The combined expression of Semaphorin4D and PlexinB1 predicts disease recurrence in colorectal cancer

The study population included patients who underwent surgery for colorectal cancer between 2008 and 2011 at the Department of Surgical Oncology, Osaka City University Graduate School of Medicine, Japan. Patients who received preoperative chemotherapy were excluded from the analysis. The patients consisted of 124 males and 102 females, with a median age of 66 years (range: 21–91 years). The clinicopathological classification was determined according to the TNM classification of malignant tumors, as described by the International Union Against Cancer (UICC) [11]. The tumor stages of the patients were graded as follows: stage I (n = 58), stage II (n = 57), and stage III (n = 111).

All tissues were fixed in 10 % formalin immediately after surgical resection and 6-μm-thick specimens were embedded in paraffin. The immunohistochemical determination of the Sema4D and PlexinB1 levels in the colorectal cancer cells was carried out according to the manufacturer’s instructions. In brief, the slides were deparaffinized in xylene and hydrated in decreasing concentrations of ethyl alcohol. The sections were then deparaffinized and incubated with 3 % hydrogen peroxide in methanol for 15 min to block endogenous peroxidase activity. The tissues were subsequently heated for 10 min at 105 °C by autoclaving in Target Retrieval Solution (Dako, Carpinteria, CA, USA). The sections were then washed in phosphate-buffered saline (PBS) and incubated in 10 % normal rabbit serum for 10 min to reduce nonspecific antibody binding. The specimens were incubated with antibodies to Sema4D for 1 h at room-temperature and antibodies to PlexinB1 overnight at 4 °C. They were then washed twice with PBS. The primary antibodies used for the immunohistochemical detection of anti-Sema4D and anti-PlexinB1 were Rabbit polyclonal antibody to Sema4D (SIGMA-Aldrich Ltd, Poole, UK, HPA015662, 1:150) and Rabbit polyclonal antibody to PlexinB1 (SIGMA-Aldrich Ltd, Poole, UK, HPA040586, 1:100). The sections were incubated with biotinylated rabbit anti-goat immunoglobulin G for 10 min, then washed twice with PBS. The slides were then treated with peroxidase-conjugated streptavidin reagent for 5 min and washed twice with PBS. Finally, the slides were incubated with diaminobenzidine (DAB) kit (Histofine SAB-PO Kit; Nichirei, Tokyo, Japan) for 180 s for Sema4D antibodies, and 150 s for PlexinB1 antibodies, then counterstained with Mayer’s hematoxylin and mounted.

We counted the total number of infiltrating inflammatory cells in the tumor stroma in three independent high-power fields (×400) for each tissue sample. The positive Sema4D staining of inflammatory cells was observed in the tumor stroma (Fig. 1). We then calculated the percentage of Sema4D-positive cells among the total number of inflammatory cells. The specimens were then divided into three grades, according to the degree of positivity as follows: grade 1 (0–25 % positive), grade 2 (26–50 % positive) and grade 3 (51–100 % positive). For the statistical analyses, grades 1 and 2 were defined as negative, and grade 3 was defined as positive [9].

Positive PlexinB1 staining of the tumor gland was observed in the cytoplasm of cancer cells (Fig. 2). The staining intensity of epithelial tumor cells (in comparison to non-tumor cells) was scored as follows: 0 (no staining), 1 (weak staining), 2 (moderate staining) and 3 (strong staining). For the statistical analyses, scores of 0 or 1 were defined as negative, and a score of 2 or 3 was defined as positive [9].

The immunohistochemical staining of colorectal cancer tissue specimens revealed that 95 (42 %) and 105 (46.4 %) of the specimens were positive for Sema4D and Plexin B1, respectively. The expression of Sema4D was significantly related to stage, the depth of tumor invasion, lymph node metastasis, lymphatic invasion, and venous invasion (Table 1). Age, gender and histology did not significantly affect Sema4D expression. The expression of PlexinB1 was significantly related to stage, depth of tumor invasion, lymph node metastasis, lymphatic invasion, and venous invasion. There were no significant differences among the other factors.

Among 85 tumors which were found to be positive for Sema4D, 63(74.1 %) were positive for PlexinB1. On the other hand, among the 131 tumors which found to be negative for Sema4D, 42(32 %) were negative for PlexinB1. There was a significant correlation between the expression of Sema4D and PlexinB1 (p < 0.001, Table 2).

Sixty-three patients (27.9 %) expressed both Sema4D and PlexinB1. There were no significant differences in the amount of infiltrated inflammatory cells around the tumor stroma between the positive Sema4D/PlexinB1 group and the negative Sema4D/PlexinB1 group (p = 0.539). We examined the correlations between the positive expression of both Sema4D and PlexinB1 and the clinicopathological findings. The expression of both Sema4D and PlexinB1 was found to be correlated with all of the clinicopathological factors that were examined, with the exception of age, gender and histology (Table 3). In addition, we evaluated the presence of any correlation between the treatment as an adjuvant chemotherapy and the expression of Sema4D and PlexinB1. Adjuvant chemotherapy was administered to patients who belonged to the high risk groups with stage II and stage III disease, except for patients who rejected any further treatment and those who demonstrated a poor performance status. As a result, no significant differences were observed between the adjuvant chemotherapy and the combination of Sema4D and PlexinB1 expression at each stage (Table 4).

The patients of both Sema4D and PlexinB1 positive groups exhibited a worse prognosis compared to the others (p < 0.001, Fig. 3). In the univariate analysis, histology, the depth of tumor invasion, lymph node metastasis, lymphatic invasion, venous invasion and positivity for both Sema4D and PlexinB1 were found to be significantly associated with overall survival (Table 5). However, a multivariate analysis demonstrated that the depth of tumor invasion (HR 2.692, CI 1.022–9.317; P = 0.044), lymph node metastasis (HR 2.304, CI 1.138–5.069; P = 0.019), and the positive expression of both Sema4D and PlexinB1 (HR 1.681, CI 1.004–2.819; P = 0.047) were an independent risk factors for worse survival (Table 5). The recurrence rate was 46 % (29/63) in patients who positively expressed both Sema4D and PlexinB1, and 16.6 % (27/163) in the other patients; this amounted to a significant difference (p < 0.001). The relapse-free survival (RFS) of the patients who positively expressed both Sema4D and PlexinB1was significantly worse than that of other patients (p < 0.001, Fig. 4). In the univariate analysis, histology, the depth of tumor invasion, lymph node metastasis, lymphatic invasion, venous invasion and positivity for both Sema4D and PlexinB1 were found to be significantly associated with RFS (Table 6). However, a multivariate analysis demonstrated that lymph node metastasis (HR 2.783, CI 1.425–5.822; P = 0.002), and the positive expression of both Sema4D and PlexinB1 (HR 1.079, CI 1.013–2.868; P = 0.044) were an independent risk factors for worse survival (Table 6).