Retained Cell–Cell Adhesion in Serrated Neoplastic Pathway as Opposed to Conventional Colorectal Adenomas

Abstract

The molecular features of serrated polyps of colorectum remain to be elucidated. The expression pattern of adhesive molecules (E-cadherin, α-catenin, and β-catenin) has not been examined in serrated neoplastic pathway. The expression of E-cadherin, α-catenin, and β-catenin were analyzed by immunohistochemistry in 32 hyperplastic polyps (HPs), 28 sessile serrated adenomas (SSAs), 37 traditional serrated adenomas (TSAs), 51 traditional adenomas (TAs), and 10 normal colonic tissues (NCs). Retained membranous expression for E-cadherin, α-catenin, and β-catenin was more frequent in HPs, SSAs, and TSAs than that in TAs (p < 0.001). Nuclear labeling of β-catenin was detected in 19.6% of TAs, but in none of HPs, SSAs, and TSAs (p < 0.001). Cytoplasmic accumulation of β-catenin was found in 3.1% of HPs, 3.6% of SSAs, and 21.6% of TSAs, significantly lower than that in TAs (60.8%, p < 0.001). The membranous co-expression of E-cadherin, α-catenin, and β-catenin was more frequent in HPs (68.8%), SSAs (60.7%), and TSAs (37.8%) than that in TAs (7.8%, p < 0.001). Cell adhesion function is retained in serrated neoplastic pathway. Wnt signaling pathway plays a less active role in the development of colorectal serrated polys than in TAs.

Most colorectal carcinomas are considered to arise from conventional adenoma based on the concept of the adenoma-adenocarcinoma sequence, which is the predominant pathogenic pathway in colorectal tumorigenesis (Hill et al. 1978). However, recent studies have highlighted an alternative “serrated polyp neoplasia pathway” (hyperplastic polyp [HP], serrate adenoma [SA], colorectal carcinoma) (Goldstein 2006; Grady 2007; Kambara et al. 2004).

The primary histologic characteristic of HPs is infolding of the crypt epithelium, leading to a saw-toothed appearance in longitudinal section and a stellate appearance on cross section. In 1990, Fenoglio-Preiser coined the term serrated adenoma (herein reported as traditional serrated adenomas or TSA) for colorectal polyps that share serrated architecture with HP, but show epithelial dysplasia identical to that seen in traditional adenomas (TAs) (Farris et al. 2008). In 1996, the term sessile serrated adenoma (SSA) was introduced to refer to a subset of serrated lesions, which cytologically resembles HPs but is distinguished from HPs on the basis of crypt dilation, branching, and horizontal spreading (Cunningham and Riddell 2006). It was proposed that HPs, TSAs, and SSAs constituted a morphologic continuum within the same neoplastic pathway (Yantiss 2007).

However, investigators are still unable to distinguish molecular abnormalities among these serrated polyps to support the hypothesis that they are associated with each other through the progressive accumulation of genetic events (Yantiss 2007). The evidence reflecting the accumulation of these genetic events along this serrated pathway is lacking.

It is well known that Wnt signal transduction involving β-catenin plays a critical role in colorectal carcinogenesis (Bienz and Clevers 2000; Sena et al. 2006). However, Wnt pathway abnormalities (APC mutation/LOH, β-catenin mutation/nuclear expression) were reported to be infrequent in the SAs (Dehari 2001; Sawyer et al. 2002; Uchida et al. 1998; Yamamoto et al. 2003). Previous studies reported that widespread or focal nuclear β-catenin expression was demonstrated in only 0% to 7% of SAs (BABA et al. 2004; Yamamoto et al. 2003). Moreover, our previous study revealed that cytoplasmic or nuclear accumulation of β-catenin was less frequent in the SAs than that in the traditional TA-carcinoma sequence (Fu et al. 2009). More evidences are needed before we can draw a conclusion that the Wnt signaling pathway involving β-catenin does not play a major role in the tumor progression of serrated neoplastic pathway.

In addition, β-catenin plays an essential role in two different cellular processes: cell–cell adhesion and Wnt signal transduction (Brembeck et al. 2006; Gottardi and Gumbiner 2004). The E-cadherin/α-catenin/β-catenin complex is the predominant complex assembly influencing the formation of cell–cell junctions (El-Bahrawy et al. 2004; Niessen and Gottardi 2008). As a key element, β-catenin participates in both functions in a mutually exclusive manner, placing it at the crossroads between cell adhesion and Wnt signaling (Bienz 2005; Gavard and Mege 2005). Both functions of β-catenin are de-regulated during the development of malignant epithelial cancers, leading both to the loss of cell–cell adhesion and to the increased transcription of Wnt target genes (Brembeck et al. 2006). Interestingly, as mentioned above, β-catenin accumulation was infrequent in SAs as opposed to that in TAs. Another study reported that only 8% (3/39) of TSAs showed reduced membrane expression of E-cadherin (Sawyer et al. 2002). These data prompt us to make the hypothesis that β-catenin, associated with E-cadherin and α-catenin, may be mainly involved in cell adhesion function, but not Wnt signaling in serrated polyps of the colorectum. However, a detailed and systemic study on the expression of adhesion molecules in serrated polyps is still lacking.

In the present study, we examined the expression of E-cadherin, α-catenin, and β-catenin in HPs, SSAs, and TSAs, and compared the findings with NCs and TAs, to determine the role of adhesion molecules in the development of serrated polyps of the colorectum.

Materials and Methods

Sample Collection

We identified 148 polyps with the following original diagnoses from the patients who underwent colonoscopic polypectomy in the Nanfang Hospital Pathology Archives for the period between February 2007 and July 2009. Specimens of 10 normal colorectal mucosas (NC) were used as controls. Using the nomenclature and criteria of Snover et al. (2005), all samples were reviewed by a senior pathologist (Y.Z.) and classified as follows: sporadic HPs (n = 32), SSAs (n = 28, 25 pure-type and 3 mixed-type), TSAs (n = 37, 32 pure-type and 5 mixed-type), and TAs (n = 51). SSA is distinguished from HP on the basis of larger size, aberrant structure, dysplasia, and predilection for proximal colon (Fig. 1) (Torlakovic et al. 2003). Clinicopathologic data for each patient were obtained from hospital records. Informed consent was obtained from all the participants. The project was approved by the Institutional Review Board.

Figure 1.

Histological appearance of a typical sessile serrated adenoma (SSA) and a traditional serrated adenoma (TSA) showing serrated architecture, hematoxylin and eosin (H&E) staining. Scale bars = 100 µm, 100 × magnification.

Immunohistochemical Staining

Indirect immunohistochemistry was performed with formalin-fixed paraffin-embedded tissue sections. After deparaffinized and rehydrated with xylene and ethanol, the sections were heated in a microwave oven in 0.01 mol/l sodium citrate buffer (pH 6.0) for 12 min to retrieve antigens. Endogenous peroxidase activity was inhibited by incubation with 0.3% hydrogen peroxidase in methanol for 10 min. Primary antibodies used were β-catenin (1:100, E-5, sc-7963, Santa Cruz Biotechnology, Santa Cruz, CA), α-catenin (1:100, α-CAT-7A4, Zymed Laboratories Inc., San Francisco, CA), and E-cadherin (1:100, HECD-1, Laboratories Inc., San Francisco, CA). Following the appropriate secondary antibodies, the labeled antigens were visualized by the development of brown pigment via a standard 3,3-diaminobenzidine protocol. Slides were then counterstained lightly with haematoxylin. Phosphate-buffered saline (pH 7.2) was used for rinsing between each step. Normal colonic tissues were included as positive controls. Normal colonic mucosa present in the sections was also used as an internal control, thus allowing direct comparison within the same tissue specimen. Staining without primary antibody was used as a negative control.

Each immunostained section was examined under light microscopy and evaluated by the first author. The pattern of immunolabeling was described as normal for E-cadherin and α-catenin when labeling was exclusively membranous with no cytoplasmic labeling and of similar intensity to adjacent normal epithelium. Abnormal labeling included discontinuous or absent membranous labeling, with or without cytoplasmic labeling (Andrews et al. 1997). In addition, membranous, cytoplasmic, and nuclear staining of β-catenin was considered separately. Membranous and cytoplasmic staining was evaluated as negative or positive. Nuclear expression was evaluated as widespread (>75% of cells per section), focal (<75% of cells per section), or absent (no nuclear staining).

Statistical Analysis

Data are presented as mean and standard deviation for continuous variables and as proportions for categorical variables. Data were analyzed using 1-way ANOVA, followed by Bonferroni test for multiple comparisons. Differences in categorical variables were determined by the chi-square or Fisher’s exact tests, as appropriate. Differences were considered significant if p < 0.05. All significance tests were 2-tailed. All statistical tests were performed using SPSS software, version 13.0 (SPSS Inc., Chicago, IL).

Results

Clinicopathologic Feature

Among the SSAs, 3 contained foci of classic HP adjacent to the SSA (mixed SSA-HP) and 25 were pure SSAs. A total of 5 cases with mixed features of TSA and focal SSA were included in the TSA category, and 32 were pure TSAs. The clinicopathologic findings of patients are shown in Table 1. There were no significant differences in patient gender or age among the five histological types, and there were no significant differences in the incidence of high-grade dysplasia between TSAs and TAs. However, SSAs were more likely to occur in the right side of the colon than the other 3 histological types (p < 0.001). All mixed SSA-HPs were right-sided. Furthermore, HPs were significantly smaller in comparison with SSAs, TSAs, and TAs (p < 0.001).

Table 1.

Clinicopathologic Characteristics of Patients

	NC	HP	SSA	TSA	TA
	(n = 10)	(n = 32)	(n = 28)	(n = 37)	(n = 51)
Gender
Female	6	19	11	19	28
Male	4	13	17	18	23
Age (year)	46.6	45.5	48.6	48.6	51.1
(range)	(27-67)	(21-77)	(29-68)	(20-66)	(29-80)
Location
Right location	3	8	22a	6	13
Left location	7	24	6a	31	38
Average size (mm)	NA	5.1b	7.5	8.5	8.6
(range)		(4-7)	(4-12)	(5-12)	(6-14)
Histologic dysplasia
LGD	NA	NA	NA	30	37
HGD				7	14

NC, normal colonic tissue; HP, hyperplastic polyp; SSA, sessile serrated adenoma; TSA, traditional serrated adenoma; TA, traditional adenoma; LGD, low-grade dysplasia; HGD, high-grade dysplasia; NA, not applicable.

^aChi-square test among five groups, p < 0.001.

^bOne-way ANOVA among five groups, p < 0.001.

Immunolabeling for E-Cadherin, α-Catenin, and β-Catenin in NCs and TAs

Expression pattern for E-cadherin, α-catenin, and β-catenin in five histological types is summarized in Table 2. Expression of E-cadherin, α-catenin, and β-catenin was clearly evident at the cell–cell boundaries of all 10 normal colonic tissues and normal mucosa adjacent to tumors. However, membranous expression of E-cadherin and α-catenin was only seen in 33.3% (17/51) and 27.5% (14/51) of all investigated TAs, respectively (Table 2, Fig. 2).

Table 2.

Membranous Expression of E-Cadherin, α-Catenin, and β-Catenin in Colonic Polys

	NC, n (%)	HP, n (%)	SSA, n (%)	TSA, n (%)	TA, n (%)
E-Cadherin
Normal	10 (100)	31 (96.9)	25 (89.3)	29 (78.4)	17 (33.3)a
Abnormal	0 (0)	1 (3.1)	3 (10.7)	8 (21.6)	34 (66.7)
α-Catenin
Normal	10 (100)	30 (93.8)	25 (89.3)	25 (67.6)	14 (27.5)a
Abnormal	1 (0)	2 (6.2)	3 (10.7)	12 (32.4)	37 (72.5)
β-Catenin
Normal	10 (100)	25 (78.1)	21 (75.0)	25 (67.6)	15 (29.4)a
Abnormal	0 (0)	7 (21.9)	7 (25.0)	12 (32.4)	36 (70.6)

NC, normal colonic tissue; HP, hyperplastic polyp; SSA, sessile serrated adenoma; TSA, traditional serrated adenoma; TA, traditional adenoma.

^aPearson chi-Square, p < 0.001.

Figure 2.

Representative immunohistochemical staining patterns for E-cadherin, α-catenin, and β-catenin in a TA showing: normal membranous expression of E-cadherin, α-catenin and β-catenin in the normal adjacent tissue (white arrows), and absence of E-cadherin and α-catenin expression in adenomatous epithelium (black arrows). Abnormal nuclear expression of β-catenin was seen in adenomatous epithelium (black arrow). Scale bars = 100 µm, 200 × magnification.

Membranous expression of β-catenin was demonstrated in 29.4% (15/51) of TAs. Cytoplasmic accumulation and widespread or focal nuclear expression of β-catenin was observed in 60.8% (31/51) and 19.6% (10/51) of all TAs, respectively (Fig. 2). Mesenchymal tissue surrounding the epithelial cells did not express E-cadherin or any of the catenins.

Immunolabeling for E-Cadherin, α-Catenin, and β-Catenin in HPs, SSAs and TSAs

Membranous expression of E-cadherin was observed in 96.9% (31/32) of HPs, 89.3% (25/28) of SSAs, and 78.4% (29/37) of TSAs, significantly higher than that of TAs (33.3%, p < 0.001). Membranous expression of α-catenin was seen in 93.8% (30/32) of HPs, 89.3% (25/28) of SSAs, and 67.6% (25/37) of TSAs, significantly higher than that of TAs (27.5%, p < 0.001) (Table 2, Figs. 3–6).

Figure 3.

Expression of E-cadherin, α-catenin, and β-catenin in a hyperplastic polyp (HP). Expression of all three components was confined to the cell borders of epithelial cells, with no expression in the surrounding mesenchymal cells. Scale bars = 100 µm, 200 × magnification.

Figure 6.

Positive membranous immunohistochemical staining of E-cadherin, α-catenin, and β-catenin in a pure traditional serrated adenoma (TSA). Scale bars = 100 µm, 200 × magnification.

Figure 4.

In a sessile serrated adenoma (SSA), E-cadherin, α-catenin, and β-catenin were expressed in the cell membrane. Scale bars = 100 µm, 200 × magnification.

Figure 5.

Traditional serrated adenoma (TSA) (lower part, T) with foci of sessile serrated adenoma (SSA) (upper part, S). Retained membranous distribution of E-cadherin, α-catenin, and β-catenin staining was seen in both the SSA part and TSA part. MP, mixed polyp. Scale bars = 100 µm, 200 × magnification.

Furthermore, normal membranous immunoreactivity of β-catenin was found in 78.1% (25/32) of HPs, 75.0% (21/28) of SSAs, and 67.6% (25/37) of TSAs, significantly higher than that in TAs (29.4%, p < 0.001) (Table 2). Widespread or focal nuclear staining for β-catenin was observed in 19.6% (10/51) of TAs, while none of the HPs, SSAs, or TSAs showed nuclear labeling for β-catenin (p < 0.001). In addition, cytoplasmic accumulation of β-catenin was only demonstrated in 3.1% (1/32) of HPs, 3.6% (1/28) of SSAs, and 21.6% (8 /37) of TSAs, significantly lower than that in TAs (60.8%, p < 0.001) (Figs. 3–6).

Remarkably, membranous co-expression of E-cadherin, α-catenin, and β-catenin in HPs, SSAs, and TSAs was more frequent than that in TAs (Table 3).

Table 3.

Co-expression Pattern of E-Cadherin, α-Catenin, and β-Catenin in Colonic Polys

	Normal Expression of all Molecules, n (%)	Reduced Expression of 1 or 2 Molecules, n (%)	Reduced Expression of all Molecules, n (%)
NC	10 (100)	0 (0)	0 (0)
HP	22 (68.8)	10 (31.3)	0 (0)
SSA	17 (60.7)	11 (39.3)	0 (0)
TSA	14 (37.8)	21 (56.8)	2 (5.4)
TA	4 (7.8)a	31 (60.8)	16 (31.4)a

NC, normal colonic tissue; HP, hyperplastic polyp; SSA, sessile serrated adenoma; TSA, traditional serrated adenoma; TA, traditional adenoma.

^aChi-square test among five groups, p < 0.001.

Discussion

Colorectal cancer (CRC) is one of the most common malignant tumors in the world, but the way of cancer progression is still unclear. There is increasing evidence that CRC evolves through a number of pathways (Jass et al. 2002). Recent studies increasingly support the existence of an alternative route for colorectal carcinogenesis through serrated neoplastic pathway (Goldstein 2006; Grady 2007; Kambara et al. 2004). Although the concept of SA has been generally accepted, a firm molecular basis for SAs has been lacking.

It’s known that Wnt signaling pathway plays an important role in colorectal carcinogenesis (Schneikert and Behrens 2007). However, the role of Wnt signaling pathway in SAs is still conflicting. In the present study, membranous location of β-catenin was found in most HPs, SSAs, and TSAs, but not in most TAs. The frequency of cytoplasmic or nuclear location for β-catenin in SAs was significantly lower than that in TAs. These findings suggest that Wnt signaling pathway is unlikely to have a major role in the serrated pathway, as opposed to its role in the traditional adenoma-adenocarcinoma sequence.

However, previous studies observed nuclear β-catenin accumulation in 9/22 (41%), 35/54 (67%), and 6/16 (38%) of SSAs (Sandmeier et al. 2009; Wu et al. 2008; Yachida et al. 2009). These results are different from those of our present study, and suggest a role for Wnt pathway in the serrated neoplastic pathway. Surprisingly, studies by Wu et al and Yachida et al found nuclear β-catenin accumulation in 0/19 and 0/12 of HPs, respectively, contrasting to 6/12 (50%) of HPs in study by Sandmeier et al. Many reasons could explain these differing results with each other and with ours, including lack of diagnostic standard for different types of serrated polyps, intrinsic tumor heterogeneity, small sample size, differences in antigen retrieval, and staining procedures used by each laboratory. Therefore, further studies are needed before a definite conclusion can be made about the role of Wnt pathway in serrated pathway.

β-catenin plays a critical structural role in cadherin-based adhesions and is also the central in the Wnt signaling cascade. Cells can regulate whether β-catenin is used independently or competitively in cell adhesion and nuclear signaling, so that the two processes are tightly coordinated and interrelated (Gottardi and Gumbiner 2004). Wnt-induced β-catenin conformational changes favor assembly into transcription complexes, while α-catenin associated β-catenin seems to favor adhesion (Gottardi and Gumbiner 2004; Harris and Peifer 2005). Interestingly, in the present study, most HPs, SSAs, and TSAs showed retained membranous staining of E-cadherin, α-catenin, and β-catenin, indicating that β-catenin tends to assemble into adhesion complex and mediate adhesion in serrated pathway. This finding also provides another evidence that SAs and TAs are distinct entities, each with their own characteristic spectrum of molecular changes.

The switch of β-catenin in adhesion or Wnt signaling relies on the regulation of distinct molecular forms of β-catenin with different binding properties, including posttranslational modification (phosphorylation) or conformational changes (closed C-terminus ) in β-catenin (Gottardi and Gumbiner 2004; Harris and Peifer 2005). Thus, it remains to be tested whether there is a conformational change in β-catenin that favors the adhesive function as opposed to a function in Wnt signal transduction in the carcinogenesis of serrated pathway.

In addition, it’s reported that E-cadherin, α-catenin, and β-catenin associated with the actin cytoskeleton to determine tissue organization and morphogenesis (Gumbiner 2000). This prompts us to hypothesize that intact immunolocalization (along the lateral cell membrane) of adhesive molecules may contribute to the typical serrated architecture of serrated polyps, thus suggesting a new characteristic feature of this type of polyps.