Prospective evaluation of serum tissue inhibitor of metalloproteinase 1 and carbonic anhydrase IX in correlation to circulating tumor cells in patients with metastatic breast cancer

In breast cancer patients, hematogenous tumor cell dissemination is a crucial step in tumor progression and blood-borne metastases account for the vast majority of breast cancer-related death. Circulating tumor cells (CTC) derived from primary tumors and metastatic sites can be detected in the circulation. Many methods for the detection of CTC have been described [1, 2]. At present, the CellSearch^® system, which combines both automated enrichment and immunostaining, is the only standardized technology that was approved by the Food and Drug Administration for the detection of CTC in patients with metastatic breast, colon, and prostate cancer [3‐5]. The detection of CTC in blood can provide prognostic information [3, 6]. Moreover, CTC detection and characterization has already improved our understanding of the complex process underlying tumor cell dissemination and metastatic progression in breast cancer. It is widely accepted now that the release of tumor cells from solid tumors requires specific mechanisms such as proteolysis and release is enhanced when tumor hypoxia occurs.

The tissue inhibitor of metalloproteinase 1 (TIMP-1) is of interest because it plays a role in tissue invasion and angiogenesis. A negative prognostic impact of serum TIMP-1 as well as tissue protein levels was described in breast cancer, colorectal cancer, and other malignancies [7‐9]. At one side, TIMP-1 inhibits matrix metalloproteinases (MMPs) and thus, may influence tumor growth and invasion. On the other side, it has been demonstrated that TIMP-1 may inhibit apoptosis in breast epithelial cells [10‐12] and promotes cell growth, tumorigenesis, and angiogenesis in different cell types, including breast carcinoma cell lines [13‐15].

Carbonic anhydrase IX (CAIX) is a metalloenzyme involved in cell adhesion, growth, and survival of tumor cells. There is strong evidence that CAIX is involved in tumor cell proliferation as inhibition of CAIX in vitro and in vivo significantly reduces growth and survival of tumor cells [16]. In several epithelial cancers, CAIX overexpression was shown to be of prognostic relevance [17‐21].

Apart from the cellular transmembrane form of CAIX, there is a soluble isoform that is released by proteolytic cleavage and can be detected in peripheral venous blood [22]. Although several reports indicate a role of serum CAIX in renal cell cancer [22, 23], information about CAIX in serum of breast cancer patients is limited [24].

In conclusion, several publications demonstrated a biologic role for TIMP-1 in breast cancer whereas information on CAIX is limited. For both markers, experimental and clinical data suggest that they might be also involved in tumor cell dissemination. However, TIMP-1 and CAIX have so far not been examined in combination with CTC measurements as a surrogate marker for hematogenous tumor cell spread. Therefore, the aim of this study was to investigate the role of TIMP1 and CAIX serum levels in association with the presence of CTC in metastatic breast cancer.

Insights into the biology of the metastatic potential of breast cancer cells are of relevance for several reasons. One clinical aspect is the identification of patients with more aggressive tumors that might benefit from more intense therapy. Another important reason is the need for an improved understanding of mechanisms leading to hematogenous tumor cell dissemination. This is of potential consequence for the development of new therapeutic approaches. Blood is often preferred over tumor tissue as it is easy to obtain. Also, repeated sampling is possible, which allows the use of markers for monitoring patients during the course of their disease. Thus, we examined the two serum factors TIMP-1 and CAIX and also investigated a correlation to the presence of CTC, which is a "real time" parameter of tumor cell dissemination.

Protein levels of TIMP-1 in tumor tissue are associated with prognosis and therapy response in patients with primary breast cancer [16‐18, 28] and with response to chemotherapy [29]. Schrohl and colleagues were able to show that elevated primary tumor levels for TIMP-1 also have a negative impact in the metastatic situation indicating that TIMP-1 has a general impact on tumor biology [29]. In an experimental study, also a role of TIMP-1 for chemoresistance was described [30]. It was recently shown in a prospective study that levels of TIMP-1 in plasma and serum obtained preoperatively from patients with primary breast cancer are associated with prognosis [31] whereas plasma levels at the time of primary surgery are not correlated with tissue concentrations [32]. Our results support the idea that also serum TIMP-1 levels reflect an enhanced ability of tumor tissues (including metastatic sites) to release cells into the circulation. In our patient cohort, higher TIMP-1 levels are associated with shorter PFS and OS in univariate analysis, supporting the biologic relevance of this factor. However, we cannot provide experimental evidence for a direct role of TIMP-1 in the release of CTC into the circulation. Rather than being involved in tumor cell release, TIMP-1 might stabilize released cells by inhibiting apoptosis through activation of survival pathways either by inhibition of MMPs or independently of MMPs. Furthermore, TIMP-1 protects tumor cells from chemotherapy-induced apoptosis [10‐12]. In addition, TIMP-1 is expressed in a variety of cell types, including tumor cells and stromal cells and detectable in most tissues and in body fluids. It is also expressed by monocytes and macrophages and all these cells might contribute to high TIMP-1 concentrations in tumor tissue. Transcriptional analysis in colorectal cancer revealed that the expression of TIMP-1 in fibroblasts is even higher than that in tumor cells [33]. TIMP-1 protein expression was also high in stromal cells closest to tumor areas [34]. Therefore, secreted TIMP-1 in serum might be derived from different cell types (metastatic tumor cells, surrounding stromal cells, infiltrating macrophages, monocytes and others) making analysis of rare single cells for TIMP-1 difficult. Moreover, CTC were detected with the CellSearch system and the only free fluorescence channel for further characterization of CTC already was occupied for HER2 detection in our study. Furthermore, tumor tissue from metastatic sites was not available for further analysis in most cases. Therefore, it was not possible to directly examine the expression of TIMP-1 and/or CAIX in the tumor tissue or on CTC.

The fact that for TIMP-1 a relevant difference in PFS was observed only in patients with less than five CTC and the OS difference between patients with elevated and non-elevated TIMP-1 was also larger in patients with low CTC numbers could indicate that serum TIMP-1 is especially relevant in patients without the detection of elevated CTC counts. This might open a clinical perspective for this marker also in the context of CTC measurements because even in metastatic breast cancer a relevant portion has CTC counts below the widely established cut off of five or more CTC in 7.5 mL blood. In patients with elevated TIMP-1, OS did not differ between patients with CTC findings above and below the cut off. Therefore, TIMP-1 does not seem to add prognostic relevance in CTC-positive patients. Currently, we do not have an explanation for the observation that patients with elevated TIMP-1 and less than five CTC/7.5 ml had the shortest PFS (Figure 2 and Table 4). In addition, it is not possible to exclude that at least some patients with elevated TIMP-1 levels who are detected to be "CTC-negative" in the CellSearch assay represent a subpopulation of patients with still undetectable CTC that have lost their epithelial characteristics in the course of epithelial-mesenchymal transition [35].

A generally accepted normal level of TIMP-1 serum concentration was not defined so far. Our study was not designed to validate cut-off values for these markers. Lipton et al. applied a cut-off of 454 ng/mL (95% of the control group) when they used control group of 49 healthy postmenopausal women to derive the serum TIMP-1 cut-off with the same assay as used in our study [7].

Altered glycolysis is a main metabolic feature of malignant cells. Combined with decreased oxidative phosphorylation it can result in acidification of the extracellular space [36, 37]. Transcription factors of the glycolytic pathway also influence cell proliferation and differentiation; disordered glycolysis and acidic milieu were therefore proposed to play a major role in the complex multistep process of carcinogenesis [37‐41]. One of the enzymes contributing to acidification of the extracellular space is CAIX. It is a transmembrane zinc enzyme catalyzing the hydration of carbon dioxide [42]. CAIX is strongly induced by hypoxia via activation of transcriptional factors such as HIF-1 [43]. CAIX is overexpressed in a variety of solid tumors with different results for its potential role in gynecologic cancers and breast cancer [44‐48]. The inhibition of this enzyme is a potiential therapeutic approach [49‐51]. In renal cell cancer, overexpression of CAIX is common and the possible role of CAIX targeting antibodies (WX-G250, Rencarex^®) is currently being evaluated in phase III trials for this entity [52]. There is some information on serum values in renal cell cancer patients showing significantly higher values in patients with metastatic disease than in patients with localized cancer. Furthermore, renal cell cancer patients with high serum CAIX before surgery were at significantly higher risk for disease recurrence than those with low preoperative values [23]. The role of serum CAIX in breast cancer has not been determined. As hypoxia is postulated to be associated with hematogenous tumor cell dissemination in breast cancer [53], we examined this factor in correlation to the detection of CTC. It appears that CAIX alone has prognostic significance in both CTC-negative and CTC-positive groups and we provide to our knowledge the first evidence for a prognostic relevance of serum CAIX. This is of potential clinical relevance for the application of new therapeutic approaches inhibiting angiogenesis or directly CAIX.

Moreover, our findings show an association between elevated CAIX serum levels and the presence of CTC which also supports the experimental findings that indicate an association between hypoxia and release of tumor cells into the circulation. Our observations also could indicate that CAIX in contrast to TIMP-1 has prognostic relevance in patients with elevated and non-elevated CTC numbers (Figure 3 and Table 4). However, we cannot definitely explain the correlation we observed because the cellular origin of serum CAIX in our patients is not clear. Elevated levels of both CAIX and TIMP-1 were seen in 18% of patients (Figure 1; Table 2). However, in the overall cohort no relevant improvement for the prognostic information concerning PFS and OS was observed. This does not deliver the rationale for a combined use as prognostic markers.

The lack of correlation between CTC detection and PFS in multivariate analysis in our cohort might be due to different treatments, therapeutic settings (1^st line, 2^nd line, 3^rd line and more), metastatic sites, and by slightly different response monitoring according to the institutional standards of the participating Breast Cancer Centers. For example, in the currently most cited publication using the CellSearch System also applied in our study [3], 47% of patients were starting their first line of therapy whereas in our cohort the rate was only 38%. In addition, the use of targeted therapies (trastuzumab and lapatinib for HER2-positive and bevacizumab for HER2-negative patients) might change the prognostic relevance of CTC detection with respect to PFS as the end point. Similarly to other studies, we did not observe a correlation between CTC detection and estrogen receptor/progesterone receptor and HER2 status of the corresponding primary tumors [3]. The fact that we did not detect a significantly higher number of CTC-positive patients having bone metastasis compared with other sites of metastasis is not in line with some other findings, but might be explained by the relatively small number of patients only with bone metastasis (n = 35) enrolled in our patient study. Taken together, this lack of standardized treatment is a potential drawback of the study However, the strength of our study is the prospective and multicenter setting and the analysis of biomarkers blinded for clinical data.

This study demonstrates with a prospective design that serum TIMP-1 and CAIX have prognostic impact in metastatic breast cancer and elevated serum levels of these invasion markers are associated with the presence of CTC-positivity. The mechanisms that lead to tumor cell dissemination from primary tumors or metastases have not been recognized in detail yet. To find out whether TIMP-1 and CAIX are involved in this process, for example by loosening cell adhesion, changing pH conditions or prolongation of CTC survival has to be further investigated in experimental model systems.