Low Expression of Claudin-1 and Presence of Poorly-differentiated Tumor Clusters Correlate with Poor Prognosis in Colorectal Cancer

Patients and Methods

Patients and specimens. A total of 344 stage II and III (TNM staging) colorectal cancer cases were reviewed. All patients underwent curative surgery at the Department of Surgical Oncology (First Department of Surgery) of the Osaka City University Graduate School of Medicine (Osaka, Japan) between 2003 and 2008. The patients' characteristics are summarized in Table I. The mean age of the patients at initial surgery was 66.8 years (range: 26-90), and 197 males and 147 females were included in this study. Of the tumors, 200 were of the colon and 144 were of the rectum. Patients with familial adenomatous polyposis or inflammatory bowel disease, or those who had died within 30 days of surgery were not included in the analysis. No patient received chemotherapy or radiotherapy before surgery. All patients were informed of the investigational nature of this study and provided written informed consent.

Pathological review. Slides of tumors stained with hematoxylin and eosin were examined by an experienced gastrointestinal pathologist. The depth of invasion (pT category), lymph node involvement (pN category), and pathological staging of all surgically-resected tumors were assessed according to the Union for International Cancer Control (UICC) TNM classification (19). Out of the 344 patients, 174 were diagnosed pathologically to have TNM stage II lesions and 170 had stage III lesions. The degree of tumor differentiation was defined based on the guidelines of the Japanese General Rules for Clinical and Pathologic Studies on Cancer of the Colon, Rectum and Anus (20). The grade of the poorly-differentiated component (POR) was assessed by using the tumor grading system based on the extent of the poorly-differentiated component (17, 18).

Follow-up. In 146 out of the 170 patients with stage III disease, adjuvant chemotherapy was given for at least six months. All patients were followed-up regularly by physical and blood examinations with mandatory screening by colonoscopy, ultrasound and/or computed tomography. The patients were observed for more than three years, and the mean duration of follow-up was 51.7 months for the survivors who were alive at the date of their last visit or who were checked by telephone.

Immunohistochemistry. Immunohistochemical staining for claudin-1 was performed on 4-μm paraffin sections of each colorectal cancer sample. Slides were stained with claudin-1 antibodies (polyclonal rabbit, 1:100, Zymed Laboratories South San Francisco, CA, USA). After deparaffinization in xylene and being rehydrated in a graded ethanol series, followed by rinsing with water, antigen retrieval was performed by microwaving the samples three times at 500 W for 5 min each in 10 mM sodium citrate (pH 6.0).

Following H₂O₂ and serum blocking, the slides were incubated with the primary antibody at room temperature overnight. The secondary antibody was a biotin-labeled goat anti-rabbit IgG (1:500, Nichirei, Tokyo, Japan). Detection was performed with the DAB kit (Histofine SAB-PO Kit; Nichirei, Tokyo, Japan) for 3 min. Slides were counterstained with hematoxylin, dehydrated, cleared and mounted on glass coverslips.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Claudin-1 expression in colorectal cancer (×400). a: Claudin-1 immunoreactivity is absent in the membrane of tumor cells in this case. b: Claudin-1 immunoreactivity is present in the membrane of tumor cells in this case.

Assessment of immunostaining. The immunostaining was classified as follows: 0, no immunostaining present; 1, <25% of cells positive; 2, 25-50% of cells positive; 3, 51-100% of cells positive. In the evaluation, only membrane-bound positivity was considered to be significant, and cytoplasmic or nuclear immunoreactivity was ignored (Figure 1).

In addition, the staining intensity was ignored. The results were analyzed by two observers who had no knowledge of the clinicopathological data at the time of the analysis.

Assessment of the extent of the POR. Firstly, the largest POR area in the tumor was determined by surveying the entire tumor with a low-magnification objective lens (step 1). Next, the area chosen as the POR was examined to determine whether it completely filled the microscopic field of a ×400 lens (step 2). If it filled the field, the tumor was classified as grade 3. When the microscopic field of the ×400 lens was not occupied by the POR area alone, we focused on cancer clusters without gland formation composed of five or more cancer cells, which were mainly observed in the selected area. The number of solid cancer nests (SCNs) in the microscopic field of a ×40 lens was counted in the area where the SCNs were most numerous (step 3). Tumors with fewer than 10 SCNs were classified as grade 1, and those with 10 or more SCNs were classified as grade 2 (Figure 2) (17, 18).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

The poorly-differentiated component (POR) in colorectal cancer a: POR grade 1. The number of solid cancer nests in the microscopic field of a ×40 lens was fewer than 10. b: POR grade 2. The number of solid cancer nests in the microscopic field of a ×40 lens was 10 or more. c: POR grade 3. The largest POR filled the field of a ×400 lens.

Statistical analysis. For the statistical evaluations, the immunohistochemical staining intensities were grouped as negative (which included scores of 0 and 1) and positive (which included scores of 2 and 3). The statistical analyses were carried out using SPSS, version 11.0.1.0, software program for Windows (SPSS Japan Inc., Tokyo, Japan). Associations between the claudin-1 expression and clinicopathological parameters were analyzed using Fisher's exact and χ² tests. The univariate survival analysis was carried out according to the Kaplan–Meier method. Differences in survival curves were assessed with the log-rank test. Multivariate analysis was performed according to the Cox regression model. Two-tailed p-values ≤0.05 were considered to be statistically significant.