Matrix Metalloproteinase 9 and the Epidermal Growth Factor Signal Pathway in Operable Non-Small Cell Lung Cancer¹

Lung cancer is the commonest cause of cancer death in Europe and the United States. Despite advances in surgery, chemotherapy, and radiotherapy in the last two decades, the death rate has remained little changed. Subjects with the same stage of disease can have markedly different rates of disease progression in NSCLC.³ This suggests that these tumors, despite having the same histological subtype, are biologically different. Immunohistochemical investigation may reveal other prognostic markers and suggest subgroups that could benefit from adjuvant therapy after surgical resection.

The EGFR is a member of the erb/HER type I tyrosine kinase receptor family involved in cell proliferation and differentiation. Epidermal growth factor, amphiregulin, and transforming growth factor-α bind to EGFR leading to receptor dimerization. Ligand binding activates the intracellular tyrosine kinase domain and the transmission of growth regulatory signals triggering DNA synthesis (1). EGFR expression is generally low in normal bronchial epithelium and is enhanced in preneoplastic and neoplastic lesions (2). EGFR has been shown to be associated with poor prognosis in NSCLC in some studies (3) but not in others (4).

Malignant cells either secrete proteinases or acquire proteinase activity from host stromal cells or inflammatory cells to allow them to break through collagenous protein barriers. MMP-9 (gelatinase B)degrades collagen type IV, the major component of the basement membrane (5). Overexpression of MMP-9 will facilitate metastatic spread. MMP levels relate to invasive and metastatic potential in vitro (6), whereas synthetic MMP inhibitors decrease the spread of cancer in vivo (7). Increased expression of MMP-9 mRNA and MMP-9 protein has been demonstrated in many solid tumors (8), including NSCLC (9, 10, 11). Higher levels of MMP-9 mRNA have been found in stage III NSCLC compared to stages I and II (10). The expression of either MMP-9 or MMP-2 (gelatinase A) confers a worse prognosis in early stage adenocarcinoma of the lung (11).

The regulatory pathway for MMPs is complex. A recent study of head and neck squamous cell lines has implicated the EGFR signal pathway in the up-regulation of MMP-9 (12). The aim of this study is to evaluate the relationship between MMP-9 and EGFR expression in NSCLC and to assess the impact of these factors on survival.

This is a retrospective study of 169 patients with stage I–IIIa NSCLC who underwent surgical excision in Glenfield Hospital (Leicester,United Kingdom) between 1991 and 1996. Patients with a postoperative survival <60 days were excluded to remove the bias of perioperative death. Follow-up ranged from 24–108 months (median, 39.8 months). Eighth-nine (52.7%) subjects died from a recurrence of their primary lung cancer. One hundred eighteen cases (69.8%) were male, and the mean age of all subjects was 64.9 years (SD, 7.51; median,66; range, 42–78). The clinicopathological features of the specimens were assessed according to the WHO classification (13) and the tumor-node-metastasis staging system (14).

Four-μm-thick formalin-fixed, paraffin-embedded sections taken from the tumor periphery were mounted on silane-coated slides. Sections were dewaxed in xylene and rehydrated through graded alcohols. Antigen retrieval was achieved by pressure cooking slides for 2 min in 10 mm citric acid buffer at pH 6. Endogenous peroxidase activity was blocked by placing sections in 2% hydrogen peroxide for 30 min. Sections were rinsed in deionized water, and then Tris-buffered saline containing 0.1% BSA. To block nonspecific staining, slides were incubated in 20% normal rabbit serum for 10 min. Sections were incubated overnight at 4°C with the primary antibody. The antibodies used were MMP-9 mouse monoclonal clone 56–2A4, which recognizes both latent and active MMP-9 at a dilution of 1:100 (Chemicon International, Ltd.), and EGFR mouse monoclonal clone EGFR.113, which recognizes the extracellular region of the EGFR molecule at a dilution of 1:20 (Novocastra Laboratories, Ltd.). Sections were washed in Tris-buffered saline and then incubated sequentially with biotinlyated rabbit antimouse IgG (DAKO Corp.) at a dilution of 1:400,followed by streptavidin combined in vitro with biotinylated horseradish peroxidase at a dilution of 1:1000 (DAKO Corp.). The reaction product was developed using diaminobenzidine tetrahydrochloride. Sections were counterstained with hematoxylin,dehydrated through graded alcohols, and mounted in resinous mountant. Known positive controls were included with each run, and negative controls had the primary antibody omitted.

The extent and pattern of reactivity for each antibody was recorded by two observers. The extent of expression was scored 0 for no staining,<20%, 20–50%, and 50–100%. A similar semiquantitative scale of 0,+, ++, or +++ was used to assess the intensity of staining in comparison to a known positive control. This scoring system was applied to membranous and cytoplasmic staining for EGFR and to tumor cell and stromal reactivity for MMP-9. Cases were called positive if staining intensity was ++ or +++ over at least 20% of the tumor. Sections were analyzed in a blinded fashion, and the results of the immunohistochemistry, tumor status, and patient outcome correlated subsequently.

Statistical analysis was performed using the SPSS software system (SPSS for Windows, Version 9.0). The χ² test was used to analyze the associations between categorical clinicopathological variables. Cancer-specific survival curves were plotted using the Kaplan-Meier method, and the log-rank test was used to assess the statistical significance of differences between groups. The joint effects of covariables that were significant at a level of 0.25 in univariate analysis were further examined via Cox regression using a forward selection procedure. The 0.05 level of significance was used for entering or removing a covariable from this model.

The clinicopathological findings are listed in Table 1. The stage of the tumor was prognostic(P = 0.0001). Tumor spread to nodes was associated with poor prognosis (P = 0.0009), and increasing nodal status was more significant (P < 0.0001). No other clinicopathological finding, including age, sex, grade, or histological subtype, was associated with outcome.

Reactivity for EGFR was present in 94 of 169 (56%) tumors. Membranous immunopositivity with or without cytoplasmic staining was seen in 55 of 169 (33%) cases (Fig. 1,A). Cytoplasmic expression without membranous staining was seen in 39 of 169 (23%) cases (Fig. 1,B). Membranous, cytoplasmic and overall EGFR expression were not associated with outcome(P = 0.13, 0.99, and 0.17, respectively; Fig. 2). Large cell and squamous cell carcinomas expressed EGFR more frequently than did adenocarcinomas (P < 0.0001), and expression was more common in the elderly (P = 0.02). Membranous EGFR expression was more frequent in squamous cell carcinomas(P = 0.008), in the elderly (P = 0.02),in poorly differentiated tumors (P = 0.03), and in males (P = 0.05; Table 2).

Reactivity for MMP-9 was observed in stromal fibroblasts, infiltrating macrophages, and localized to the cytoplasm of tumor cells and was frequently more intense at the infiltrating edge of the tumor (Fig. 1, C–E). MMP-9 tumor cell expression was recorded in 88 of 169(52%) cases and conferred a poor prognosis (P =0.0023; Fig. 3,A). MMP-9 tumor cell expression was associated with poor outcome in stage II disease(P = 0.03) but did not reach significance in either stage I (P = 0.08) or stage IIIa (P =0.87) disease. MMP-9 stromal expression was seen in 79 of 169 (47%)cases and was not prognostic (P = 0.86; Fig. 3,B). Stromal staining was found more frequently in large cell and squamous cell carcinoma histological subtypes(P = 0.003). Stromal and tumor cell MMP-9 were frequently co-expressed (P = 0.015). There were no other associations with clinicopathological findings for either tumor cell or stromal MMP-9 expression (Table 3).

Tumor cell co-expression of MMP-9 and EGFR was found in 62 of 169(37%) cases (P < 0.0001). Membranous EGFR and tumor cell MMP-9 were also co-expressed in 38 of 169 (22%) cases(P = 0.002). There was no association between cytoplasmic EGFR and tumor cell MMP-9 expression (P =0.18). There was a trend toward co-expression for EGFR with stromal MMP-9 (P = 0.06). Tumor cell co-expression of EGFR and MMP-9 was associated with a poor outcome (P = 0.0001;Fig. 4,A), as were those with co-expression of membranous EGFR and MMP-9 (P = 0.0019;Fig. 4 B). Cytoplasmic EGFR and tumor cell MMP-9 co-expression was not associated with poor outcome. EGFR and MMP-9 tumor cell co-expression was associated with poor outcome in stage I disease (P = 0.002) but did not reach significance in either stage II (P = 0.09) or stage IIIa disease(P = 0.74).

Cox proportional hazards regression analysis was used to define biological markers with independent predictive value with respect to cancer-specific survival (Table 4). Nodal status (P = 0.0006) and tumor cell MMP-9 expression(P = 0.027) were the only significant independent prognostic factors.

MMPs are part of the proteolytic cascade that degrades the ECM and allows the migration of tumor and endothelial cells. In particular,MMP-9 is a gelatinase capable of forming gaps in the basement membrane to facilitate invasion and metastatic spread. Using paraffin-embedded NSCLC tissue sections, our study has demonstrated both stromal and cytoplasmic tumor cell MMP-9 expression. Tumor cell expression was associated with a poor prognosis on univariate (P =0.0023) and multivariate (P = 0.027) analysis,especially in early stage disease. This provides further evidence in support of an important role for proteases in the malignant process. The presence of MMP-9 confirms NSCLC as a potential target for a general inhibitor of metalloproteinases such as marimastat (15) or the more specific gelatinase inhibitor CT-1746 (16), both of which are currently undergoing clinical trials. Expression of MMP-9 alone or in conjunction with other MMPs may be developed as a prognostic marker for NSCLC.

Membranous and overall EGFR expression occurred more commonly in squamous cell carcinoma, and membranous EGFR expression was found more frequently in poorly differentiated tumors, findings that have been shown in previous studies (17). The pattern of EGFR immunostaining was not found to be prognostic in our study. There was a significant relationship between the presence of tumor cell MMP-9 reactivity and EGFR expression, both membranous and total reactivity(P = 0.002 and P < 0.0001,respectively). A similar relationship has recently been described in head and neck squamous cell lines (12). The EGFR ligands epidermal growth factor, amphiregulin, and transforming growth factor-α have been shown to induce the expression of MMPs (18, 19). Together, these findings suggest that the EGFR signaling pathway may contribute to the metastatic process by specifically up-regulating expression of MMP-9 and promoting tumor invasion. The precise mechanism involved is not understood; however, the potential pathways could involve mitogen-activated protein kinase (20), Ets, AP-1 (21), or the cell-cell adhesion molecule system of E-cadherin and β-catenin (22, 23).

In summary, we have found a strong association between EGFR and MMP-9 expression in NSCLC, which suggests that the EGFR signaling pathway may up-regulate MMP-9 expression. Further studies are required to evaluate the possible roles of mitogen-activated protein kinase and E-cadherin in this pathway. We also found MMP-9 expression in tumor cells confers a poor prognosis, especially when there is co-expression of EGFR. Proteases are novel targets for therapeutic intervention, and the use of metalloproteinase inhibitors in NSCLC should be further evaluated.