Lack of MMP-9 expression is a marker for poor prognosis in Dukes’ B colorectal cancer

Background Matrix metalloproteinases (MMPs) play a role in cancer progression by degrading extracellular matrix and basement membranes, assisting in tumour neovascularization and in supporting immune response in cancer. Methods We studied the prognostic value of immunohistochemical expression of MMP-2, MMP-8, and MMP-9 in a series of 619 colorectal cancer patients using tissue microarray specimens. Results Of the samples, 56% were positive for MMP-2, 78% for MMP-8, and 60% for MMP-9. MMP-9 associated with low WHO grade (p < 0.001). In univariate analysis of Dukes’ B tumours, MMP-9 negativity associated with poor survival (p = 0.018), and MMP-9 positivity was an independent prognostic marker in multivariate analysis of these tumours (p = 0.034). Conclusion Negative MMP-9 expression can predict poor prognosis in Dukes’ B colorectal tumours and may prove useful for identifying patients, who should be offered adjuvant treatment.


Background
Colorectal cancer (CRC) is the third most common malignancy in the world [1]. The most important prognostic factor in colorectal cancer is stage of disease at diagnosis. Other factors for poor prognosis are venous and lymphatic invasion, deficient healthy tissue margins at surgery, obstruction or perforation of the bowel, and poor differentiation at histology [2][3][4][5]. In addition to tumour-specific factors, also host responses, such as intra-or peritumoral inflammation and desmoplasia or reactive lymph nodes can suppress tumour spread and predict better outcome [6][7][8]. In cancer progression and spread, tumours must invade their surrounding tissues, basement membranes (BM), and extracellular matrix (ECM), must avoid the host's immunoresponse, and must ascertain their circulation by neovascularization. In all these phenomena, matrix metalloproteinases (MMPs) play an important role [9].
MMPs are zinc-dependent endopeptidases capable of degrading both BM and ECM proteins and extracellular adhesion molecules. MMP-9, also known as gelatinase B, is able to degrade collagens IV (the main component of BM) and V, gelatins, and elastin. MMP-2, also known as gelatinase A, is able to degrade the same substrates as MMP-9, but also collagens I, VII, X, fibronectin and procollagenase-3 [9]. MMP-8, also called collagenase-2, is able to degrade collagens I, II and III [10].
In this study, we studied the prognostic value of MMP-2, MMP-8 and MMP-9 immunoexpression in colorectal cancer.

Patients
Sample material came from 643 consecutive patients who underwent surgery for CRC at the Department of Surgery, Meilahti Hospital, Helsinki University Central Hospital, between 1982 and 1998. Of these, 18 patients were excluded on account of incorrect final diagnosis (9 patients) or synchronous multiple tumours (9 patients). Six cases were excluded because of insufficient archival surgical tissue specimens. Finally, 619 cases remained, 330 of them male. Tumour staging was performed according to the modified Dukes' classification [19], with 91 tumours classified as Dukes' stage A, 226 as stage B, 161 as stage C, and 145 as stage D. For patients' clinicopathological characteristics, see Table 1.
Clinical data were retrieved from patient records, and survival and cause of death data until March 2011 from the Population Register Centre of Finland, and Statistics Finland.
For MMP-9, we also stained tissue microarray (TMA) specimens from a validation series of 213 CRC patients treated between 1998-2001, 82 of them male; 31 tumours were classified as Dukes' stage A, 70 stage B, 69 stage C, and 41 stage D, with 201 of tumours being adenocarcinomas; with 7 being grade I, 161 grade II, 37 grade III, and 4 grade IV. Clinical data was retrieved, as for the main cohort, until October 2011.

Tissue samples and preparation of TMA blocks
Formalin-fixed and paraffin-embedded surgical tissue samples were collected from the archives of the Department of Pathology, University of Helsinki. Histopathologically representative regions of tumour specimens were defined and marked on H&E slides. Three cores from each tumour block were sampled with 1.0 mm punchers with a manual tissue microarrayer (Tissue Arrayer 1, Beecher Instruments Inc., Silver Spring, MD, USA). Three series of blocks were constructed, all including one sample from each patient. From each block, 4-μm slides were cut and processed for immunohistochemistry.
Immunohistochemistry of MMP-2, MMP-8, and MMP-9 The Scoring of immunostainings MMP-2, MMP-8, and MMP-9 immunostainings were scored by two independent investigators (S.K. and J.H.) without knowledge of clinical outcome. In case of different scores, the consensus score was determined. The interobserver variation was low, below 5%. The intensity of cytoplasmic staining in cancer cells was evaluated: strong intensity, scored as 3, moderate as 2, and weak as 1. Absence of positivity in all spots was scored as 0 ( Figure 1). Spots without cancer cells were excluded. For further analysis, the patients were divided into two groups, negative (0) versus positive (score 1-3).

Statistical analysis
Correlations between stainings and clinicopathological variables were assessed with the χ2 test or Fisher's exact test when applicable. The Mann-Whitney U-test was applied to determine the correlation between age and the immunoreactivity. Life-tables were calculated with the Kaplan-Meier method.
In the validation series, 5-year survival was 64.8% in MMP-9 positive patients compared to 63.5% in those MMP-9 negative, (p = 0.418). In subgroup analysis of Dukes' stages there were no significant differences in survival according to stage.

Discussion
It is clinically relevant to identify a prognostic marker in Dukes' B (stage II) disease. This stage is classified as local disease, but 20% will still die from this, with surgery curing 80%. We urgently need biomarkers to identify those 20% at risk in order to offer them adjuvant treatment.
Here we show that MMP-9 tumour expression is an independent prognostic factor in Dukes' B colorectal cancer. In the other stage groups, no significant difference in survival emerged. In Dukes' B, the difference was so great, that it affected analysis of the whole cohort. In Dukes' B and C of the validation series, we noticed poorer prognosis among patients with MMP-9 negative tumours, but the differences were not significant.
The overall survival was better in the validation series. The difference in survival may have several reasons. First, pathologic routines have changed and more lymph nodes are nowadays evaluated, which may lead to stage migration. Second, the surgical technique has changed, especially in rectal cancer treatment. Third, the more frequent use of adjuvant treatment might have improved survival.
Our results differ from those in a recent study of stage II CRC in which high MMP-9 expression associated with higher recurrence rate, shorter disease-free survival, and also shorter disease-specific survival, but association with disease-specific survival was not significant in multivariate analysis [21]. In CRC patients, high MMP-9 expression has been associated with liver metastasis [22]. We also analyzed high versus low immunoexpression, but found no association with survival (data not shown). Elevated MMP-9 mRNA levels have been associated with poor disease-free and overall survival [14]. On the other hand, some studies using immunohistochemistry have failed to show any correlation between MMP-9 and survival or clinicopathological parameters [16,23]. In other cancer forms, such as lung cancer, head and neck squamous cell carcinoma, and gastric cancer, MMP-9 associates with poor prognosis [24][25][26]. Interestingly, in early breast cancer, elevated MMP-9 associates with better prognosis [27].
In cancer, the host response, like the intra-or peritumoral inflammation reaction or desmoplasia, can predict better prognosis [6][7][8]. Theoretically, some matrixdegrading proteinases may play a defensive role by supporting the immune/inflammatory response. Stromal expression of MMP-9 inversely associates with liver metastasis and tumour infiltration in CRC [28]. Stromal MMP-9 positivity also inhibits metachronous haematogenic metastasis in Dukes' B and C colorectal cancer patients [29]. Here, we did not evaluate stromal MMP-9 expression, because we used tissue arrays with punches taken from cancerous regions, and our samples were unsuitable for reliable stroma evaluation. The TMA is not as reliable as analysis of whole tissue sections because of intra-tumoural heterogenity of immunohistochemical expression. However, the results of evaluation of TMAs and those of whole tissue sections have been shown to be in concordance [30]. Tumour expression of MMP-2 has been associated with poor prognosis in colorectal [11][12][13]31], gastric [25], ovarian [32], and breast cancer [33]. Here, we found no correlation between MMP-2 immunoexpression and clinicopathological variables or prognosis of CRC. Results similar to ours have emerged in CRC [34]. Thereby the role of MMP-2 in CRC is still controversial.
In breast cancer, MMP-8 changes metastatic potential in vitro [15], and in melanoma and lung cancer it inhibits metastasis formation by modulating cancer cell invasion and adhesion [14]. It was expected that MMP-8 might be a marker for improved prognosis in colorectal cancer, but we noticed no association between MMP-8 immunoexpression and survival.

Conclusion
The role of MMP-9 immunoexpression in colorectal cancer is dual; it plays a role in matrix degradation enabling tumour invasion, but it also seems to act as a supportive factor for hosts' defensive mechanisms against cancer spread. Here we show that immunoexpression of MMP-9 is a promising prognostic marker in Dukes' B (Stage II) CRC, the group of patients for whom we need new markers to identify those at risk who need adjuvant treatment.

Ethics approval
The study has been approved by the Surgical Ethics Committee of Helsinki University Central Hospital (Dnro HUS 226/E6/06) and National Supervisory Authority for Welfare and Health.