Virchows Archiv

, 453:521

Aberrant expression of an “intestinal marker” Cdx2 in pyloric gland adenoma of the gallbladder

Authors

    • Department of PathologyKurashiki Central Hospital
  • Kenji Notohara
    • Department of PathologyKurashiki Central Hospital
  • Masayoshi Fujisawa
    • Department of PathologyHimeji Red Cross Hospital
Original Article

DOI: 10.1007/s00428-008-0680-z

Cite this article as:
Wani, Y., Notohara, K. & Fujisawa, M. Virchows Arch (2008) 453: 521. doi:10.1007/s00428-008-0680-z
  • 136 Views

Abstract

The aim of this study was to survey Cdx2 expression in pyloric gland adenoma (PGA) of the gallbladder. We reviewed 29 PGA cases, ten (34.4%) and seven (24.1%) of which showed intestinal metaplasia (IM) and squamous morule (SM), respectively. The immunostaining for Cdx2, beta-catenin, MUC5AC, MUC2, MUC6, and M-GGMC-1 was performed and scored (0 = negative, 1+ = <10%, 2+ = 10% to <30%, 3+ = 30% to <50%, 4+ = 50% to <70%, 5+ = 70–100%). Although its scores were relatively low (1+ or 2+), Cdx2 was frequently expressed in 27 cases (93.1%). Not only goblet and/or Paneth cells were positive but also non-IM cells in PGAs, as opposed to the lack of staining in the background mucosa. Cdx2 scores were not correlated with those of IM (p = 0.485) and MUC2 (p = 0.868). Of note, Cdx2 was positive in foci of SM in all seven cases, and there was a significant difference in Cdx2 scores between PGAs with and without SM. Furthermore, the p value of scores between Cdx2 and beta-catenin was 0.051, and both mean labeling indices (LIs) were correlated (r = 0.736). With Cdx2, higher morular LIs than glandular LIs were observed (p = 0.001). Finally, we concluded that aberrant Cdx2 expression in PGAs is closely associated with nuclear beta-catenin expression and SM in contrast with IM.

Keywords

Cdx2Beta-cateninPyloric gland adenomaGallbladderImmunohistochemistry

Introduction

The Cdx2 protein is an intestine-specific homeobox gene transcription factor that plays an important role in regulating the differentiation of the intestinal epithelium [1]. The expression in adult normal tissue is restricted to the nuclei of the intestinal epithelium from the duodenum to the rectum [2, 3]. Its expression, however, is not limited to normal mucosa because it has also been reported in colonic adenocarcinoma [2, 4]. Furthermore, Cdx2 expression has been documented in adenocarcinomas at other sites such as the stomach, pancreas, urinary bladder, and ovary; this expression has been interpreted as intestinal differentiation [48].

In gallbladder, Cdx2 expression has been observed in the mucosa in chronic cholecystitis, cholelithiasis, and adenocarcinoma and has been found to be related to MUC2 expression [912]. These phenomena have also been interpreted as intestinal differentiation. In tubular adenoma, the pyloric gland-type (or pyloric gland adenoma; PGA) of the gallbladder, which is composed of lobules that contain closely packed pyloric type tubular glands, only a report of Cdx2 has been published [13]. According to that, none of the 58 PGA cases showed Cdx2 expression.

We have clarified that Cdx2 protein is frequently aberrantly expressed in endometrial lesions with squamous differentiation, especially morular-type differentiation, and strongly correlates with nuclear expression of beta-catenin [14]. Furthermore, Saegusa et al. proved that in cell lines of endometrial carcinomas, overexpression of the active form of beta-catenin results in a significant increase in endogenous Cdx2 expression at both mRNA and protein levels [15]. They concluded that an association between Cdx2 and beta-catenin signaling might participate in an induction of transdifferentiation of endometrioid carcinoma cells. In PGA, squamous morules (or spindle cell metaplasias; SMs) are also present in 23% of cases [16]. SM is defined as having discrete nests or uniformly demarcated sheets of nonkeratinizing (immature) squamous cells, which are oval or spindle-shaped cells appearing bland and uniform and lacking prominent nucleoli [1719]. These cells sometimes contain biotin-rich optically clear nuclei [18, 20]. Mitotic figures are usually not seen. Immunohistochemically, nuclear expression of beta-catenin, estrogen-beta receptor, p53, cyclin D1, and low MIB-1 index is shown in the foci of SM [18, 21] In addition, the gene mutation of beta-catenin in PGA has been established as well as in endometrial lesions [22, 23]. Given that PGA and endometrial lesions have SM in common, aberrant Cdx2 expression in PGA can also be presumed in a mode other than intestinal differentiation.

The aim of the present study was to survey Cdx2 expression in PGA of the gallbladder and to determine whether Cdx2 expression is related to intestinal differentiation, beta-catenin expression, or SM.

Materials and methods

We retrieved 29 cases of PGAs from the files of Kurashiki Central Hospital and Himeji Red Cross Hospital. We reviewed HE slides to note the size of the lesion and the presence or absence of IM and SM. With IM, a small number of admixed cells (Paneth cells and/or goblet cells) were also noted. In immunohistochemistry, paraffin-embedded sections cut at 4 μm were deparaffinized and rehydrated, and endogenous peroxidase was blocked in 3% hydrogen peroxidase. In all specimens, immunostainings were performed for beta-catenin and for mucin markers such as MUC5AC for gastric foveolar-type epithelium, MUC6, M-GGMC-1 for the pyloric-gland type, and MUC2 for the intestinal goblet-cell type, using suitable antigen retrieval for each, and detection was carried out with the Envision plus detection system (DAKO, Kyoto, Japan; Table 1). Diaminobenzidine was employed as a chromogen. The immunostaining was scored as follows: 0 = negative, 1+ = <10% positive cells, 2+ = 10% to <30%, 3+ = 30% to <50%, 4+ = 50% to <70%, 5+ = 70–100%. In a closer survey of the relationship between Cdx2 and beta-catenin, the number of positive nuclei for Cdx2 and beta-catenin was counted in the well-labeled high power fields (HPFs). The mean labeling indices (LIs; %) were calculated by counting 1,000 cells in three different HPFs. Moreover, with SM cases, morular and glandular LI (%) were also calculated by counting positive cells in all foci of SM and in the surrounding glandular cells (at least 1,000 cells per a focus of SM).
Table 1

Immunohistochemical panel

Primary antibody

Clone

Company

Dilution

Antigen retrieval

Cdx2

CDX2–88

BioGenex, San Ramon, CA, USA

1:100

MW (EDTA, pH 7.0)

Beta-catenin

17C2

Novocastra, Newcastle, UK

1:50

MW (citrate, pH 6.0) + PC

MUC2

Ccp58

Novocastra, Newcastle, UK

1:100

MW (citrate, pH 6.0)

MUC6

CLH5

Novocastra, Newcastle, UK

1:100

MW (citrate, pH 6.0)

M-GGMC-1

HIK1083

Kanto Chemical, Tokyo, Japan

1:25

MW

MUC5AC

CLH2

Novocastra, Newcastle, UK

1:100

MW (citrate, pH 6.0)

MW microwave, PC pressure cooker

All statistical analyses were carried out using the SPSS software package version 15.0 (SPSS Inc., Chicago, IL, USA). Comparative data were analyzed using Pearson’s chi square test, Linear-by-Linear association, Spearman’s correlation coefficient, and the Mann–Whitney U test. A p value of <0.05 was considered statistically significant.

Results

PGA ranged in size from 3 to 16 mm (mean 8.2 mm). Seven cases (24.1%) contained foci of SM. Ten cases (34.4%) had small but distinct foci of goblet cells and/or Paneth cells. Of these, only two cases contained both metaplasias. The immunohistochemical results are listed in Table 2. In the background mucosa of all 18 cases assessable, Cdx2 expression was completely absent. Although its scores were low (1+ or 2+), Cdx2 was frequently expressed in 27 cases (93.1%). Not only goblet and/or Paneth cells, but also other cells in PGAs were found to be positive for the marker, in contrast with the absence of staining in the background mucosa (Figs. 1a–c, 2a–d). Of note, Cdx2 was also positive in foci of SM in all seven cases (100%; Fig. 2f,h). Statistically, Cdx2 scores were not correlated with those of IM (p = 0.485) and MUC2 (p = 0.868; Tables 3 and 4; Fig. 1g). However, only a small number of goblet and/or Paneth cells in five cases were positive for MUC2 (17.2%), where Cdx2-positive cells were also seen (Table 2; Fig. 2b). The Cdx2 score was not statistically correlated with that of MUC6 (p = 0.871), M-GGMC-1 (p = 0.878), and MUC5AC (p = 0.499; Table 4; Fig. 1e,f,h). In contrast, there was a significant difference in Cdx2 scores between PGAs with and without SM (Table 3). Furthermore, the p value of scores between Cdx2 and beta-catenin was just around a cut-off level (0.051; Table 4; Fig. 1d); however, both of the mean LIs, which were counted in well-labeled fields, were correlated (r = 0.736; Fig. 3). With both Cdx2 and beta-catenin in SM cases, the morular LIs were significantly higher than the glandular LIs (p = 0.001, 0.038, respectively; Fig. 4).
https://static-content.springer.com/image/art%3A10.1007%2Fs00428-008-0680-z/MediaObjects/428_2008_680_Fig1_HTML.gif
Fig. 1

PGA without IM (A HE ×100; B HE ×400). PGA showing the expression of mucin markers MUC6 (C, ×200), M-GGMC-1 (D, ×400), and MUC5AC (E, ×200), lacking MUC2 expression (F, ×200). Likewise, beta-catenin (G, ×400), nuclear Cdx2 expression (H, ×400) is identified in the minority of glandular cells

https://static-content.springer.com/image/art%3A10.1007%2Fs00428-008-0680-z/MediaObjects/428_2008_680_Fig2_HTML.gif
Fig. 2

PGA with IM (AD) and SM (EH). Cdx2 stain is seen in both Goblet (arrow) and Paneth cells containing granules (arrow head; A HE ×200; B, ×400), despite there being no staining in the background mucosa (C HE ×100; D, ×200). In morular foci (E HE ×200), Cdx2 expression (F, ×200; H, ×200) closely resembles nuclear expression of beta-catenin (G, ×200)

https://static-content.springer.com/image/art%3A10.1007%2Fs00428-008-0680-z/MediaObjects/428_2008_680_Fig3_HTML.gif
Fig. 3

Correlation between mLIs of Cdx2 and beta-catenin in PGA. Statistical analysis was performed with Spearman’s correlation coefficient

https://static-content.springer.com/image/art%3A10.1007%2Fs00428-008-0680-z/MediaObjects/428_2008_680_Fig4_HTML.gif
Fig. 4

Comparison of Cdx2 between glandular and morular LIs in PGA. The data are means ± SD values. Statistical analysis was performed with the Mann–Whitney U test

Table 2

Immunohistochemical results

Antibody

Score

0

%

1+

%

2+

%

3+

%

4+

%

5+

%

Cdx2

2

6.9

19

65.5

8

27.6

0

0.0

0

0.0

0

0.0

Beta-catenin

1

3.4

11

37.9

5

17.2

9

31.0

3

10.3

0

0.0

MUC6

0

0.0

0

0.0

2

6.9

5

17.2

3

10.3

19

65.5

M-GGMC-1

0

0.0

1

3.4

2

6.9

10

34.5

8

27.6

8

27.6

MUC2

24

82.8

5

17.2

0

0.0

0

0.0

0

0.0

0

0.0

MUC5AC

1

3.4

16

55.2

9

31.0

3

10.3

0

0.0

0

0.0

Table 3

Comparison of Cdx2 scores between PGA cases with and without metaplasia

 

Cdx2

0

1

2

p Value

PGA with IM

0

8

2

0.485

PGA without IM

2

11

6

 

PGA with SM

0

2

5

0.015*

PGA without SM

2

17

3

 

Statistical analysis was performed with Pearson’s chi square test (p < 0.05)

PGA pyloric gland adenoma, IM intestinal metaplasia, SM squamous morule

Table 4

Correlations of scores between Cdx2 and other markers

 

Cdx2

Beta-catenin

p = 0.051

MUC2

p = 1.000

MUC6

p = 0.871

M-GGMC-1

p = 0.878

MUC5AC

p = 0.499

Statistical analysis was performed with Linear-by-Linear Association

Discussion

In gallbladder, Cdx2 expression, a well-known intestinal marker, has been observed in the mucosa in chronic cholecystitis, cholelithiasis, and adenocarcinoma and has been found to be related to MUC2 expression [912]. These phenomena have also been interpreted as intestinal differentiation. As to PGA, frequent Cdx2 expression (92%) has also been considered to be specific to PGA without any expression in the background mucosa of 18 cases. Our results were completely different from the data by Nagata et al., who described that no 58 PGA cases showed Cdx2 expression [13]. The reason for the dissociation of these results remains unclear. In their study, the number of PGA cases with SM or IM was not elucidated. They might have experienced some PGA cases with a small number of Cdx2-positive cells, corresponding to our “Score 1” cases because they considered that Cdx2 expression was positive when the percentage of positively stained cells was 10% or more. In general, PGA contains tumor cells showing differentiation to goblet cells, Paneth cells, or endocrine cells, which are often admixed, and these admixed cells mean a kind of intestinal metaplasia of PGA [16]. It seems plausible that Cdx2 expression may correspond to the number of metaplastic cells. Indeed, in the present study, Cdx2 expression was focally identified in areas showing goblet and/or Paneth cell-type intestinal differentiation, but it was very limited. PGAs with Paneth cells or endocrine cells showed unquestionable positivity for Cdx2, whereas Cdx2 did not correlate to MUC2 expression (all but five were completely MUC2-negative). This miscorrelation between the two markers is unique and has never been reported in other gallbladder lesions (chronic cholecystitis or carcinoma) [10, 11]. In addition, Cdx2 expression was not confined to the areas or cases clearly showing intestinal metaplasia. Considering the identification of Cdx2 in most PGA cases without IM, Cdx2 expression in PGAs does not largely seem to indicate true intestinal differentiation. Furthermore, neither does it to correlate to expressions of MUC6 and M-GGMC-1 (pyloric phenotypes) nor MUC5AC (gastric foveolar phenotype).

Of note, Cdx2 was positive in foci of SM in all seven cases, and there was a significant difference in scores between cases with and without SM (p = 0.015). The correlation of scores between Cdx2 and beta-catenin was just around a cutoff level value (p = 0.051), and the correlation of mean LIs between Cdx2 and beta-catenin was also identified. Comparing the mean LIs for both markers, LIs of beta-catenin were likely to higher than those of Cdx2. However, foci of SM in all seven cases expressed both markers. In addition, morular LIs of Cdx2 were significantly higher than glandular LIs (p = 0.001), as seen in beta-catenin (p = 0.038). These observations in PGA have never been reported previously. Recently, we clarified that Cdx2 protein is aberrantly expressed in endometrial lesions with squamous differentiation, especially morular-type differentiation, and strongly correlates with nuclear expression of beta-catenin [14]. In addition to some previous reports with beta-catenin mutation in PGAs [22, 23], our results of the correlation of mean LIs between the two and their common expression in SM foci raise the intriguing possibility that the mutation of beta-catenin genes promotes Cdx2 expression, as Saegusa et al. indicated in endometrial lesions [15]. They proved that in cell lines of endometrial carcinomas, overexpression of the active form beta-catenin results in a significant increase in endogenous Cdx2 expression at both mRNA and protein levels. They concluded that an association between Cdx2 and beta-catenin signaling might participate in induction of transdifferentiation of endometrioid carcinoma cells. Further molecular study in PGAs will be necessary to confirm this speculation. In any event, Cdx2 expression has previously only been considered to be “intestinal differentiation” in gastrointestinal and pancreatobiliary systems. We will describe a new aspect of Cdx2 expression related to SM and stable beta-catenin, probably due to the latter gene mutation. Our study can also offer the expression of Cdx2 as one of the traits of PGA.

In summary, we documented frequent Cdx2 expression in PGA of the gallbladder. Unlike other gallbladder lesions (chronic cholecystitis, cholelithiasis, and adenocarcinoma), IM in itself plays a minor role of Cdx2 expression in PGA. Moreover, Cdx2 expression did not correlate to MUC2 expression. Besides, significantly high levels of Cdx2 expression in morular foci and a correlation of mean LIs between Cdx2 and beta-catenin are noteworthy. We, therefore, concluded that aberrant Cdx2 expression in PGAs is closely associated with beta-catenin expression and SM.

Acknowledgements

The authors thank Mr. Hironori Kajitani and Ms. Akemi Minamimoto for their technical support. They also thank Mr. Yoshiharu Yamamoto, Clinical Research Center, Kurashiki Central Hospital, for providing assistance with the statistical analyses.

Disclosure/Conflict of Interest

The authors have no conflict of interest to declare.

Copyright information

© Springer-Verlag 2008