This article is mentioned in:

Introduction

Colorectal adenocarcinoma (CRA) is a common type of malignant tumor. The 5-year survival rate of CRA is ~65%, and the tumor stage, lymph node (LN) status, tumor grade and lymphatic and venous invasion are critical morphological prognostic factors (1,2). Although advances have been made while studying the molecular basis of this disease, the spectrum of genes that reveal altered expression in CRA, as well as their role in the disease, remain unclear (3,4). Therefore, more sensitive CRA biomarkers that are capable of predicting prognosis and guiding effective targeting therapy are required.

The ezrin gene is a member of the ezrin-radixin-moesin (ERM) cytoskeleton-associated protein family and is involved in a wide variety of cellular processes. It is one of the components of cell-surface structures involved in cell adhesion to the extracellular matrix, and has been implicated in membrane-cytoskeleton interactions (5,6). The ezrin protein correlates with tumor invasiveness, metastasis and clinical prognosis in numerous types of human cancer, including CRA (7–9). The molecular characteristics of the ezrin protein may be important during tumor progression (10); however, the clinical significance of these characteristics in human cancer requires clarification. The aim of the present study was to investigate the correlation between ezrin expression, clinicopathological parameters and the prognosis of CRA patients. Our data suggested that cytoplasmic ezrin expression is associated with tumor metastasis and poor survival in CRA patients, and may be a useful marker for predicting disease progression and prognosis in CRA patients.

Materials and methods

Samples

In total, 186 cases of routinely processed and diagnosed CRA with strict follow-up were randomly selected from patients who underwent surgery between 2004 and 2007 in the Liaoning and Shanghai regions of China. Pathological parameters, including age, gender, grade, the presence of nodal metastasis, clinical stage and survival data, were carefully reviewed in all cases. The ages of the patients ranged from 25–82 years, with a mean age of 59.35 years. The male:female ratio was 119:67. Staging was performed according to the TNM and the International Federation of Gynecology and Obstetrics (FIGO) classifications of colonic and rectal carcinomas; according to FIGO, 97 cases were classified as stages I–IIA (early stage). According to the Union for International Cancer Control (UICC) criteria 7th Edition and the WHO classification (Pathology and Genetics of Tumors of the Digestive System), 89 cases were classified as stages IIB–IV (advanced stage). In addition, 99 cases were defined as well-differentiated and 87 cases were defined as moderately or poorly differentiated. All adjacent normal colorectal mucosal tissues from the cancer resection margin were included and none of the patients had received chemotherapy prior to surgery. All 186 patients were followed-up for 5 years or until they succumbed to the disease. At the end of follow-up, 139 patients had survived. Informed consent was obtained from each patient prior to commencing the study and the research protocols were approved by the Ethics Committee of the University Hospitals.

Immunohistochemistry for ezrin in paraffin-embedded tissues

Tissue sections (4 μm) were prepared on silane-coated slides (Sigma, St. Louis, MO, USA). Immunostaining kits were purchased from DakoCytomaton Inc. (Glostrup, Denmark) and Nichirei Inc. (Tokyo, Japan). Tissue sections were deparaffinized, rehydrated and incubated with 3% H2O2 in methanol for 15 min at room temperature in order to eliminate endogenous peroxidase activity. The antigen was retrieved at 95ºC for 20 min by placing the slides into a 0.01 M sodium citrate buffer (pH 6.0). The slides were subsequently incubated with a primary ezrin antibody (1:50, BD Biosciences Pharmingen, San Diego, CA, USA) at 4°C overnight. Following incubation at room temperature for 30 min with a biotinylated secondary antibody, the slides were incubated with a streptavidin-peroxidase complex (BD Biosciences Pharmingen) at room temperature for 30 min. Immunostaining was developed using chromogen, 3,3′-diaminobenzidine and counterstained with Mayer’s hematoxylin. We used mouse IgG isotope controls, which demonstrated negative staining. Furthermore, the positive tissue sections were processed omitting the primary antibody (mouse anti-ezrin) as negative controls.

Analysis and interpretation of staining

Immunoreactivity was independently evaluated by two researchers who were blinded to the patient outcome. The evaluation was based on the extent and intensity of the staining (12). The ezrin staining intensity was scored as follows: 0, negative; 1, weak; 2, moderate; and 3, strong. Staining extent was scored as follows: 0, 0%; 1, 1–25%; 2, 26–50%; 3, 51–75%; and 4, 76–100%, depending on the percentage of positively stained cells. The sum of the staining intensity and the staining extent scores was used as the final staining score. The specimens were divided into three groups according to their 5-year scores: 0–1, negative (−); 2–4, weakly positive (+); and 5–7, strongly positive (++).

Statistical analysis

Statistical analyses were performed using SPSS 17.0 (SPSS Inc., Chiacgo, IL, USA). The correlation between ezrin expression and clinicopathological characteristics was evaluated using the χ2 and Fisher’s exact tests. The survival rates following tumor removal were calculated using the Kaplan-Meier method and the difference in survival curves was analyzed using the log-rank test. Multivariate survival analysis was performed on all significant characteristics and was measured by univariate survival analysis with the Cox proportional hazard regression model. P<0.05 was considered to indicate a statistically significant difference.

Results

Expression of ezrin protein in CRA

The expression of ezrin protein revealed an immunohistochemical cytoplasmic staining pattern in CRA. The difference between the strongly positive rate (++) in the CRA (15.6%, 29/186) and adjacent normal mucosal tissues (25.3%, 47/186) was not deemed to be statistically significant (P=0.083); however, the positive (+ and ++) rate in CRA (61.3%, 114/186) was significantly lower than that found in the adjacent normal mucosal tissues (91.9%, 171/186; P=0.000; Table I and Fig. 1)

Figure 1 Immunohistochemical staining of ezrin in colorectal adenocarcinoma (CRA). (A) CRA with no ezrin expression; (B) CRA with weak ezrin expression; (C) CRA with marked ezrin expression; (D) Ezrin is not expressed in metastatic lymph nodes, but is potently expressed in the adjacent normal lymph node tissue. (Original magnification, ×100).

Table I Ezrin protein expression in CRA.

Table I

Ezrin protein expression in CRA.

		Ezrin (n)
Diagnosis	No.	−	+	++	Positive rate (%)	Strongly positive rate (%)
Carcinoma	186	72	85	29	61.3a	15.6
Normal mucosa	186	15	124	47	91.9	26.3

a P<0.01, compared with the normal mucosa group.

{ label (or @symbol) needed for fn[@id='tfn2-mmr-08-01-0061'] } CRA, colorectal adenocarcinoma.

Clinicopathological and prognostic significance of ezrin expression

To evaluate the role of ezrin protein in CRA progression, we analyzed the correlation between the expression of ezrin protein and major clinicopathological features. Table II and Fig. 3 demonstrate that the positive rate of ezrin expression in cases with a large tumor was 52.1% (37/71), which was significantly lower than in cases with a smaller tumor (67.0%, 77/115; P=0.044). The positive rates of ezrin expression in cases without serosal invasion (68.3%, 69/101) or LN metastasis (70.3%, 78/111) were significantly higher than in CRA cases with these factors (P=0.032 and P=0.002, respectively). The lymph node ratio (LNR) refers to the examination of the positive LN rate, and ezrin expression was identified to statistically correlate with the LNR. Cases with a high LNR (≥0.7) had a lower positive rate of ezrin expression (36.4%, 8/22) than cases with a low LNR (<0.7; 64.6%, 106/164; P=0.011). Finally, the positive rate of ezrin expression was 47.2% (42/89) in late-stage CRA, which was significantly lower than in early-stage cases (74.2%, 72/97; P=0.000). However, the differences among ezrin expression, age, gender and tumor grade were not statistically significant (P>0.05, respectively).

Figure 3 Kaplan-Meier analysis of the overall survival rate in 186 CRA patients with or without ezrin expression in relation to serosal invasion, tumor grade, lymph node ratio (LNR) status and tumor stage. The P-value was calculated by comparing all four groups (log rank test). (A) CRA cases with poor differentiation concomitant with the non-expression of ezrin were associated with the worst 5-year survival rate, significantly worse than CRA with poor differentiation only (P=0.000). (B) CRA with serosal invasion concomitant with the non-expression of ezrin had the lowest 5-year survival rate and a lower 5-year survival rate than CRA with serosal invasion only (P=0.000). (C) CRA cases with a high LNR concomitant with the non-expression of ezrin were associated with the worst 5-year survival rate, but not significantly worse than CRA with a high LNR only (P=0.156). (D) Late-stage CRA concomitant with the non-expression of ezrin was associated with the worst 5-year survival rate, which was significantly worse than only late-stage CRA (P=0.000). +, positive; −, negative; CRA, colorectal adenocarcinoma.

Table II Univariate analysis of ezrin protein expression and various risk factors in 186 patients with colorectal adenocarcinoma.

Table II

Univariate analysis of ezrin protein expression and various risk factors in 186 patients with colorectal adenocarcinoma.

		Ezrin expression, n (%)
Characteristic	No.	−	+ and ++	OR (95% CI)	P-value
Age (years)				0.230 (0.680–2.227)	0.297
>59	82	34 (41.5)	48 (58.5)
≤59	104	38 (36.5)	66 (63.5)
Gender				0.901 (0.488–1.663)	0.429
Male	119	45 (37.8)	74 (62.2)
Female	67	27 (40.3)	40 (59.7)
Tumor grade				0.585 (0.323–1.059)	0.057
Well	99	32 (32.3)	67 (67.7)
Moderate and poor	87	40 (46.0)	47 (54.0)
Tumor size				0.537 (0.293–0.985)	0.044a
≤5 cm	115	38 (33.0)	77 (67.0)
>5 cm	71	34 (47.9)	37 (52.1)
Serosal invasion				1.917 (1.054–3.484)	0.032a
Present	85	40 (47.1)	45 (52.9)
Absent	101	32 (31.7)	69 (68.3)
LN metastasis				2.561 (1.393–4.708)	0.002b
Present	75	39 (52.0)	36 (48.0)
Absent	111	33 (29.7)	78 (70.3)
LNR				3.198 (1.267–8.072)	0.011a
≥0.7	22	14 (63.6)	8 (36.4)
<0.7	164	58 (35.4)	106 (64.6)
Clinical stage				3.223 (1.740–5.971)	0.000b
I–II	97	25 (25.8)	72 (74.2)
IIIa–IV	89	47 (52.8)	42 (47.2)

{ label (or @symbol) needed for fn[@id='tfn3-mmr-08-01-0061'] } LN, lymph node; LNR, lymph node rate.

a P<0.05,

b P<0.01.

Ezrin expression level combined with tumor grade, serosal invasion status and stage affects the prognosis of CRA

To confirm the role of ezrin expression in CRA progression, we analyzed the 5-year survival rate of 186 CRA patients using the Kaplan-Meier method and discovered that CRA patients with no ezrin expression had a lower 5-year survival rate than those with positive ezrin expression (P=0.000; Fig. 2). Furthermore, we analyzed the correlation between other factors (age, gender, tumor size, grade, serosal invasion, LN, LNR and tumor stage) and the 5-year survival rate in CRA and discovered that tumor grade, LNR, serosal invasion and tumor stage were key factors associated with 5-year survival rate (P=0.016, P=0.004, P=0.015 and P=0.006, respectively). Further combination analysis revealed that CRA with poor differentiation combined with no ezrin expression had the lowest 5-year survival rate, and this was significantly lower than that in cases with ezrin expression (Fig. 3A; P=0.000). Furthermore, CRA without serosal invasion combined with positive ezrin expression had the highest 5-year survival rate, and CRA with serosal invasion combined with no ezrin expression had a lower 5-year survival rate than cases with a positive level of ezrin expression (Fig. 3B; P=0.000). The 5-year survival rate of CRA patients with a high LNR did not correlate with the cases, whether it was accompanied by ezrin expression or not (Fig. 3C; P=0.156). However, the 5-year survival rate of CRA patients with a low LNR did correlate with the cases, whether it was accompanied by ezrin expression or not (Fig. 3C; P=0.008). Tumor stage is the most important histological prognostic factor in CRA (11) and we revealed that CRA patients with a high stage (IIIa–IV) and no ezrin expression had a significantly lower 5-year survival rate than those patients with ezrin expression (Fig. 3D; P=0.002). Therefore, the non-expression of ezrin may be a poor prognostic marker for CRA with poor differentiation, serosal invasion, high LNR and at a late stage.

Figure 2 Kaplan-Meier analysis of the overall survival rate in 186 CRA patients in relation to ezrin protein expression. Patients with no ezrin expression had a significantly lower 5-year survival rate (P=0.000). +, positive; −, negative; CRA, colorectal adenocarcinoma.

Non-expression of ezrin is an independent prognostic factor in CRA, as determined by the Cox proportional hazard regression model

Table III shows the univariate and multivariate analyses performed using the Cox proportional hazards model. The LNR (HR, 0.589; 95% CI, 0.369–0.939; P=0.026) and tumor stage (HR, 0.655; 95% CI, 0.487–0.880; P=0.005) were demonstrated to be independent prognostic factors for survival in CRA. Notably, the non-expression of ezrin emerged as a significant independent prognostic factor in CRA (HR, 0.562; 95% CI, 0.404–0.783; P=0.001).

Table III Univariate and multivariate survival analyses (Cox regression model) of various factors in 186 patients with CRA.

Table III

Univariate and multivariate survival analyses (Cox regression model) of various factors in 186 patients with CRA.

Factor	B	SE	Wald	HR (95% CI)	P-value
Univariate survival analyses
Age (years)	−0.016	0.148	0.011	0.984 (0.737–1.315)	0.916
Gender	0.028	0.153	0.035	1.029 (0.763–1.388)	0.852
Tumor size (cm)	−0.156	0.151	1.068	0.856 (0.636–1.150)	0.301
Tumor grade	−0.120	0.147	0.663	0.887 (0.665–1.183)	0.416
Serosal invasion	−0.038	0.149	0.065	0.963 (0.719–1.289)	0.799
LN metastasis	−0.274	0.148	3.447	0.760 (0.569–1.015)	0.063
LNR	−0.622	0.228	7.464	0.537 (0.344–0.839)	0.006b
Clinical stage	−0.445	0.147	9.127	0.641 (0.480–0.855)	0.003b
Ezrin	−0.605	0.152	15.901	0.546 (0.406–0.735)	0.000b
Multivariate survival analyses
Age (years)	0.043	0.153	0.078	1.043 (0.774–1.408)	0.781
Gender	0.078	0.166	0.221	1.081 (0.781–1.498)	0.638
Tumor size (cm)	−0.076	0.165	0.213	0.927 (0.671–1.281)	0.644
Tumor grade	−0.115	0.150	0.580	0.892 (0.664–1.197)	0.446
Serosal invasion	0.246	0.164	2.230	1.278 (0.926–1.765)	0.135
LN metastasis	−0.276	0.154	3.232	0.759 (0.561–1.025)	0.072
LNR	−0.530	0.238	4.945	0.589 (0.369–0.939)	0.026a
Clinical stage	−0.424	0.151	7.862	0.655 (0.487–0.880)	0.005b
Ezrin	−0.576	0.169	11.628	0.562 (0.404–0.783)	0.001b

{ label (or @symbol) needed for fn[@id='tfn6-mmr-08-01-0061'] } CRA, colorectal adenocarcinoma; LN, Lymph node; LNR, Lymph node rate.

a P<0.05,

b P<0.01.

{ label (or @symbol) needed for fn[@id='tfn9-mmr-08-01-0061'] } B, coefficient; SE, standard error; Wald, Wald statistic; HR, hazard ratio; CI, confidence interval.

Discussion

The ezrin gene is located on chromosome 6q25.2-q26. The full length mRNA is 3166 bp, encoding a protein that consists of 585 amino acids. In 1981, Fehon et al(5) separated and purified the protein in the small intestinal epithelial cell brush border of a chicken. The ezrin gene belongs to the ERM family and addresses all types of epithelium, shares a homology with the amino terminal membrane-binding domain of erythrocyte band 4.1, and is also involved in membrane-cytoskeleton interactions (13). Ezrin is capable of interacting with several membrane proteins, including CD44, CD43 (14), intercellular adhesion molecule-1, intercellular adhesion molecule-2 and phosphatidylinositol-bisphosphate (15). Ezrin, moesin and radixin are often all expressed on the intramembrane. Ezrin may alter the relationship between the cell membrane and the cytoskeleton and promote the formation of microvilli and aphosphyses, which move cells. The main function of ezrin is to interact with p85, the regulatory subunit of PI3-kinase (PI3K), involved in determining the survival of the epithelial cells by activating the PI3K/Akt pathway (16). Subsequently, the regulation of adhesion, migration and invasion were observed and shown to be important to tumor development and progression (17).

Two theories exist with regard to the role of ezrin in tumor invasion and metastasis. One states that ezrin is the promotion factor for tumor invasion and metastasis and is overexpressed in a variety of invasive carcinoma tissues. In 2004, Yu et al(18) and Khanna et al(19) reported that ezrin is crucial for the metastasis of rhabdomyosarcoma and bone sarcoma in children; they transferred the Vil2 gene (ezrin coding gene) into low-transfer ability cell lines and discovered that the transfer capability of the tumor cell was greatly improved, and the creation of lung metastases was simple. This demonstrates that the overexpression of ezrin in tumor cells may provide the cell with a high transfer activity. Contrary to the upregulation of ezrin in tumor cells, another hypothesis is that ezrin is downregulated in tumor cells and inhibits tumor invasion and metastasis. Karmakar and Das (20) observed that in the human chorioepithelioma cell line JEG-3 cultivated by IL-IB, ezrin, E-cadherin and β2 serial protein expression levels were downregulated, CD44 expression was enhanced, adhesion between the JEG-3 cells and the matrix was increased, cell-cell adhesion abated and aggressiveness was enhanced. Hiscox and Jiang (21) cultured the colon cancer cell line in vitro and observed that a lack of ezrin decreased tumor cell aggregation, increased the separation phenomenon, broadened the intercellular space, formatted pseudopodia, increased the ability to move, decreased intercellular adhesion, increased matrix adhesion and enhanced aggressiveness. Our study suggests that ezrin expression is deregulated in CRA and that the non-expression of ezrin exerts a profound effect on cellular proliferation, cell adhesion, invasion and aggressiveness during cancer growth and metastasis.

Tumor staging is an important histological feature of CRA prognosis, but knowledge of its associated molecules is very limited. In this study, no ezrin expression was identified in a third of CRA cases, which correlated with the tumor size, serosal invasion status, LNR and the pathological stage. Additionally, the non-expression of ezrin in early-stage CRA was three times lower than in the late-stage cases. Furthermore, the cases had a lower 5-year survival rate, and more notably, late-stage CRA accompanied by ezrin non-expression resulted in a poorer survival rate than that in late-stage CRA only. These results indicate that tumor stage and the non-expression of ezrin may indicate the survival of patients. Additionally, tumor staging was found to be an independent prognostic factor, and we demonstrated that the non-expression of ezrin is associated with the poor prognosis of late-stage CRA.

According to the recommendations of the American Joint Committee on Cancer (AJCC) and the National Cancer Institute, ≥12 LNs should be examined in order to confirm the absence of nodal involvement in colorectal cancer patients (1,3), due to the fact that the number of LNs examined is a reflection of the aggressiveness of the surgical dissection and positive pathological identification (22). Additionally, the number of positive LNs examined in colorectal surgery may be associated with the outcome for the patient. Telian and Bilchik (23) reported that the LNR, rather than LN number, is an important positive prognostic factor for colorectal cancer. Wang et al(24) concluded that LNR is an independent survival predictor after adjusting for the age of the patient, tumor size, tumor grade, ethnicity, number of positive LNs and total number of LNs collected. Therefore, LNR ≥12 may be a better prognostic indicator than the number of positive LNs. Consistent with other reports, we confirmed that a high LNR (≥12) is an independent prognostic factor for patient survival in CRA. Moreover, CRA with a low LNR had 3-fold ezrin expression compared with a high LNR in CRA. Additionally, CRA with a decreased LNR without ezrin expression had the lowest 5-year survival rate, even though it was not statistically lower than a high LNR with ezrin expression.

In conclusion, we identified ezrin as a potential biomarker to evaluate the tumor progression and prognosis of CRA. The non-expression of ezrin was more commonly observed in cases with poor CRA prognostic factors, leading to late-stage tumors and short survival times. Ezrin has been suggested as a novel therapeutic method for selectively targeting cancer cells, particularly for late-stage CRA.

References