Expression of genes encoding type IV collagen-degrading metalloproteinases and tissue inhibitors of metalloproteinases was detected in various human tumor cells, including colorectal cancer [18]. Transcripts for MMP-2 were more frequently expressed in mesenchymal tumor cells than in epithelial tumor cells. Their activity was regulated by TIMPs, which are induced mainly in stromal cells [18]. The immunohistochemical expression of MMP-2 in CRC tumors is significantly higher than that in adjacent normal tissues [13, 19]. However, little is known about the comparison of the levels of MMP-2 and TIMP-2 in the sera of CRC patients with an expression of MMP-2 and its inhibitor in colorectal cancer tissue. Therefore, the aim of the study was to examine the serum levels of MMP-2 and TIMP-2 in CRC patients and to compare the serum concentrations of those proteins with their presence and intensity expression in colorectal cancer cells, interstitial infiltrate inflammatory cell, and normal colorectal cells. We have also assessed the significance of MMP-2 and TIMP-2 as prognostic factors of CRC patients’ survival.
In our study, the serum levels of MMP-2 and TIMP-2 in CRC patients and healthy controls were assessed using an ELISA method. We revealed that the concentrations of both proteins tested were significantly lower in cancer patients than in healthy subjects, which is in agreement with our previous results [14] as well as with the investigation of Waas et al., who have shown that plasma pro-MMP-2 levels were lower in colorectal cancer patients than in healthy controls [20]. Moreover, in the study of Oberg et al., the serum levels of the MMP-2/TIMP-2 complexes were significantly lower in CRC patients as compared to healthy blood donors; however, the serum levels of free MMP-2 and total TIMP-2 were significantly higher in comparison with control group [21].
In the present paper, the concentrations of MMP-2 and TIMP-2 in the sera of CRC patients revealed neither any significant correlation with tumor stage, tumor size, and nodal involvement nor with the presence of distant metastases, although there was a tendency to decrease the levels of proteins tested in more advanced tumors. This tendency is opposite to the results obtained by of Angenete et al. who revealed that plasma levels of MMP-2 were higher in CRC patients with distant metastases [22]. However, our findings are in accordance with the study of Langenskiöld et al. who had found that plasma MMP-2 in CRC patients with T4 tumors was lower than that in T3 and T1 + T2 subgroups [23]. The decrease of serum MMP-2 and TIMP-2 in more advanced tumor stages might be caused by the formation of MMP–TIMP complexes in colorectal tumor progression. The differences between the results observed by cited authors and those obtained in our present study might be the effects of the complex function of MMP-2 and TIMP-2. It was suggested that MMP-2 inhibition occurs at high levels of TIMP-2, whereas low concentrations of the inhibitor are associated with the activation of MMP-2 [24, 25].
It was suggested that in cancer tissue, the enhanced expression of MMPs and TIMPs on the surface of inflammatory cells might be a result of a host response induced by tumors [26]. In the present paper, the tissue expression of MMP-2 and TIMP-2 was observed in cancer and interstitial inflammatory infiltrate cells as well as in normal colorectal epithelium. The expression of both proteins was higher in cancer cell than in normal tissue, where 15 % of normal colorectal cells expressed the presence of TIMP-2 and only 1 % of normal cells were MMP-2 positive, which is in line with the study of Gershtein et al., who revealed that the content of MMP-2 in tumors of patients with colorectal cancer was significantly increased in comparison with the adjacent normal mucosa [19]. Our results are also in accordance with the study of Wu et al., who demonstrated an increased expression of MMP-2 in CRC tissues, but negative in normal colorectal tissues [6], and with the paper of Kim et al. that also indicated higher levels of MMP-2 protein in colon and rectal tumor tissues than in corresponding paired normal mucosa, although, in their study, TIMP-2 tissue levels were significantly lower in cancer than in normal tissue [27]. Similarly, opposite results were obtained by Li et al., who revealed that TIMP-2 expression in CRC tissues was significantly lower than that in normal tissues [13]. Moreover, in our study, the highest percentage of positive reactions of MMP-2 was observed in interstitial inflammatory infiltrate cells. Obtained results may suggest that the main source of MMP-2 in cancer tissue is rather interstitial inflammatory cells from tumor microenvironment than malignant tumor cells. Our results may also confirm the role of inflammation in carcinogenesis of colorectal cells, which was suggested by other authors and in our previous study [28–30].
Interestingly, in both types of cells, the higher percentages of positive immunoreactivity were shown for TIMP-2 than for MMP-2; in cancer cells, we observed TIMP-2 immunostaining in 87.5 % cases in comparison with 23.6 % of MMP-2 positivity, whereas 75 % of inflammatory cells were TIMP-2 positive in comparison with 50 % of inflammatory cells with MMP-2 immunoreactivity. Our observations are in line with the results of Schwandner et al., who demonstrated that MMP-2 tumor expression was positive in 35 % of epithelial malignant cells, whereas there was an expression of TIMP-2 in 47 % of cases in rectal carcinoma [31].
In our current study, the correlations between expression of both proteins tested and clinicopathological features of CRC were evaluated. We did not confirm any significant correlation between MMP-2 and TIMP-2 expression in colorectal cancer cells or normal cells and tumor stage, tumor size (T), nodal involvement (N), presence of distant metastases (M), or resectability of tumor. In inflammatory cells, the expression of MMP-2 revealed a significant correlation with M factor and was higher in patients without distant metastases. The significant differences were also found for TIMP-2 expression in inflammatory infiltrate cell; the percentage of positive reactions decreased in patients with higher tumor size, number of lymph node involved, and metastases to distant organs. The opposite results were obtained by Jung et al., who demonstrated that the upregulation of TIMP-2 in submucosal colorectal cancer tissue was positively correlated with adjacent lymphatic vessel invasion and lymph node metastasis in submucosally invasive colorectal carcinoma [32].
Additionally, we compared serum levels of proteins tested with their tissue expression. We demonstrated that MMP-2 serum levels were significantly higher in patients with positive expression of this enzyme in cancer cells and in CRC patients with positive expression of TIMP-2 in inflammatory and normal cells. The observations are in agreement with the results of Spearman correlation test, which shows a significant positive association between serum MMP-2 concentration and expression of its inhibitor in the same types of cells as well as between serum levels of both proteins. These findings suggest a complex role of MMP-2/TIMP-2 network in colorectal cancer development and metastasis.
Strong expression of many MMPs has been related to poor survival of CRC patients. The expression of various TIMPs has been associated with both a beneficial and a poor outcome. Thus, there is a need to further clarify the significance of MMPs and TIMPs in CRC. In the present study, we also investigated whether the serum concentrations of MMP-2 and TIMP-2 and intensity of their expression in various types of cell might be useful prognostic factors for survival of CRC patients. The univariate regression analysis indicated only the presence of MMP-2 expression in normal colorectal cells as a significant prognostic factor for patients’ survival, although not independent on tumor stage. Our observations indicate that the expression of MMP-2 in normal mucosa of CRC patients could also be relevant as well to the outcome in cancer cells. Our results are also in line with the study of Langers et al. [33], who have shown that a high expression of MMP-2 in the normal colorectal mucosa was associated with reduced survival of CRC patients.
Additionally, our results show the tumor stage, nodal involvement, presence of distant metastases, and tumor resectability, but not serum MMP-2 nor TIMP-2 as the significant prognostic factors in CRC patients in univariate analysis. Obtained results are in agreement with our previous study [14], where preoperative serum MMP-2 and TIMP-2 levels were not prognostic markers for CRC patients. Similar observations were also demonstrated by Waas et al. [5], who showed that preoperative plasma pro-MMP-2 levels had no potential value as prognostic markers in colorectal cancer.
In the present paper, we defined the ROC AUC for the proteins tested to assess a potential clinical significance of MMP-2 and TIMP-2 in the diagnosis of CRC. We found that AUC for serum MMP-2 (0.740) was higher than for TIMP-2 (0.643), and both were significantly higher than AUC = 0.5. These results suggest the potential clinical usefulness of pretreatment serum MMP-2 and TIMP-2 as tumor markers in CRC.
In conclusion, the aim of the present study was to assess the serum levels of MMP-2 and TIMP-2 as well as tumor tissue expression of these proteins in patients with colorectal cancer. We have found the significant associations between decreased MMP-2 expressions in inflammatory cells and distant metastases as well as between tumor invasion and reduced expression of TIMP-2 in inflammatory interstitium. Moreover, the immunoreactivity of MMP-2 and TIMP-2 in colorectal cancer tissue correlated with the expression of these proteins in the interstitial inflammatory infiltrate cells as well as with the serum levels of TIMP-2 in CRC patients. The area under ROC curve for serum levels of proteins tested was higher for MMP-2 than for TIMP-2, what indicates a possible clinical significance of serum MMP-2 in the diagnosis of CRC, while the tissue expression of MMP-2 as a prognostic factor of patients’ survival. Our results suggest a complex network of interactions between tumor, its microenvironment, and stromal cells, but this issue requires further investigations.