INTRODUCTION

Matrix metalloproteinases (MMPs), a family of zinc-dependent endopeptidases, degrade the basement membrane and extracellular matrix, facilitating cell migration, tumor invasion, and metastasis (1, 2, 3, 4, 5). There are at least 20 human MMPs, divided into the collagenases, gelatinases, stromelysins, and membrane-type MMPs (MT-MMPs; 1, 2, 3, 4, 5). Tissue inhibitors of metalloproteinases (TIMPs) are the major endogenous regulators of MMPs and consist of four homologous members (TIMP1–4; 6, 7, 8). TIMPs are multifunctional proteins that inhibit cell invasion in vitro and tumorigenesis and metastasis in vivo (6). Although each TIMP appears capable of inhibiting several MMPs, these proteins exhibit preferential inhibitory capacity; for example, TIMPs1 and 2 selectively inhibit MMP9 and 2, respectively (9).

Increased expression of MMPs has been associated with poor prognosis and shortened patient survival in a variety of malignancies including carcinomas of the esophagus (10), stomach (11), colon (12), breast (13), pancreas (14), lung (15), kidney (16), and ovary (17). TIMP expression has been associated with both tumor suppressor or antimetastatic effects and tumor-promoting effects in selected cancers (18, 19, 20). MMP and TIMP expression in prostate cancer has been recently reviewed (21). Both MMPs and TIMPs have been characterized in prostate cancer cell lines (22, 23, 24, 25) and clinical samples from prostate cancer patients (26, 27, 28, 29, 30, 31, 32, 33), with conflicting results. Similarly, serum levels of circulating MMPs and TIMPs have shown variable capability of predicting disease progression (34, 35, 36, 37). The aim of the current study was to evaluate the immunohistochemical expression of MMP2 and TIMP2 in prostate cancer and determine whether the expression of these markers correlates with prognostic variables, including patient survival.

MATERIALS AND METHODS

Patients and Specimens

One hundred thirty-eight randomly selected prostatic adenocarcinomas (PACs) treated by radical retropubic prostatectomy obtained from the files of the Albany Medical Center Hospital between 1987 and 1997 were included in this study. All hematoxylin and eosin–stained slides from each case were reviewed, and tumors were graded according to the Gleason system (38) and staged according to TNM criteria (39). Multiple blocks were identified based on the presence of adequate tumor and the representative nature of the overall grade. For statistical evaluations, tumors with Gleason scores of 6 or lower were considered as low grade, and tumors with Gleason scores of 7 or higher were considered as high grade. Statistical analysis was also performed using a three-tier scheme isolating tumors with Gleason score of 7. Serum PSA levels as measured by the Hybritech Tandem method (Hybritech) were obtained from review of the patients’ medical records. Postoperative PSA of >0.4 ng/mL on two consecutive occasions after prostatectomy was considered as biochemical evidence of disease recurrence.

Immunohistochemistry

To analyze for the expression of MMP2 and of TIMP2 proteins, contiguous 4-μm sections were cut from a single block of formalin-fixed, paraffin-embedded tissue randomly chosen from those initially identified; sections were placed on charged slides. After deparaffinization, primary antibody incubation was performed by an automated system (Ventana Medical Systems, Tucson, AZ) for MMP2 and manually for TIMP2. Pertinent details regarding antibodies and staining procedure are summarized in Table 1. The remainder of the staining procedure included incubation with a biotinylated anti-mouse secondary antibody, diaminobenzidine substrate, and hematoxylin counterstain and was performed on the Ventana ES automated immunohistochemistry system. Negative-control slides were incubated with isotype-matched immunoglobulin in parallel with each batch of staining to confirm the specificity of the antibodies.

TABLE 1 Antibodies and Immunohistochemical Procedure
Full size table

Staining Interpretation

Staining results were interpreted without prior knowledge of clinical and pathologic parameters by two observers using a consensus method. For all markers, both the intensity of staining and approximate percentage of positive tumor cells were considered in the semiquantitative assessment, as previously published (40, 41). Briefly, the distribution of positive staining in the tumors was graded as focal (≤10%), regional (11–50%), and diffuse (>50%). The staining intensity was subjectively scored as weak, moderate, or intense. Staining patterns of moderate diffuse, moderate regional, intense regional, and intense diffuse were considered as increased expression of each protein.

Quantitative DNA Analysis

Quantitative analysis of DNA content was determined for each case using 5-μm tissue sections stained by the Feulgen reaction and evaluated by the CAS 200 image analyzer (Tripath Corp., Burlington, NC), as previously described (42, 43).

Statistical Analysis

Statistical comparisons were performed using Stata software (Stata Corp, College Station, TX). Correlation between protein expression and pathologic variables was performed using the χ2 univariate analysis. Survival curves for all univariate analyses were assessed using the Kaplan-Meier method. Overall survival was defined as the interval between surgery and postsurgical biochemical disease recurrence. Multivariate analyses of clinicopathologic parameters, including survival, were performed using the Cox proportional hazards model. The level of significance was set at .05.

RESULTS

Of the 138 PACs, there were 75 (54%) low-grade and 63 (46%) high-grade tumors. At prostatectomy, there were 78 (57%) organ-confined tumors (Stages I and II) and 60 (43%) advanced-stage (Stages III and IV) cancers. Of the 77 cases tested for total DNA content, 52 (68%) were diploid, and 25 (32%) were nondiploid. A total of 131/138 (95%) had sequential serum PSA follow-up information available. Of these 131 patients, 50(38%) had biochemical postsurgical disease recurrence.

Immunohistochemistry and Statistical Analysis

Immunostaining pattern for both proteins was cytoplasmic, with tumor cells showing moderate to intense positivity, as opposed to relatively weaker expression in the benign elements, which served as internal control in each case. One hundred five of 138 (76%) PACs expressed MMP2, and 113/138 (82%) expressed TIMP2. There was an overall significant coexpression of MMP2 and TIMP2 in 94/138 (68%) PACs (P = .01; Fig. 1). The co-expression of MMP2 and TIMP2 correlated with advanced tumor stage (P = .05) and reached near-significance as a univariate predictor of disease recurrence (P = .07). TIMP2 expression individually correlated with advanced tumor stage (P = .04; Fig. 2) and reached near significance with disease recurrence (P = .06; Fig. 3). MMP2 expression was also more frequent in the PACs that recurred, although this value did not reach statistical significance (P = .07). On univariate analysis, neither MMP2 nor TIMP2 expression correlated with tumor grade (using either the two- or three-tier scheme) or DNA ploidy.

FIGURE 1
figure 1

An example of the same case of prostatic adenocarcinoma showing co-expression of MMP2 (A) and TIMP2 (B) proteins. (3,3′-diaminobenzidine; hematoxylin, 200 ×).

Full size image
FIGURE 2
figure 2

A Stage 3 prostatic adenocarcinoma with perineural invasion expressing TIMP2 protein (A) in comparison to its absence in tumor glands of a Stage 2 cancer (B). (3,3′-diaminobenzidine; hematoxylin, 200 ×).

Full size image
FIGURE 3
figure 3

Kaplan-Meier estimates showing correlation between TIMP2 protein expression and postsurgical disease recurrence (P = .05).

Full size image

On multivariate analysis, only tumor stage (P = .009) and DNA ploidy status (P = .03) independently predicted disease recurrence.

DISCUSSION

MMP expression has been reported to be low or undetectable in most benign elements but is substantially increased in a majority of human malignancies (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21). Analysis of both primary and metastatic tumors has shown increased relative MMP expression at the metastatic site, supporting a role in tumor migration and spread (44). Additionally, increased MMP levels have been reported in the plasma and urine of patients with a variety of advanced malignancies (45). Cancer outcome studies have also shown that increased expression of MMPs is associated with shortened patient survival (10, 12, 46). Aberrant expression of MMPs in prostate cancer was first described using in situ hybridization in 1991 (31). MMP7 expression has been linked to prostate cancer pathologic stage and incidence of metastasis (33). Using both Northern analysis and in situ hybridization, Still and co-workers (27) linked increased MMP2 and TIMP2 to high tumor grade and advanced tumor stage of the disease. Increased MMP2 expression has also been associated with high tumor Gleason score (28). Finally, increased MMP expression also has been implicated in the development of prostate cancer, as evidenced by increased levels found in carcinomas versus benign prostatic hypertrophy and prostatic intraepithelial neoplasia (47, 48).

Serum measurements of MMPs in prostate cancer have yielded conflicting correlations with disease outcome. Several studies have found a correlation between circulating MMPs (MMP1,2,3) and circulating TIMPs (TIMP1,3) and advanced or progressive disease (34, 36); others have failed to confirm this association (35).

In the present study, increased immunohistochemical co-expression of MMP2 and TIMP2 was associated with advanced tumor stage and reached near-significance as a predictor of disease recurrence. TIMP2 expression correlated with tumor grade and predicted disease recurrence on univariate analysis but was not an independent predictor when tumor stage and DNA ploidy status were included in the multivariate analysis model.

At the time of their discovery, TIMPs were considered to be tumor suppressor proteins. Recombinant TIMP2 was shown to inhibit invasion of HT 1080 fibrosarcoma cells in vitro (49). Increased TIMP expression has been associated with decreased tumor growth, invasiveness, and metastasis in a variety of prostate cancer and non–prostate cancer cell lines (22, 23, 24, 25, 49, 50, 51, 52). However, the results of the current study, demonstrating a poor prognostic significance in prostate cancer for increased TIMP2 expression, are contrary to the original tumor suppressor role hypothesized for TIMPs and more in line with recent evidence documenting a multifunctional complex role for TIMPs. Nemith et al. (53) described the growth-promoting abilities of TIMP2 in several human cell types, including fibroblasts, keratocytes, lymphocytes, and stem cells. Increased TIMP1 and TIMP2 mRNA levels have been correlated with tumor stage, lymph node metastasis, and shortened survival in patients with carcinomas of colon (18), breast (19), and bladder (54). Our findings of the poor prognostic role of increased TIMP2 expression in prostate cancer also concur with the data of Kugler et al. (16), which demonstrated a correlation of increased TIMP2 levels with aggressive phenotype in renal cell carcinoma.

Although the paradoxical positive effect of TIMP in tumor progression is not completely understood, the tumor-promoting activity may be due either to proteolytic degradation of ECM or direct influence on cell survival and growth. TIMP2 is reported to regulate matrix degradation, acting through a membrane type MMP (MT1-MMP; 55, 56). MT1-MMP is a key enzyme in tumor angiogenesis and metastasis, hydrolyzes a variety of ECM components, and is a physiologic activator of pro-MMP2 (57). TIMP2 forms a complex with MT1-MMP and pro-MMP2 on the cell surface, promoting hydrolysis of pro-MMP2 to its active form (MMP2) and resulting in degradation of ECM. Also, formation of this complex decreases the autocatalysis of MT1-MMP, resulting in increased levels of its active form. It also has been reported that some TIMPs can directly affect cell growth and survival, independent of their actions on MMPs. Stimulation of cell growth by TIMPs is thought to be mediated by c-AMP–dependent activation of protein kinase A (58) and increased tyrosine phosphorylation (59). Cell survival is prolonged by the TIMP1-mediated upregulation of anti-apoptotic protein bcl-XL and by decreased NF Kappa B activity (60). Several additional factors that may play key roles in the TIMP promotion of cancer progression include the following: local TIMP concentration, cellular distribution, association with pro-MMPs, and presence of TIMP receptors (61, 62).

In view of their important role in tumor invasion and metastasis, inhibitors of MMP activity have been investigated as a method of preventing or decreasing tumor spread. Clinical trials involving batimastat (British Biotech), a potent, broad-based inhibitor of MMPs 1, 2, 3, and 9 (63), and marimastat (British Biotech), a second-generation, water-soluble synthetic MMP inhibitor, have been associated with clinical responses in pancreatic, pulmonary, ovarian, and mammary carcinomas (63). In preclinical studies of prostate cancer, anti-tumor effects of MMP inhibitors (doxycycline and chemically modified tetracyclines), both in vitro and in vivo, have been reported (21).

In conclusion, co-expression of MMP2 and TIMP2 proteins bears prognostic significance in patients with prostate cancer and supports a potential therapeutic role for synthetic MMP inhibitors. The paradoxical poor prognostic significance of TIMP2 expression warrants further investigation into the complex MMP–TIMP interactions and into the role of TIMPs in tumor evolution and spread.