Coexpression of Actin-Related Protein 2 and Wiskott-Aldrich Syndrome Family Verproline-Homologous Protein 2 in Adenocarcinoma of the Lung

Patients and Methods

Patients and tumors. Between 1993 and 1998, 338 patients underwent resection of adenocarcinoma of the lung at Tokyo Medical University Hospital, 115 of them were selected as subjects of this study on the basis of the following criteria: pathologic stage IA to IIIA, pneumonectomy or lobectomy with systemic lymph node dissection, and >60 months of follow-up of patients who did not experience tumor recurrence. One patient with a history of previous cancer, four patients who had other cancers of the lung whose a histologic type was different from the primary lung adenocarcinoma, and three patients who died of causes other than lung cancer were excluded from the analysis. All clinical data were collected by reviewing the patients' medical charts. The archived blocks of formalin-fixed, paraffin-embedded tissue from these patients were retrieved from the Department of Diagnostic Pathology.

Disease-free survival was defined as the period from the date of surgery until the date when tumor recurrence was detected, or, if not detected, until the date of the most recent medical examination. Tumor recurrence was diagnosed when multiple lesions in the lung or other organs were detected by X-ray examination, computed tomography, or gallium scintigraphy or a single new lesion was confirmed as to be a recurrence based on histologic examination. Signed informed consent to use their specimens was obtained from all patients in accordance with the institutional guidelines.

Table 1 shows the clinicopathologic characteristics of the 115 patients, 51 of whom developed recurrent disease and 36 of whom died of it. The median follow-up period was 60.4 months (1-132 months). Pathologic staging of the primary tumor was done according to a system based on the tumor-node-metastasis classification (19). The WHO criteria were used for histologic classification (20). Regarding adjuvant chemotherapy, 38 of 77 patients with completely resected pathologic stage I received orally a combination of uracil and tegafur (250 mg tegafur/m² body surface area/d) for 2 years. Other 33 patients in stage I did not receive any chemotherapy. In other pathologic stages (stage IIA, IIB, and IIIA), main regimen was cisplatin/vindesine chemotherapy (from two to four courses of cisplatin 70-100 mg/m², i.v. day 1; vindesine 3 mg/m², i.v. days 1 and 8) with or without oral uracil and tegafur therapy.

Immunohistochemistry. One or two representative blocks of each tumor were selected for immunohistochemical analysis of Arp2 and WAVE2. After cutting 4-μm-thick sections from all paraffin blocks and placing them on silane-coated slides, mirror sections, comprising a pair of consecutive slices, were prepared to examine cells for colocalization of Arp2 and WAVE2. The sections were placed on the slides so that their adjacent surfaces faced upwards, and appeared as reversed images of each other, the same as an object and its image in a mirror. The cut sections were dried overnight at 40°C, and then deparaffinized in xylene and rehydrated through a graded ethanol series. Endogenous peroxidase activity was blocked by immersing the slides in 0.3% hydrogen peroxide in methanol for 20 minutes. The slides were then placed in 1 mmol/L EDTA (pH 8.0) and heated in an autoclave at 110°C for 10 minutes for antigen retrieval. After cooling, the pairs of mirror sections were incubated overnight at room temperature with the primary antibodies, one member of each pair with a goat polyclonal anti-human Arp2 antibody, at a dilution of 1:50 (k-15, Santa Cruz Biotechnology, Santa Cruz, CA), and the other member of each pair with a goat polyclonal anti-human WAVE2 antibody, at a dilution of 1:50 (c-14, Santa Cruz Biotechnology).

After incubation, the slides were washed with 0.05 mol/L Tris buffer (pH 7.6) and reacted for 30 minutes with a secondary rabbit anti-goat immunoglobulin antibody [included in the Simple Stain MAX-PO(G) kit, Nichirei, Tokyo, Japan]. This antibody is labeled with a polymer prepared by combining amino acid polymers with peroxidase and a secondary antibody that has been reduced to the Fab′ fragment, and thus this method does not require an avidin-biotin reaction, which may cause nonspecific staining. Peroxidase staining was done with a solution of 3,3′-diaminobenzidine tetrahydrochrolide in Tris buffer [0.05 mol/L (pH 7.6)] containing 0.006% hydrogen peroxide. Mayer's hematoxylin was used as a counterstain.

Each pair of mirror section was compared carefully whether Arp2 and WAVE2 staining reveal the same localization in the tumor and the same intracytoplasmic distribution within the tumor cell. If their localization and intracytoplasmic distribution is clearly identical in >1% of cancer cells, the case is judged to be positive for coexpression of Arp2 and WAVE2 irrespective of localization pattern in the tumor, or intensity of intracytoplasmic staining.

Statistical methods. Disease-free survival was analyzed by the Kaplan-Meier method, and the statistical significance of differences was determined by log-rank test. Correlations between the patients' clinicopathologic features and staining results were analyzed by the χ² test. Multivariate analysis was done using Cox's proportional hazard regression model. We considered differences to be significant at the 0.05 level or below.

Results

Immunohistochemistry of Arp2 and WAVE2 in mirror sections. Immunoreactivity for both Arp2 and WAVE2 was evident in the cytoplasm of the adenocarcinoma of the lung cells. The intracytoplasmic staining pattern varied from dotted to granular or homogeneous. The Arp2-positive cells and WAVE2-positive cells were scattered or focally accumulated within the tumor tissue. Comparison of mirror sections separately stained for Arp2 and WAVE2 showed that the staining pattern and distribution of the cells that were positive for each antigen were identical in the tissue of some cancers. Figure 1 shows that the staining pattern and distribution of Arp2-positve cells were identical to those of the WAVE2-positive cells in two cases (A and B, C and D). Similar results were obtained in 78 (67.8%) of the 115 tumors that we judged to be positive for coexpression of Arp2 and WAVE2. WAVE2 was expressed in eight tumors that did not express Arp2 (Table 2 ), and neither Arp2 nor WAVE2 was expressed in 29 tumors.

No coexpression was detected in noncancerous tissue.

Relationship between clinicopathologic features and coexpression of Arp2 and WAVE2. The correlations between coexpression of Arp2 and WAVE2 and clinicopathologic features are shown in Table 3 . The frequency of coexpression was 46.8% (22 of 47) in stage IA and significantly lower than in the other stages (P < 0.0001). The frequency of coexpression in bronchioloalveolar carcinoma (BAC) was low (14.3%, 3 of 21), and the difference between the frequency in BAC and non-BAC was significant (P < 0.0001). Coexpression was detected more frequently in cases with lymph node metastasis (N_1-3; 86.1%, 31 of 36) than in N₀ cases (59.5%, 47 of 79; P = 0.0046). However, there were no significant correlations between coexpression and T category or vascular or lymphatic invasion.

Relationship between patient outcome and coexpression of Arp2 and WAVE2.Figure 2A shows the Kaplan-Meier disease-free survival curve of patients with tumors coexpressing of Arp2 and WAVE2 and not coexpressing them, and the difference was significant (P < 0.0001). There was also a significant difference between them in overall survival curve (P < 0.0001; Fig. 2B).

Multivariate Cox regression analysis to identify independent factors affecting patient outcome showed that coexpression of Arp2 and WAVE2 (P = 0.0039), N factor (P < 0.0001), and T factor (P < 0.0053) had significant prognostic value for disease-free survival (Table 4 ).

We analyzed the differences in the Kaplan-Meier disease-free survival curves between positive and negative groups at each stage, and found that the disease-free survival curves significantly differed between the positive and negative groups only in stage IA (Fig. 2C; P < 0.0110).

Discussion

Immunohistochemical staining of mirror sections in this study showed expression of both Arp2 and WAVE2 in the same adenocarcinoma of the lung cells, and the clinical data indicated that coexpression of Arp2 and WAVE2 was an independent risk factor for tumor recurrence. The state of coexpression of Arp2 and WAVE2 would determine whether strong adjuvant chemotherapy should be done especially at stage IA; coexpression was also shown to affect the overall survival rate of the 115 patients and was significantly correlated with lymph node metastasis. These results suggest that coexpression of Arp2 and WAVE2 is involved in a mechanism that augments the malignant potential of the tumor cells.

To analyze the relationship between coexpression and mild chemotherapy, which is mainly composed of oral uracil and tegafur in this study, we divided 77 cases of stage I into two cases with positive coexpression and with negative coexpression. In 45 positive coexpression cases, patients who received chemotherapy had better disease-free survival than those who received no chemotherapy although the difference was not significant (P = 0.163). In 32 patients with negative coexpression tumor, there was no significant difference between patients who received chemotherapy and those who did not (P = 0.946), and both patients had good disease-free survival irrespective of adjuvant chemotherapy. Further clinical examination is needed but we suppose that mild adjuvant chemotherapy in this study has a small effect on the disease-free survival compared with coexpression of Arp2 and WAVE2.

It has been established that WAVE2 binds to the Arp2/3 complex in vivo (14), thus activating it to initiate polymerization of actin to form lamellipodia. We think that WAVE2 binds to the Arp2/3 complex in lung adenocarcinoma cells coexpressing both proteins and that the binding is a cellular signal to form lamellipodia, which drive cell migration. Thus, adenocarcinomas of the lung coexpressing Arp2 and WAVE2 are thought to acquire the ability to migrate into surrounding tissue, which is one of the essential steps required for invasion and metastasis.

The mechanism of directional cell movement that involves interaction between Arp2/3 complex and WAVE2 is common to mammalian cells and not unique to malignant cells. However, because our study clearly showed the absence of coexpression of Arp2 and WAVE2 in normal bronchoalveolar cells, we speculate that coexpression of Arp2 and WAVE2 is enhanced in cancer cells and increases their potential for developing lamellipodia and motility.

We investigated whether the number of cells coexpressing Arp2 and WAVE2 in a tumor affects patient outcome but failed to find any evidence of a correlation (data not shown). We also divided 78 coexpression-positive cases into three groups by following two variables: distribution patterns in the tumor (scattered, focal, and intermediate) or intracytoplasmic staining patterns (dotted, granular, and homogenous). But we failed to find any evidence of correlation between patients' prognosis and immunohistochemical staining patterns, such as distribution or intracytoplasmic staining pattern (data not shown). We think that one reason for failing to find a correlation may have been technical problems with immunostaining, such as a decrease in the antigenicity of Arp2 and/or WAVE2 during formalin fixation. Another reason may have been misleading results because of an imbalance in the distribution of coexpressing cells in the tumors. Still another possibility is that binding between Arp2/3 complex and WAVE2 may have some other function besides lamellipodium formation, such as in cell signaling. Although further investigation seems necessary to determine the significance of the interaction between Arp2/3 complex and WAVE2, such an interaction is required for at least some adenocarcinoma of the lung cells to acquire invasive or metastatic potential. Recent genetic studies of k-ras and p53 in both BAC and atypical adenomatous hyperplasia suggest a multistep mechanism of carcinogenesis of adenocarcinoma of the lung (21–23) in which atypical adenomatous hyperplasia changes to BAC, which, in turn, progresses to non-BAC. A clinicopathologic study of small adenocarcinomas has shown that the boundary between a low and high malignant character lies between BAC and non-BAC (24). We have shown a significantly lower rate of Arp2/WAVE2 coexpression in BAC than in non-BAC, and the frequency of cancer cells exhibiting coexpression in cases of mixed-subtype adenocarcinoma was much lower in the area of BAC portion than in the non-BAC portion. Therefore, it is conceivable that Arp2/WAVE2 coexpression is a phenotype of tumor cells that acquire invasive potential.

Lung cancer is currently one of the most common causes of cancer deaths in advanced nations. Because many patients experience tumor recurrence after surgery, even when the tumor is small or lymph node metastasis is absent, adjuvant chemotherapy is required for the majority of patients with early-stage lung adenocarcinoma. Coexpression of Arp2 and WAVE2 is expected to be a useful clinical indicator of high-risk group of early lung adenocarcinoma patients, and may be valuable for decision-making as to whether chemotherapy is required after surgery. Avoidance of unnecessary chemotherapy would improve the patients' quality of life after surgery.