Prognostic value of cyclin B1 in patients with esophageal squamous cell carcinoma

Authors


Abstract

BACKGROUND

It has been reported that p53 regulates the G2-M checkpoint transition through cyclin B1, and it has been suggested that p53 plays an important role in the development and progression of various malignancies. The objective of the current study was to clarify the role of the cell cycle regulators, cyclin B1 and p53, in patients with esophageal squamous cell carcinoma (ESCC).

METHODS

Tissue samples from 71 patients with ESCC were included in the current study. Expression levels of cyclin B1 and p53 in samples of normal squamous epithelium, dysplasia, and tumor cells from patients with ESCC were analyzed by immunohistochemistry.

RESULTS

Several cells in the basement layer of normal epithelium expressed cyclin B1. The number of cyclin B1 positive cells tended to increase as the degree of dysplasia increased from low grade to high grade. More than 20% of tumor cells were cyclin B1 positive in 38 patients (49.3%). Several clinicopathologic parameters, including macroscopic configuration (P < 0.01), pathologic tumor status (P < 0.05), pathologic lymph node status (P < 0.001), pathologic metastatic status (P < 0.01), tumor stage (P < 0.0001), and invasion of lymphatic vessels (P < 0.05), were correlated with the overexpression of cyclin B1. Elevated expression levels of cyclin B1 also were correlated with a poor prognosis in patients with ESCC in univariate analysis (P < 0.0001) and multivariate analysis (P = 0.0135). In contrast, p53 expression exhibited no significant correlation with the level of cyclin B1 expression and was not associated with prognostic parameters in patients with ESCC.

CONCLUSIONS

These findings suggest that cyclin B1 is involved in the pathogenesis of carcinoma of the esophagus and that elevated levels of cyclin B1 expression, but not p53 expression, may indicate a poor prognosis for patients with ESCC. Cancer 2002;94:2874–81. © 2002 American Cancer Society.

DOI 10.1002/cncr.10542

Esophageal carcinoma is one of the common malignancies with a high mortality rate, especially in males and in less developed areas of the world. 1 To date, many studies have been performed in an attempt to clarify the underlying molecular mechanisms that determine the biologic behavior of esophageal carcinoma.

The p53 gene is a tumor suppressor gene that plays an important role in cell cycle regulation. However, despite investigation, the role played by p53 in the pathogenesis and progression of esophageal carcinoma remains unclear. 2–6p53 regulates the G2-M transition, a critical cell cycle checkpoint, and p53-deficient or p53-mutant cells undergo cell cycle arrest in the G2 phase.7, 8 Wild type p53 regulates Gadd45 protein, which inhibits the activity of the cdc2/cyclin B1 complex. Thus, the G2-M checkpoint that acts to prevent malignant transformation is controlled through p53 by regulation of intracellular cyclin B1 levels and activity.9–11

Recently, several studies reported that the level of cyclin B1 expression may be an important factor in the development of some neoplasms. It was reported that cyclin B1 overexpression had a prognostic influence in patients with early-stage nonsmall cell lung carcinoma and also was correlated with astrocytoma grade. 12, 13 The involvement of cyclin B1 in carcinogenesis also was supported by studies of patients with malignant and premalignant lesions of the breast and colon.14 In contrast, it is interesting to note that cyclin B1 overexpression in patients with either prostatic or gastric adenocarcinoma was of no prognostic significance.15, 16 There are scant data currently available regarding cyclin B1 in patients with esophageal carcinoma. The only series reported was by Murakami et al., and those authors concluded that a more comprehensive study with a large series of patients was required.17

The correlation of cyclin B1 expression with p34cdc2 or with the proliferation markers Ki-67 or proliferating cell nuclear antigen has been investigated in several neoplasms by immunohistochemistry. 18, 19 However, to our knowledge, there are no correlative immunohistochemical studies of cyclin B1 and p53 expression in any neoplasm. Therefore, in the current study, we investigated expression levels of cyclin B1 and p53 in patients with esophageal squamous cell carcinoma (ESCC) using immunohistochemistry to gain insight into the effect of cyclin B1 and p53 on the biologic behavior of esophageal tumors and to determine whether these expression levels have clinicopathologic implications.

MATERIALS AND METHODS

Patients

Seventy-one patients with ESCC, including 63 males and 8 females with a mean age of 63.8 years (range, 43–84 years) were included in the study. Patients underwent surgical treatment in the Department of Surgery II at Oita Medical University between January, 1990 and December, 1997. Surgical treatment consisted of resection of the esophagus with lymph node dissection without preoperative supplemental therapy. Patients who were without evidence of macroscopic residual carcinoma, except for distant metastasis, were included in the study.

Resected specimens were classified according to the International Union Against Cancer TNM classification system. 20 In addition, invasion of lymphatic and blood vessels was assessed microscopically.

Immunohistochemical Staining

Resected specimens were fixed in 10% buffered formalin for 24 hours and embedded in paraffin. To prevent bias, whenever possible, paraffin blocks for the current study were selected from the nonrepresentative area that did not show disease progression: for example, pT status or vessel invasion. Sections (4 μm) were placed on silane-coated slides. After deparaffinization and rehydration, endogenous peroxidase activity was blocked by incubation in 3% hydrogen peroxide for 20 minutes. Tissue sections were then autoclaved at 121 °C in 10 mM citrate buffer, pH 6.0, for 10 minutes for antigen retrieval. After cooling at room temperature for 30 minutes, the specimens were incubated with normal rabbit serum for 15 minutes at room temperature. They were then incubated with either anticyclin B1 monoclonal antibody (NCL-Cyclin B1; Novocastra, Newcastle, United Kingdom; 1:30 dilution) or anti-p53 monoclonal antibody (DO-7; Dako, Carpinteria, CA; 1:50 dilution) for 12 hours at 4 °C. Immunohistochemical staining was performed using a standard avidin-biotin-peroxidase complex technique with the SAB-PO (M) kit (Nichirei, Tokyo, Japan) using 3,3′-diaminobenzidine as the chromogen, and nuclei were counterstained with hematoxylin. Positive control tissues included samples of tonsil (cyclin B1) and gastric carcinoma (p53), and negative control tissues included the omission of the primary antibodies.

The number of p53 and cyclin B1 positive cells was determined by counting at least 1000 tumor cells for each in three randomly selected fields. Two independent observers who were blinded to clinical information evaluated the immunohistochemical staining. The mean score was used for the evaluation of immunohistochemical staining after the confirmation of the correlation between the scores counted by two observers. Immunostaining for p53 was scored according to the number of p53 positive tumor cells. A tumor was considered negative if < 10% of nuclei were p53 positive and was considered positive if > 10% of nuclei were p53 positive. With regard to cyclin B1 immunostaining, a cut-off value of 20% cyclin B1 positivity was chosen to indicate positive tumors according to a previous report 12; i.e., if > 20% of cells were positive for cyclin B1, then the tumor was considered positive.

Clinicopathologic Parameters

The parameters used in the current study included patient age and gender; location, configuration, size, and histologic grade (differentiation) of the tumor; TNM classification; and the degree of lymphatic or blood vessel invasion present. Concerning age and tumor size, the current series was divided in two groups by mean values. Tumor location was determined by the position of the main lesion; accordingly, the current series was divided into three groups with tumors located in the upper esophagus, middle esophagus, or lower esophagus. Tumors also were divided into three group based on tumor configuration: early lesions (flat spread), protuberant lesions, or ulcerating lesions.

Statistical Methods

The Pearson correlation coefficient test was employed to confirm the correlation between the scores of two observers. The correlation between cyclin B1 expression and clinicopathologic factors was analyzed using the Mann–Whitney U test or the chi-square test. The correlation between cyclin B1 and p53 expression was analyzed using the chi-square test and the Student t test.

The Kaplan–Meier method was used to measure the prognostic influences from the month after surgery. The differences were examined using the log-rank-test. The Cox proportional hazards model was used in the multivariate analysis. A P value < 0.05 was considered statistically significant.

RESULTS

Immunohistochemical Staining

In normal squamous epithelium, cyclin B1 was localized in the nucleus and was expressed only in several cells of the basal layer (Fig. 1a). By contrast, in dysplastic epithelium, cyclin B1 was localized in the cytoplasm, and the numbers of cyclin B1 positive cells increased compared with normal epithelium. Furthermore, areas of high-grade dysplasia tended to have greater numbers of cyclin B1 positive cells compared with areas of low-grade dysplasia (Fig. 1b,c).

Figure 1.

Photomicrographs showing immunohistochemical expression of cyclin B1 in normal squamous epithelium (a), in low-grade dysplasia (b), in high-grade dysplasia (c), and in squamous cell carcinoma of the esophagus (ESCC) (d). Cyclin B1 was localized in the nuclei in normal squamous epithelium. In contrast, cytoplasmic staining for cyclin B1 was observed mainly in samples of dysplasia and ESCC. (e) Photomicrograph showing immunohistochemical overexpression of p53 in ESCC. Nuclei of ESCC revealed a positive reaction for p53.

With respect to the ESCC samples, cyclin B1 expression was observed in the cytoplasm, and the positive rate for cyclin B1 positivity ranged from 1.5% to 73.1% (Fig. 1d). Thirty-five samples (49.3%) were cyclin B1 positive. Regarding p53, its expression was observed in the nucleus of either normal squamous epithelium, dysplasia, or ESCC, and 45 ESCC samples (63.4%) were positive in the current study (Fig. 1e).

Correlation with Clinicopathologic Parameters

Increased expression of cyclin B1 exhibited a significant correlation with the following clinicopathologic parameters: tumor configuration (P < 0.01), pT status (P < 0.05), pN status (P < 0.001), pM status (P < 0.01), tumor stage (P < 0.0001) and the presence of lymphatic vessel invasion (P < 0.05) (Table 1). Ulcerating esophageal tumors tended to exhibit cyclin B1 positivity, whereas early tumors were predominantly cyclin B1 negative. Regarding the pT status, we noted that the proportion of positive tumors increased with increasing depth of invasion. In addition, cyclin B1 positive tumors tended to exhibit metastatic spread; i.e., they were associated with pN and pM status. Finally, cyclin B1 positive tumors tended to invade lymphatic vessels.

Table 1. Immunohistochemical Status of Cyclin B1 in Patients with Esophageal Carcinoma
Clinicopathologic parameterNo. with cyclin B1 overexpression (%)P value
High (> 20%)Low (< 20%)
  • SCC; squamous cell carcinoma.

  • a

    Chi-square test.

  • b

    Fisher exact probability test.

  • c

    Mann–Whitney u test.

Age (yrs)
 < 6418 (25.4)21 (29.6)
 ≥ 6417 (23.9)15 (21.1)0.56a
Gender
 Male29 (40.8)34 (47.9)
 Female6 (8.5)2 (2.8)0.15b
Location
 Upper esophagus4 (5.6)4 (5.6)
 Middle esophagus17 (23.9)18 (25.4)
 Lower esophagus14 (19.7)14 (19.7)0.99a
Configuration
 Early type5 (7.0)17 (23.9)
 Protuberant type3 (4.2)1 (1.5)
 Ulcer formation type27 (38.0)18 (25.4)< 0.01a
Greatest dimension (mm)
 < 45.720 (28.2)23 (32.4)
 ≥ 45.715 (21.1)13 (18.3)0.56a
Grade of SCC differentiation
 Well9 (12.7)11 (15.5)
 Moderate12 (16.9)19 (26.8)
 Poor14 (19.7)6 (8.5)0.083a
pT status
 IS0 (0.0)3 (4.2)
 16 (8.5)14 (19.7)
 26 (8.5)4 (5.6)
 323 (32.4)15 (21.1)< 0.05c
pN status
 Positive27 (38.0)13 (18.3)
 Negative8 (11.3)23 (32.4)< 0.001a
pM status
 Positive13 (18.3)2 (2.8)
 Negative22 (31.0)34 (47.9)< 0.01b
Stage
 00 (0.0)3 (4.2)
 I3 (4.2)11 (15.5)
 II8 (11.3)12 (16.9)
 III11 (15.5)8 (11.3)
 IV13 (18.3)2 (2.8)< 0.0001c
Lymphatic invasion
 Positive28 (39.4)20 (28.2)
 Negative7 (9.9)16 (22.5)< 0.05a
Vascular invasion
 Positive12 (16.9)11 (15.5)
 Negative23 (32.4)25 (35.2)0.74a

Correlation between Cyclin B1 and p53 Expression

No significant correlation was revealed between cyclin B1 and p53 expression using the chi-square test (P = 0.17) (Table 2). In addition, the Student t test revealed no statistical differences between the p53 positive and p53 negative groups (29.1% + 20.3% vs. 24.9% ± 21.1%, respectively; P = 0.38).

Table 2. Correlation between p53 and Cyclin B1 Overexpression
GroupNo. with p53 overexpression (%)P value
> 10%< 10%
  • a

    Chi-square test.

  • b

    Student t test.

Tumors with cyclin B1 overexpression
 > 20%25 (35.2)10 (14.1)
 < 20%20 (28.2)16 (22.5)0.17a
Rate of positive cells for cyclin B129.1 ± 20.324.9 ± 21.10.41b

Univariate Analysis of Prognostic Parameters

Of the conventional histopathologic factors, we found that histologic grade of ESCC (P < 0.05), pT status (P < 0.01), pN status (P < 0.0001), pM status (P < 0.0001), tumor stage (P < 0.0001), and lymphatic vessel invasion (P < 0.0001) all exhibited prognostic implications. Blood vessel invasion (P = 0.073) tended to indicate a shorter survival but did not reach statistical significance. Increased cyclin B1 expression was correlated significantly with a poor prognosis in patients with esophageal carcinoma when the cut-off value for positivity was set at > 20% (P < 0.0001) (Fig. 2a). Indeed, cut-off values of > 30%, > 40%, and > 50% also were correlated significantly with patient outcome, with values of P < 0.0001. However, the 10% cut-off value did not reach statistical significance (P = 0.075). In contrast, p53 positivity had no prognostic implication in the current esophageal carcinoma series (P = 0.86) (Fig. 2b).

Figure 2.

Graphs illustrating survival after surgery. (a) The group of patients with high cyclin B1 expression exhibited a poor prognosis compared with the group of patients with low cyclin B1 expression (P < 0.0001). (b) No prognostic implication was revealed in the p53 immunohistochemical analysis (P = 0.86).

Multivariate Analysis of Prognostic Parameters

Conventional prognostic parameters and cyclin B1 and p53 expression also were analyzed using multivariate analysis. Distant metastasis (P = 0.0292), tumor stage (P = 0.0248), and cyclin B1 expression (P = 0.0135) all were independent prognostic parameters (Table 3).

Table 3. Multivariate Analysis of Prognostic Outcome
ParameterRisk ratio95% Confidence limitP value
pT (0–1 vs. 2–3)0.53010.2073–1.63280.246
pN (negative vs. positive)1.44310.5556–4.22560.4623
pM (negative vs. positive)2.39181.0947–5.18630.0292
Stage (0, I vs. II–III)10.44461.3244–219.14080.0248
Lympatic invasion2.3780.8188–8.79910.1168
Blood vessel invasion1.59750.7355–3.38660.2314
p53 Overexpression0.83350.4302–1.67320.5987
Cyclin B1 overexpression0.61810.4077–0.90750.0135

DISCUSSION

Reports indicate that p53 regulates the G2-M phase transition by decreasing cyclin B1 levels. 7–11 Our previous report suggested that p53 played an important role in the pathogenesis of esophageal carcinoma through cell cycle regulation.21 However, the role of cyclin B1 in carcinogenesis and various clinicopathologic parameters in patients with esophageal carcinoma remains unclear.

Cyclin B1 positive cells were confined predominantly to the basement layer in normal squamous epithelium. The first major finding of this study was that the number of cyclin B1 positive cells increased in proportion to the grade of dysplasia. However, it is interesting to note that not all areas of high-grade dysplasia exhibited a high frequency of immunohistochemical reactivity for cyclin B1. In some patients with hepatocellular carcinoma, the aberration of cyclin B1 may be related to malignant transformation. 22 It is possible that cyclin B1 may act as a malignant transforming factor in esophageal carcinogenesis comparable to its proposed role in the development of hepatocellular carcinoma.

With respect to the potential role of elevated levels of cyclin B1 expression in the progression of esophageal carcinoma, it is noteworthy that, in the current study, cyclin B1 expression was correlated with some conventional pathologic parameters, such as macroscopic tumor configuration, pT status, pN status, pM status, tumor stage, and lymphatic invasion. Murakami et al. reported that cyclin B1 was implicated in tumor invasion and venous invasion in patients with ESCC, 17 although our results are not in total agreement with that report, possibly because of differences in the cut-off value, which we set at 20% compared with their cut-off value of 5%. In fact, prognostic analysis in the current study revealed that there was no significant difference in prognosis if the cut-off value was set at 10%.

In contrast, elevated levels of cyclin B1 expression reflected a poor prognosis in patients with esophageal carcinoma in univariate and multivariate analysis when the cut-off value was set at 20%. It is interesting to note that our study produced results comparable to the results reported previously by Murakami et al., despite the difference in the cut-off value used. In the current study, elevated levels of cyclin B1 expression had a prognostic implication in patients with ESCC if the cut-off value was set at 20%, 30%, 40%, or 50%. This suggests that elevated levels of cyclin B1 expression may be an important prognostic parameter in patients with ESCC. Soria et al. also reported the clinical implications of cyclin B1 overexpression in patients with early-stage, nonsmall cell lung carcinoma, and especially in patients with squamous cell carcinoma, when the cut-off value was set at 15%. 20 However, it is interesting to note that the levels of cyclin B1 expression reportedly had no prognostic implication in patients with gastric or prostatic carcinoma. This may have been secondary to histologic differences between squamous cell carcinoma and adenocarcinoma. In patients with breast carcinoma, cyclin B1 reportedly regulates the G2-M checkpoint and plays an important role in carcinogenesis. The implications for both disease progression and clinical outcome of cyclin B1 expression may be peculiar to patients with squamous cell carcinoma.

Recently, it was reported that the p53 dependent G2-M cell cycle checkpoint is regulated by a reduction in cyclin B1 level mediated by the Gadd45 protein and that G2-M arrest was not observed in p53-deficient or p53-mutated cells in vitro. 7–11 In contrast, the correlation between p53 expression and clinical outcome in patients with esophageal carcinoma remains unclear. Therefore, we analyzed the expression of p53 as well as cyclin B1 in the current study.

In marked contrast to the prognostic implications of cyclin B1 overexpression, p53 overexpression had no implication in patients with esophageal carcinoma in the current study. Furthermore, we found no significant correlation between p53 and cyclin B1 overexpression in the chi-square test or the Student t test. Two possibilities were suggested based on these data: It is possible that cyclin B1 affects disease progression and prognosis directly and that p53 regulates disease progression indirectly, because the expression of cyclin B1 and cdc2 activated cdc2 kinase constitutively, with the cells passing through G2-M regardless of p53 expression. 11 Another possibility is that mutation of the p53 gene is more critical than elevated levels of immunohistochemical expression, because they are not related directly to each other.2, 4, 5 Because p53-mutated or p53-deficient cells do not undergo G2-M arrest, further analysis of the p53 mutation status will be required.7, 8

In conclusion, the current data support an important role for elevated levels of cyclin B1 expression in carcinogenesis, and analysis of cyclin B1 expression may be useful in the assessment of disease progression and prognostic outcome after patients undergo surgery for ESCC. We believe that these finding may provide patients with ESCC opportunities for adjuvant therapy in the future, because cyclin B1 may be a potential therapeutic target in patients with ESCC.

Acknowledgements

The authors thank Miss Naomi Kawano, Miss Yoko Iwata, and Miss Kaori Soe in the Department of Surgery II, Oita Medical University, for technical assistance with immunohistochemical staining.

Ancillary