Background
Breast cancer is the most common cancer affecting women in the world today. It is the leading cause of cancer related death for women aged between 35 and 55 years worldwide. One in nine women will suffer from breast cancer during her life and in excess 130 thousand women die from breast cancer each year [
1]. According to histological features invasive breast cancers are classified into three groups: well differentiated (grade 1, G1), moderately differentiated (grade 2, G2) and poorly differentiated (grade 3, G3) tumors. Distant metastases are the principal cause of death. An essential process in forming distant metastases is the degradation of the extracellular matrix allowing tumor cells to invade local tissue, intravasate and extravasate blood vessels and build new metastatic formations. This process is primarily influenced by the activity of proteinases secreted by the tumor. Currently, at least four classes of proteinases are known: serine proteinases, aspartatic proteinases, cystein proteinases and matrix metalloproteinases [
2‐
4]. Collectively, these proteinases are capable of breaking down all components of the extracellular matrix. Under physiological conditions (e.g. tissue remodeling, angiogenesis, ovulation, wound healing) there is a precise regulation between proteolytic degradation and regulatory inhibition of proteolysis [
2‐
5]. This physiological balance seems to be disrupted in cancer. Matrix metalloproteinases (MMPs) are up regulated in almost every type of cancer and their expression is often associated with a poor prognosis for patients [
6,
7]. Previous studies have shown the expression and activity of MMPs to be linked to an advanced stage of breast cancer, increased invasion of tumour cells and building of metastatic formations [reviewed in [
8]].
MMPs are a family of structural and functional related endopeptidases. They are, with exception of MMP-11, secreted as inactive zymogens and activated outside the cell by other activated MMPs or serine proteases (e.g trypsin, plasmin, kallikrein) [
2‐
4]. For their activation, a proteolytic removal of the propeptide-domain is required. This enables access to the catalytic site of the MMPs. The cleavage of the extracellular matrix (ECM) by activated MMPs facilitates the invasion of tumor cells as well as the release of ECM bound growth factors (e.g. of insulin like growth factors and fibroblast growth factors). Further, some of the resulting ECM-protein fragments can feature new biological functions (e.g. cleavage of laminin-5 or collagen type IV results in uncovering of their cryptic site which can promote migration of different cell types) [
2‐
4].
Currently, 23 members of the MMP family are known in humans. According to their substrate specificity, they are divided into six subclasses: collagenases, gelatinases, stromelysins, matrilysins, membrane-type MMPs and others [
2]. So far, most investigators have focused on the expression profile of the two gelatinases, MMP-2 and MMP-9, which are able to degrade type IV collagen. [
9‐
11]. Type IV collagen is abundant in basement membranes separating the epithelial cells from the underlying stroma. Increased expression and activity of MMP-2 and -9 in tumors leads to the degradation of basement membranes, an essential step in tumor invasion. In this respect, a correlation between a high expression of MMP-2 and reduced survival in breast cancer patients [
12] as well as an association of the tumor grade with increased levels of MMP-9 in breast cancer tissue [
13] was described. Efficient reduction of MMP-2 and -9 levels was observed during
in vitro treatment of MCF-7 breast cancer cells with the aromatase-inhibitor letrozole suggesting that this inhibitor suppresses both breast cancer growth and invasion [
14].
For MMP-1, -7, -9, -13 and -14 an association between their high expression and a shortened relapse free survival in breast cancer patients was found [
15]. A high expression of MMP-9 and -11 in breast cancer tissue was also detected by Northern blot analysis [
16]. Vizoso et al. showed a correlation between high expression levels of these two MMPs and a higher rate of distant metastases using immunohistochemistry [
15]. In addition, the expression of MMP-2, -8, -9, -10, -11 and -13 mRNA in breast cancer tissue was identified by RT-PCR [
17]. Gonzalez et al. found MMP-1, -7 and -13 to be expressed at higher levels in androgen receptor positive breast cancer cells using immunohistochemistry [
18]. This suggests, that androgen receptors might be able to up regulate MMPs and contribute to a higher invasive potential of breast cancer cells. Using Northern Blot analysis, a higher expression of MMP-2, -7, -9 and -11 mRNA was shown in breast cancer tissue in comparison to normal breast tissue [
19]. In addition, a higher content of MMP-1 and -9 protein was detected in breast cancer tissue when compared to normal breast tissue by ELISA technique [
20]. Using substrate zymography, MMP-1, -2, -3 and -9 demonstrated a higher activity in tumor tissues compared to healthy samples [
5]. Expression of MMP-3 was shown to be higher in stromal tissue surrounding the epithelial tumor cells than in tumor cells themselves [
21]. A correlation between positive nodal status and the expression of MMP-14 and -15 mRNA was found by Ueno et al. [
22]. Haupt et al. found MMP-14 to be expressed predominantly in preinvasive lesions of breast cancer using RT-PCR [
21]. To our knowledge, currently there are no data available for the remaining MMPs regarding their expression in breast cancer tissue in literature.
Thus, the aim of our study was to investigate, if a pattern of MMPs could be identified, whose expression is related to tumor grade in breast cancer. With this objective, we analyzed the expression of all human MMPs known so far in a panel of normal breast and breast cancer samples by semiquantitative RT-PCR, Western Blot and immunohistochemistry. Further, in order to create an independent and reproducible model system for the in vitro analysis of the regulation of MMP expression in breast cancer cells, we analyzed the MMP-pattern in four breast cancer cell lines frequently used in basic breast cancer research (MCF-7, BT-20, MDA-MB-468, ZR 75/1). For an overview, data concerning MMP expression in breast cancer cell lines published so far are summarized in Additional file
1: MMP expression in different breast cancer cell lines [
23‐
27].
Methods
Tissue samples
The study was performed with approval of the Ethics Committee of the University of Wuerzburg, Germany. To establish an expression profile of MMPs typical for breast cancer, we compared normal breast tissue to breast cancer tissue with higher grading (G2 and G3). We excluded well-differentiated breast cancer (G1) tissue, because material of G1-tumors was rare. MMP expression was analyzed in five normal breast and twenty breast cancer tissue samples. The normal breast tissue samples were obtained from patients who underwent reductive mammoplasty for cosmetic reasons (tissue was analyzed histologically to exclude that there were any forms of malignancy or other pathological findings; data not shown). Written informed consent was obtained from each patient according to our institutional regulations. Twenty samples of breast cancer tissue (ten G2 and ten G3 tumors) were obtained during surgical removal of the tumor at the department of Gynaecology and Obstetrics at the University Hospital of Wuerzburg, Germany. Some parts of the samples were immediately frozen in liquid nitrogen and stored at -80°C, other parts were fixed in formalin and embedded in paraffin. Patients' data are listed in Table
1. Tissue was selected in order to contain approximately equal amounts of stromal tissue (as analyzed by HE-staining). Therefore stroma was present to nearly the same extent in each tissue sample that was analyzed by RT-PCR and Western blot analysis. Tissue samples were examined for content of benign and/or malign glandular tissue by HE-staining also. Only samples with high density of benign lobuli/acini or tumour cell mass without signs of necrosis were used for further analysis.
age at diagnosis (y) | | |
mean | 64,7 | 71,8 |
minimum | 42 | 52 |
maximum | 91 | 90 |
T (tumor size) | | |
1 | 4 | 3 |
2 | 5 | 2 |
3 | - | - |
4 | 1 | 5 |
N (nodal status) | | |
0 | 6 | 1 |
1 | 2 | 5 |
2 | 1 | 2 |
X | 1 | 2 |
M | | |
0 | 10 | 6 |
1 | - | 4 |
L | | |
0 | 7 | 2 |
1 | 3 | 8 |
multifocal | 4 | 3 |
inflammatory | 1 | 4 |
invasive ductal | 8 | 9 |
invasive ductal +DCIS | 2 | - |
invasive lobular | - | 1 |
receptor status | | |
ER +/PR+ | 3 | 4 |
ER +/PR - | 1 | 1 |
ER -/any PR | 5 | 3 |
n.d. | 1 | 2 |
Her2neu 2+/3+ | 2 | 3 |
Her2neu negative | 5 | 2 |
n.d. | 3 | 5 |
Cell culture
Cell lines (MCF-7, MDA-MB-468, BT 20, ZR 75/1) were obtained from Cell Lines Service (Eppelheim, Germany) [
28]. Characteristics of the cell lines are listed in Table
2[
29‐
36]. Briefly, cells were cultured in a mixture of DMEM/Ham's F-12 (PAA, Coelbe, Germany) supplemented with 10% FCS and 10 ng/ml gentamycine at 37°C in the presence of 5% CO
2. Cells were cultured in 75 ml culture-flasks (Biochrom, Berlin, Germany) as monolayer culture and harvested at 80–90% confluency using a cell-scraper (Biochrom) or accutase (PAA) treatment. Separated cells were resuspended in phosphate-buffered saline (PBS) and washed twice. Cells were then counted and checked for viability using trypan blue and either subjected to immunocytochemistry or immediately frozen as dry pellets at -20°C for further analysis. All cell preparations used had a viability of >95%.
Cell type | Adeno-carcinoma | Adeno-carcinoma | Invasive ductal carcinoma | Adeno-carcinoma |
Origin | Metastasis (pleural effusion) | Metastasis (pleural effusion) | Primary tumor | Metastasis (ascites fluid) |
Estrogen-/progesterone-receptor | +/+ | -/- | -/- | +/+ |
Invasive potential | Low | Low | Low | - |
Metastatic potential | -/+ | Low | - | - |
RNA extraction and cDNA synthesis
Frozen blocks (1 cm2) of normal and tumor tissue were cut into sections of 6 μm; sections corresponding to 30 mg were collected in a sterile microtube and subjected to RNA isolation. In case of cultured cell lines, 106 cells were used for RNA extraction. Total RNA was extracted using RNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturers' instruction. RNA was eluted in 60 μl RNase free water and stored at -20°C. Total RNA was reverse transcribed at 42°C for 1 h in a 20 μl reaction mixture using the RevertAid H Minus First Strand cDNA synthesis kit (Fermentas, St. Leon-Roth, Germany) and terminated by heating the samples at 70°C for 10 min. Synthesized cDNA was stored at -20°C for further expression analysis.
Semiquantitative RT-PCR
Expression analysis of MMPs was performed using self-created gene specific primers. Primer sequences and PCR conditions are summarized in Additional File
2. In general, conventional PCR reaction was performed in 25 μl volumes containing template DNA, 2.5 U Taq polymerase, 10× buffer with 1.5 mM MgCl
2 (all Eppendorf, Hamburg, Germany), 200 μM dNTPs (Fermentas), 0.4 μM of both, forward and reverse primers and formamide, which was used optionally at a final concentration of 4%. PCR conditions were optimized for each primer-pair. Amplification reactions were performed using a Px2 thermal cycler (Techne, Staffordshire, U.K.) and consisted of following steps: 94°C for 5 min, 28–32 cycles at 94°C for 30 sec; optimized annealing temperatures for 30 sec and 72°C for 10 min. The amount of cDNA was normalized to the intensity of the PCR products of the housekeeping gene (PBGD) [
37]. All PCR products were separated on 1% agarose gels and visualized using GelRed (Biotium, Inc., Hayward, CA). Intensity of GelRed luminescence was measured using ImageJ software (NIH, Bethesda, USA). All RT-PCRs were performed in triplicates.
Western blotting
For protein extraction, 20 mg of cryo-cut tissue samples or respectively 106 cells were lysed in pre-cooled Ripa-buffer (Pierce, Rockford, Ilinois) containing phosphatase inhibitors (Phosphatase Inhibitor Cocktails Set II, Calbiochem, Germany), proteinase inhibitors (complete, Roche, Germany) and 2,5 mM DTT (Dithiothreitol, Sigma, Taufkirchen, Germany) as reducing agent. The mixture was incubated on ice for 30 min, combined with vortexing every 10 min. Cell lysates were clarified of cell debris by centrifugation at 14.000 g for 5 min through a QIAshredder spin column assembly (Qiagen, Hilden, Germany). Afterwards, the samples were mixed in 5× loading buffer (Fermentas), denatured at 95°C for 5 min, chilled on ice and stored at -20°C for further analysis.
Protein concentration was determined by the Bradford-method [
38] using comassie brilliant blue (Roti-Quant; Roth, Karsruhe, Germay). Samples were subjected to electrophoresis on a 10% polyacrylamide gel (SDS-PAGE) and blotted onto a nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany) for 45 min at 10 V using a semi-dry-transfer unit (PeqLab, Erlangen, Germany). Membranes were stained with ponceau-red to verify that the proteins were blotted. To avoid non-specific binding, membranes were blocked with 5% nonfat milk protein in PBS/Tween at RT for 1 hour. Subsequently, the membranes were incubated with the primary antibody diluted in 2% nonfat milk and PBS/Tween at 4°C for 18 hours. As internal loading control, anti-β-actin primary antibody was used. Clones, sources and dilutions of the primary antibodies used herein are summarized in Table
3. After washing with PBS, the membranes were incubated with species specific horseradish peroxidase-conjugated secondary antibodies (listed in Table
3) for 60 min at RT. Immunoblots were visualized by home made "enhanced chemiluminescence" ECL [
39]. Resulting images were quantified using ImageJ software (NIH, Bethesda, USA).
Table 3
Antibodies for Western blot and immunohistochemistry
MMP-1 | WB, IHC | latent and active | rabbit | polyclonal | 1:750 | 1:100 | Biozol |
MMP-2 | WB, IHC | latent and active | rabbit | polyclonal | 1: 1000 | 1:100 | Abcam |
MMP-3 | WB, IHC, ICC | latent and active | mouse | SPM 293 | 1: 500 | 1:50 | Abcam |
MMP-7 | WB | latent and active | mouse | 111433 | 1: 500 | - | Abcam |
MMP-8 | WB | latent and active | mouse | 115-13D2 | 1: 1000 | - | Chemicon |
MMP-9 | WB | latent and active | mouse | 9D4.2 | 1: 500 | - | Chemicon |
MMP-10 | WB, IHC, ICC | latent and active | mouse | IVC5 | 1: 500 | 1:100 | Chemicon |
MMP-11 | WB | latent and active | mouse | SL 3.01 | 1: 500 | - | Abcam |
MMP-11 | IHC, ICC | latent and active | mouse | SPM 199 | - | Prediluted | Biozol |
MMP-12 | WB | latent and active | rabbit | polyclonal | 1: 1000 | - | Abcam |
MMP-13 | WB, IHC | latent and active | mouse | 87512 | 1: 500 | 1:100 | R&D |
MMP-14 | WB, IHC, ICC | latent and active | rabbit | polyclonal | 1: 500 | 1:75 | Abcam |
MMP-15 | WB, IHC, ICC | latent and active | rabbit | polyclonal | 1: 500 | 1:100 | Abcam |
MMP-19 | WB | latent and active | rabbit | polyclonal | 1: 3000 | - | Biozol |
MMP-19 | IHC, ICC | latent and active | rabbit | polyclonal | - | Prediluted | Biozol |
MMP-23 | WB | latent and active | rabbit | polyclonal | 1: 1000 | - | Abcam |
MMP-24 | WB | latent and active | rabbit | polyclonal | 1: 1000 | - | Abcam |
MMP-27 | WB | not specified | rabbit | polyclonal | 1: 1000 | - | Abcam |
MMP-28 | WB | not specified | rabbit | polyclonal | 1: 1000 | - | Abcam |
β-actin | WB | b-actin | mouse | M/Abcam 8226 | 1: 10.000 | - | Abcam |
Immunohistochemistry
For immunohistochemistry, tissue samples were cut at 2 μm from formalin-fixed, paraffin-embedded tissue blocks, placed on adhesive treated slides (Superfrost, Langenbrinck, Emmendingen, Germany) and dried overnight at room temperature. Paraffin sections were dewaxed twice with xylene and rehydrated in a graded series of ethanol and in distilled water. The sections were stained without pretreatment for antigen demasking. Only in the case of MMP-19, slides were pretreated in the microwave oven in a 10 mM sodium citrate buffer solution (pH 6.0) for 10 minutes (750 W/s). Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide in methanol for 10 minutes. To reduce non-specific binding capacity of the tissue, slides were treated with a solution of human immunoglobulin (Beriglobin; Aventis Behring, Marburg, Germany) in phosphate buffered saline (PBS, dilution 1:50) for 15 minutes at room temperature. Afterwards the sections were incubated overnight at 4°C with one of the respective primary antibodies against MMP-1, -2, -3, -10, -11, -13, -14, -15 and -19 diluted in antibody diluent (DAKO, Hamburg, Germany; antibodies are listed in Table
3). The other antibodies listed in Table
2 did not produce sufficient staining results for immunohistochemistry and could therefore not be analyzed herein. After washing with PBS, the sections were incubated with horseradish-peroxidase (HRP)-labeled LSAB2 kit (streptavidin-biotin system; DAKO). Peroxidase activity was developed with diaminobenzidine (DAB; DCS, Hamburg, Germany) as a substrate for 5 min, which resulted in brown staining. The slides were counterstained with haematoxylin, dehydrated in graded ethanol, embedded in Vitro Clud (Langenbrinck) and analyzed using a light microscope Othoplan (Leica, Germany).
Immunocytochemistry
For immunocytochemistry, cells harvested with accutase were diluted at 1 × 106 cells/ml. 20 μl of each cell line was applied on APES-treated slides, air dried and fixed in a 2% formalin solution (diluted in PBS) for 15 min. Cells were then permeabilized by incubation in 0.05% Triton X100 in PBS for 10 min followed by a single wash in PBS. Blocking of non-specific binding capacities, incubation and detection of primary antibodies as well as counterstaining and embedding was performed as described for immunohistochemistry. All incubation steps apart from the primary antibodies were carried out at room temperature.
Antibodies were validated for this staining technique using cell lines of different cancer sources (breast, cervix, placenta, ovary and endometrium) with known MMP expression in RT-PCR and/or WesternBlot. With the staining protocol applied, only MMP-3, -11, -14, -15 and -19 produced reproducible results.
Data analysis and statistics
The intensity of GelRed luminescence and protein expression in Western Blot images was quantified densitometrically using ImageJ software (NIH, Bethesda, USA) and normalized in respect to the corresponding fragment concentration of the ubiquitously expressed genes PBGD and β-actin. Four different expression levels were considered in respect of their densitometrical value. Value 0 was considered to be no expression. Values between 1 and 19 were considered as very weak ((+)), between 20 and 49 as weak (+), between 50 and 79 as moderate (++) and between 80 and 100 as high (+++) expression.
Box-plots were generated using GraphPad Prism 4.0 Software (GraphPad Software, La Jolla, USA). Comparison of expression values between the groups was performed by the non-parametric two-tailed Mann-Whitney-U-test and P-values < 0.05 were considered as statistically significant.
Discussion
Breast cancer is an epithelial tumor with high invasive and metastatic potential. Poorly differentiated tumors (G3) seem to have a higher invasive potential resulting in a higher frequency of lymph node metastases and lymphangiosis carcinomatosa than well differentiated tumours (G1) [
40]. These aspects were also found in the G3 tumor patient group investigated in this study (Table
1).
In our study we focused on the expression analysis of human specific matrix metalloproteinases (MMPs) in breast cancer specimens. The expression analysis showed three different expression patterns: high, low or equal expression of MMPs in breast cancer tissue compared to normal breast tissue. For the three collagenases, MMP-1, -8 and -13, the first type of the expression pattern was identified. Our data about MMP-1 are in correlation with previously published results [
5,
18,
20,
41]. We found the increase of the expression of latent and active protein forms of MMP-1 to be related to higher tumour grade. This finding fits with previous published data, reporting on a shortened relapse-free-survival in breast cancer patients with high expression of MMP-1 [
15]. Published data about the site of the MMP-1 production are contradictory [
42]. However, we showed that MMP-1 expression is clearly restricted to the tumour cells, thus confirming the findings of Iwata et al. [
41]. For MMP-8, some studies indicate its inhibitory effect on building metastatic formations [
17,
43], and no correlation between MMP-8 expression in breast cancer and negative prognostic factors was found using ELISA [
44]. In contrast to this, we found a significant higher expression of pro- and active MMP-8 in breast cancer compared to normal breast tissue. A correlation between the high expression of MMP-13 in breast cancer tissue and a higher rate of distant metastases, poor prognosis and in addition a higher expression of MMP-13 in early stage tumors was shown [
15,
45]. In accordance to this, we also identified higher amounts of MMP-13 mRNA and pro-form in breast cancer tissue. Similar to the findings of Zhang and colleagues [
45], we identified MMP-13 protein to be predominately expressed in the cytoplasm of tumor cells. The expression of the two gelatinases, MMP-2 and MMP-9, in breast cancer is well investigated in many studies using different methods [
13,
16,
19,
20,
41,
46]. Our Western blot data showed a significant increase of proMMP-2 expression in tumor tissue, which is in accordance with data described [
15,
19,
46]. In addition, MMP-2 staining was specific for nuclei and cytoplasm of tumor cells. This is in accordance with other studies, verifying MMP-2 to be an unfavorable prognostic factor in breast cancer [
5,
13]. Our findings about the high expression of MMP-9 in the specimens analyzed are in accordance with data published by Przybylowska and colleagues, who found increased levels of MMP-9 to correlate with G3 breast cancer [
20]. In contrast, studies of Jones and Rhako failed to demonstrate any association between the expression of MMP-9 in carcinoma or stromal cells and clinicopathological parameters using immunohistochemistry [
47,
48]. These differences might be due to different detection methods used. For the matrilysin MMP-7 only few contradictory data are available suggesting that its role in breast cancer has not been brightly investigated yet [
15,
16,
19,
42]. We also found divergent data concerning the expression of MMP-7 in our tissue samples. While the expression of MMP-7 mRNA was significantly lower in breast cancer tissue when compared to normal breast tissue. Equal to higher expression levels of MMP-7 pro and active forms in breast cancer tissue were detected. This could be due to a regulation on the translational level, resulting in low levels of MMP-7 mRNA in breast cancer tissue and – as effect of a higher transcriptional rate in breast cancer tissue – higher protein levels of MMP-7 in breast cancer tissue. In accordance to this, the expression of the latent form of MMP-7 was stronger than the expression of its active form in the analyzed samples, like in the case of MMP-23 and -24, as mentioned above. Although the amount of MMP-10 (stromelysin 2) on mRNA level was equal in all analyzed samples, expression of its pro and active protein form was significantly higher in breast cancer samples.
Using immunohistochemistry we found a nuclear staining for MMP-1, -2 and -10, which is unusual, because MMPs are considered to be cytoplasmatic or membrane-bound proteins. The nuclear localization of the abovementioned MMPs could thus be an indication of further new functional roles of MMPs. In accordance to this, an atypical nuclear staining was also observed by Ip et al. for MMP-14 in hepatocellular carcinoma [
49]. In this immunohistochemical study the nuclear localization of MMP-14 was associated with aggressive tumor features including poor prognosis.
In accordance to our data, that clearly show a significant increase in the expression of MMP-11 mRNA, propeptide and peptide in breast cancer compared to normal breast tissue, the studies of Pacheco and Kossakowska reported the same expression pattern for this matrix metalloproteinase [
16,
19]. For MMP-15 (MT2-MMP) no correlation between its expression, positive axillary node status, distant metastases and size of breast tumor was detected by immunohistochemistry [
22]. However, our data show a high expression of its protein in breast cancer tissue for both, latent and active forms. For another membrane bound MMP, MMP-24, only the size of the latent protein was known and its expression was found to be significantly higher in breast cancer tissue. A putative active bandage was detected at about 55 kDa; this could possibly be an active breakdown product of MMP-24, although the precise size of the active form of MMP-24 is not known till now. The expression of its latent form was stronger than the expression of the putative active form in the analyzed samples. The absence of activated MMPs in tumor samples is not unique; this phenomenon was also observed for MMP-9 in ovarian cancer [
50]. Expression and activity of proMMP-9, but not its active form, was shown in the aggressive form of this tumor and could reflect the presence of inflammatory cells, which promote tumor progression.
In an immunohistochemical study comparing normal breast tissue and mammary gland tumors, Djonov and colleagues found MMP-19 to be expressed in all benign lesions, whereas no expression was found in tumor tissue [
51]. However, we found MMP-19 protein to be higher expressed in cancer tissue although its mRNA was equally expressed in all analyzed samples. These different findings might be due to different antibodies used in both studies. The same expression pattern was identified for the latent and active forms of MMP-12, -27 and -28. Active MMP-23 protein showed also a higher expression in G2 compared to normal breast specimens. However, similar to MMP-24, expression of its active form was weaker than of the latent form. Although the expression of MMP-27 mRNA was lower in breast cancer samples, its protein showed a diverse expression pattern in the analyzed samples.
To date there are very few publications on the most recently described MMP-27 and -28, and only very little information about their protein sizes.
Showing a statistical significant difference in the expression level between normal breast and breast cancer tissue grade 2 and 3 it seems possible, that some of the putative bandages of MMP-27 and -28 are playing a role in breast cancer development. Because till now no precise sizes of their breakdown products and putative active and inactive forms are known, these differences in the expression level between normal breast and breast cancer tissue can give us a first sign that these two MMPs can be involved in tumor progression.
A strong association between the expression of MMP-14 by stromal cells and poor prognosis for the patients was found by Vizoso et al. using immunohistochemistry [
15]. However, in our study we found MMP-14 mRNA to be equally expressed in normal breast and breast cancer tissue. These differences might be due to the different expression profiles of MMP-14 on mRNA and protein level. Therefore MMP-14 mRNA could possibly be expressed on equal levels in breast cancer tissue and normal breast tissue. Differences in expression might also be caused on the translational level. Using immunohistochemistry we found MMP-14 protein to be localized in the cytoplasm of tumor cells with only slight additional staining of the surrounding stroma cells, whereas no staining in normal breast tissue could be detected. One possible reason might be a higher translational rate of MMP-14 in breast cancer tissue. MMP14 plays an important role in the activation of other MMPs [
52]. For example, MMP-14 activates proMMP-2 at the cell surface [
52]. In our study, MMP-2 mRNA was ubiquitously expressed in all analyzed samples, but – however – without a clear correlation to the expression of its activator, MMP-14.
For the remaining MMPs there are only limited or no data available in the literature about their expression in breast cancer tissue. Some studies documented the expression of MMP-3 in breast cancer [
5,
21,
42]. In one of those, significantly higher amounts of proMMP-3 and equal amounts of its active form in breast cancer tissue compared to normal breast tissue were found using zymography [
5]. These findings are in accordance with the results obtained in our study. Furthermore, only a weak staining of MMP-3 was found in tumor cells using immunohistochemistry, which thus correlates with published results [
5,
41]. In contrast to this Haupt et al. found MMP-3 to be located in the stromal tissue that surrounds the epithelial tumor cells [
21]. Ueno and colleagues could not detect MMP-16 (MT3-MMP) in breast cancer tissue using Northern blot analysis [
22], whereas we detected its mRNA in normal and breast cancer tissue by RT-PCR. These differences in MMP-16 detection could be due to the different sensitivity of the methods used. MMP-17 (MT4-MMP) showed the same expression pattern in all tissues analyzed. For MMP-26 no expression could be found in normal mammary glands, whereas a high expression could be identified in precursor lesions (DCIS) using immunohistochemistry. Further, on tumor progression to invasive carcinomas the expression level of MMP-26 was described to decrease again [
53]. This is partly in accordance to our results, as we did not detect MMP-26 mRNA in breast cancer tissue G2 and G3, though we could not detect MMP-26 mRNA in normal breast tissue either. However, PCR products obtained by amplification of the genomic DNA, which was used as positive control, showed that PCR properly worked. As MMP-26 could not be detected by PCR, we did not investigate its expression on protein level. Expression of MMP-21 and MMP-25 was very low on mRNA level, whereas MMP-20 showed no expression. Therefore we did not investigate them in further.
Expression of MMPs in breast cancer cell lines
Since there is obvious evidence about the influence of MMPs on the development of breast cancer, we studied their expression pattern in four different breast cancer cell lines, which could be used as a model system for the analysis of the regulation of MMP expression in this tumor entity. Our analyses showed, that the MDA-MB-468 cell line expresses sixteen out of twenty- three MMPs analyzed on mRNA and protein level. Using immunocytochemistry, we could confirm our data obtained by RT-PCR and Western Blot analysis in this cell line, as we found a weak staining for MMP-3 in the cytoplasm and nuclei of the tumor cells. A moderate to strong expression of MMP-11 and -14 was identified in nearly all MDA-MB-468 cells, too. MMP-15 and -19 expression was clearly restricted to the cytoplasm of tumor cells using immunocytochemistry.
This is partly in contrast to the results obtained by Gimbernardi et al. [
9] and Grant et al. [
54], who studied the expression of MMP-1, -2, -3, -7, -8, -9, -10, -11, -12, -13, -14, -15, -16, -17, -18 and -20 by RT-PCR in 84 normal or human cancer cell lines, including some breast cancer cell lines such as BT 20, MCF-7, MDA-MB-468, ZR 75/1 and T47D. They observed the expression of MMP-9, -10, -14 and -15 in MDA-MB-468 cells only, whereas we could detect the expression of MMP-1, -2, -3, -8, -11, -16, -17 and -19 in this cell line too. In contrast to their findings, we did not detect any expression of MMP-7 and MMP-10 on mRNA level. These differences might be due to differences in the sensitivity of the methods and primers used and/or to cell culture conditions. Differences in MMP expression profiles caused by different cell culture conditions were recently shown [
55]. Kousidou et al. observed a higher expression of MMP-1 and MMP-11 in MCF-7, BT 20 and ZR 75/1 cell lines when cultured in serum-free media, whereas expression of MMP-9 in these three cell lines was higher when serum was added to the medium [
55]. Similar to this, in our serum-cultured cells, we observed a weak to high expression for MMP-1, -2, -13, -15, -17, -23 and -28 in BT 20 cells, whereas Gimbernardi et al. detected MMP-15 only in this cell line. For the remaining MMPs, no expression in BT-20 cells was detected in our study.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
AK drafted the manuscript, set up the experiments, collected the data, analyzed and interpreted the results. UK participated in the study design, interpretation of the results and finalization of the manuscript. JD participated in editorial support. MK carried out the Western blot analysis and immunohistochemistry. JA participated in the study design, experimental concept, interpretation of the results and drafting of the manuscript. All authors read and approved the final manuscript.