized in this distinction.

DNp63 is not helpful in the separation of ductal adenocarcinoma from normal ducts. Unlike in the other exocrine organs prostate and breast,14, 15, 16, 17, 18, 19 immunohistochemical expression of DNp63 is not applicable in the pancreas to distinguish normal ducts from invasive carcinoma. In fact, the lack of DNp63 labeling in this study confirms the absence of a basal/myoepithelial cell layer in this organ.

This study also confirms our previous finding that there are no identifiable stem cells residing within the pancreatic ducts.35 Some authors believe that pancreatic ducts contain stem cells,36 which are presumed to be the counterparts of ‘oval’ cells in the liver (Carolyn C Compton, September 1999, personal communications).37 There are indeed rare, scattered nonductal cells within the ductal epithelium. These were originally identified by Feyrter,38 who referred to them as ‘helle Zellen’. In our experience,35 these cells usually prove to be endocrine in nature by immunohistochemistry and they probably do not represent stem cells. In this study, these cells have failed to label with DNp63, which is reported as a ‘stem cell marker’ by some authors,14, 22 hence bringing further doubt to the ‘stem’ cell nature of these cells.

This study has also shown that those incidental microcysts within normal pancreatic parenchyma, which generally measure no more than 2–3 mm, are metaplastic in nature. They are typically lined by attenuated cells that cytologically resemble centroacinar cells (neither ductal nor acinar in morphology), and they contain amorphous acidophilic muco-secretory plugs characteristic of enzymatic concretions, suggesting that these cysts actively communicate with the acini, and may represent dilatation of the centroacinar system. For this reason, we have been referring to these cysts as ‘centroacinar microcyst’. The nature of these microcysts, which are relatively common incidental findings, has been difficult to determine. It is also difficult to determine whether they are related to the recently described acinar cell cystadenoma.39 In this study, these microcysts were consistently positive for DNp63 in contrast to the rest of the pancreatic parenchyma. This suggests that the cells lining these cysts, especially those that are basally located, had undergone squamous transformation, that is, a metaplastic process.

In conclusion, DNp63 expression in the pancreas appears to be limited to squamous differentiation and may aid in distinguishing squamous/transitional metaplasia from PanINs. It is also a marker for squamous differentiation in invasive carcinoma. Unlike the breast or prostate, there are no DNp63-expressing basal/myoepithelial cells in the pancreas; therefore, this marker is not applicable in the distinction of invasive carcinoma from benign ducts.