In vitro and in vivo MMP gene expression localisation by In Situ-RT-PCR in cell culture and paraffin embedded human breast cancer cell line xenografts

Background Members of the matrix metalloproteinase (MMP) family of proteases are required for the degradation of the basement membrane and extracellular matrix in both normal and pathological conditions. In vitro, MT1-MMP (MMP-14, membrane type-1-MMP) expression is higher in more invasive human breast cancer (HBC) cell lines, whilst in vivo its expression has been associated with the stroma surrounding breast tumours. MMP-1 (interstitial collagenase) has been associated with MDA-MB-231 invasion in vitro, while MMP-3 (stromelysin-1) has been localised around invasive cells of breast tumours in vivo. As MMPs are not stored intracellularly, the ability to localise their expression to their cells of origin is difficult. Methods We utilised the unique in situ-reverse transcription-polymerase chain reaction (IS-RT-PCR) methodology to localise the in vitro and in vivo gene expression of MT1-MMP, MMP-1 and MMP-3 in human breast cancer. In vitro, MMP induction was examined in the MDA-MB-231 and MCF-7 HBC cell lines following exposure to Concanavalin A (Con A). In vivo, we examined their expression in archival paraffin embedded xenografts derived from a range of HBC cell lines of varied invasive and metastatic potential. Mouse xenografts are heterogenous, containing neoplastic human parenchyma with mouse stroma and vasculature and provide a reproducible in vivo model system correlated to the human disease state. Results In vitro, exposure to Con A increased MT1-MMP gene expression in MDA-MB-231 cells and decreased MT1-MMP gene expression in MCF-7 cells. MMP-1 and MMP-3 gene expression remained unchanged in both cell lines. In vivo, stromal cells recruited into each xenograft demonstrated differences in localised levels of MMP gene expression. Specifically, MDA-MB-231, MDA-MB-435 and Hs578T HBC cell lines are able to influence MMP gene expression in the surrounding stroma. Conclusion We have demonstrated the applicability and sensitivity of IS-RT-PCR for the examination of MMP gene expression both in vitro and in vivo. Induction of MMP gene expression in both the epithelial tumour cells and surrounding stromal cells is associated with increased metastatic potential. Our data demonstrate the contribution of the stroma to epithelial MMP gene expression, and highlight the complexity of the role of MMPs in the stromal-epithelial interactions within breast carcinoma.


Background
Human breast carcinoma (HBC) is the most predominant cancer in females worldwide. Breast cancers can be classified histologically based upon the types and patterns of cells of which they are composed. Carcinomas can be invasive, extending into the surrounding stroma or noninvasive, confined to the epithelial cells of ducts or lobules [1]. In conjunction with other protease systems such as the serine, cysteine and aspartyl proteases, members of the matrix metalloprotease (MMP) family of proteolytic enzymes degrade constituents of the extracellular matrix surrounding invasive breast carcinomas [2]. Currently, 28 human MMPs have been identified and classified according to both their substrate specificities and structural similarities. There are four major subgroups: i) interstitial collagenases; ii) gelatinases; iii) stromelysins; and iv) the membrane-type (MT) -MMPs [3,4]. Collectively, MMPs degrade all extracellular matrix proteins as well as a growing number of key regulatory genes such as cytokines, growth factors, cell surface receptors and adhesion molecules [5,6]. Although MMPs are expressed by tissues at various stages of development, they are typically absent in normal cells of the adult organism [2], and the high frequency at which MMP transcripts or proteins are detected in invasive tumours has implicated these enzymes in the establishment, growth, invasion and/or metastasis of tumours [6]. The expression of MMPs is usually tightly regulated, and a number of studies provide evidence of carefully controlled MMP involvement in developmentally regulated processes such as ovulation, embryogenic growth and differentiation, and organ development [6][7][8].
In general terms, MMP activity is regulated at least at three levels: transcription/translation, proteolytic activation of the zymogen, and inhibition of the active enzyme [2] with upregulation of each of these associated with pathological events [6]. Importantly in all mammalian cells except polymorphonuclear (PMN) leucocytes, most MMPs are not stored intracellularly but are rapidly secreted after biosynthesis and post-translational processing [9]. As a result it is difficult to identify their cellular origin using standard immunohistochemical techniques [10].
MT1-MMP is one of the membrane bound MMPs. In vivo, MT1-MMP expression has been localised to the stroma surrounding breast tumours [11,12], whilst in vitro, our recent data confirms previous studies where basal levels of MT1-MMP have been shown to be higher in the more invasive MDA-MB-231 cells as compared to the less invasive MCF-7 cells [13,14]. MT1-MMP has been shown to activate pro-gelatinase-A (proMMP-2) in human breast carcinoma cells [15] and can also activate proMMP-13 [16]. MT1-MMP degrades native type I collagen, fibronectin, laminin, fibrin, gelatin and cartilage proteoglycan core protein [2] and has also been classified as an interstitial collagenase [3]. MMP-1 is the most ubiquitously expressed interstitial collagenase [3]. MMP-1 cleaves fibrillar collagens including collagens types I, II and III, resulting in cleavage products that are rapidly denatured at body temperature to gelatin, the substrate preferred by the gelatinases (MMP-2 and MMP-9) [2]. MMP-1 is produced by a wide variety of normal cells (eg stromal fibroblasts, macrophages and endothelial cells) [3], and is involved in tissue remodelling and repair [3,17]. It has been implicated in matrix invasion by the MDA-MB-231 HBC cells in vitro [18,19], but has a low incidence of expression in the tumour cells of breast carcinomas [2,11]. MMP-3 is a member of the stromelysin sub-family, which show broad substrate specificity, degrading type IV collagen, laminin, fibronectin and proteoglycans. MMP-3 is expressed in areas of tissue growth [20], focally expressed around invasive cells in the stromal component of breast tumours, and expressed in both benign and malignant breast phenotypes [19].
As MT1-MMP, MMP-1 and MMP-3 have all been implicated in the processes involved in human breast cancer, this study utilised HBC cell lines both in culture and grown as xenografts in nude mice, to investigate gene expression changes of MT1-MMP, MMP-1 and MMP-3 in vitro and in vivo using IS-RT-PCR. In order to demonstrate IS-RT-PCR detection to examine MMP expression level changes, we utilised Concanavalin A (Con A), an agent known to modulate MMP expression, and to upregulate at least one MMP [15,21,22], in our in vitro studies. Nude mouse xenografts are heterogenous, containing neoplastic human parenchyma with mouse stroma and vasculature with many histologic features of the original tumour maintained [23]. As xenograft growth varies with tumour type, the ability to mimic human tumours in vivo provides a reproducible model system that can be correlated to the human disease state. We have previously applied the IS-RT-PCR technique to examine in vitro gene expression of MT1-MMP in MDA-MB-231 HBC cells [21], while others have examined expression of MMP-2, MMP-9 and their inhibitors TIMP-1 and TIMP-2 in cervical cancer [24]. Thus the objective of this study was to utilise the unique IS-RT-PCR methodology to examine the localised gene expression of members of the MMP family of proteases implicated in breast cancer. Our findings here with IS-RT-PCR demonstrate the detection of in vitro and in vivo gene expression patterns of MMPs in HBC cell lines, and suggest a contribution from the stroma to MMP expression in breast carcinomas.

Cell culture
HBC cell lines MDA-MB-231, MCF-7, ZR-75-1 and MDA-MB-453, were originally obtained from the ATCC (Virginia, USA) and maintained by the Lombardi Cancer Center Shared Cell Culture Resource. They were grown as a monolayer culture in RPMI 1640 media (Invitrogen, USA) supplemented with 10% foetal calf serum (Biowhittaker, USA) in the presence of 5% CO 2 . Cells (approx. 1 × 10 5 cells/slide) were passaged onto two eight-chambered chamber slides (Lab-Tek, Nunc, USA) and allowed to attach for 8-12 hrs. Cells were then grown for a further 16-24 hrs in RPMI 1640 with 10% foetal calf serum in the presence or absence of Con A (25 µg/ml, Sigma, USA).

Xenograft tissue
Tumour xenografts of the MDA-MB-231, MDA-MB-435, MCF-7 and Hs578T HBC cell lines were xenografted into intact female mice as described previously [25]. Animal studies were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals with approval of the Georgetown University Medical Center Animal Ethics Committee. The tissue was excised, and material from the xenografted tumours fixed in 10% neutral buffered formalin (Fisher Scientific, USA) and embedded in paraffin for routine histological analysis. The Lombardi Cancer Center Tissue Resource performed sectioning of the paraffin blocks, and tissue sections of 5 µm were mounted onto Pro-Bond +/-slides (Fisher Scientific, USA).

RNase digestion
Following pre-treatment all slides were rinsed in RNase buffer (10 mM HEPES, 20 mM NaCl, 1 mM EDTA) without enzyme (2 × 2 mins). Negative control sections undergoing RNase digestion were then overlayed with RNase Digest mixture containing RNase cocktail (8 mg/ml RNase A and 160 U/ml RNase T1 final volume, Bresatec, Australia) in RNase buffer, placed in a humid chamber and incubated at 37°C for 8 hrs. Untreated slides were overlayed in RNase buffer and also incubated at 37°C for 8 hrs.

DNase pre-treatment
Prior to DNase digest, all slides were rinsed in DNase buffer (0.1 M C 2 H 3 O 2 Na, 5 mM MgSO 4 pH 5.0, 2 × 2 min). DNase buffer (200 µl) containing 1.5 U/ml final volume of RNase-free DNase (Roche Diagnostics, USA) was added to each slide and these were incubated overnight at 37°C in a humid chamber. After incubation the slides were rinsed thoroughly in DNase buffer without enzyme, followed by washes in sterile water and dehydration through a graded ethanol series (50%, 70%, and 90% in 0.9% saline, 100% × 3 mins each).

Primers
Primers encoding MT1-MMP, MMP-1 and MMP-3 (Bresatec, Australia) were designed using the Amplimer program, product design checked by the Amplify v1.2 program, and BLASTed (NCBI, USA) for homology to the mRNA of these genes, with significant sequence homology demonstrated between the human and mouse MMPs examined. Primers encoding human β-actin were used as a positive control (Research Genetics, USA). Primer sequences are listed in Table 1. All primers were tested using traditional RT-PCR on mRNA extracted from MDA-MB-231 and MCF-7 HBC cells using the High Pure RNA Isolation Kit (Roche Diagnostics, USA). Reverse transcription was performed using 20 ng of isolated mRNA, by AMV reverse transcriptase (Promega Corporation, USA) at 50°C for 30 minutes, followed by PCR amplification using Taq DNA Polymerase (Perkin-Elmer Corporation, Applied Biosystems Division, USA). The linear phase of amplification was determined to be between cycles 20-25 using a standard two-step PCR cycle.

In situ-Reverse Transcription-Polymerase Chain Reaction (IS-RT-PCR)
Prior to the IS-RT-PCR procedure Gene-Frame gaskets (Advanced Biotechnologies, UK) were placed around the sections. IS-RT-PCR was performed on consecutive sections for each xenograft: section A) RNase/DNase negative  Reverse transcription was performed as previously reported [18] with the following modifications: dCTP concentration was 100 µM, biotin-dCTP 80 µM, MnCl 2 10 mM and 7.5 U of rTth DNA Polymerase (Perkin-Elmer Corporation, Applied Biosystems Division, USA) was used in the 30 µl mix added to each slide. Linear amplification efficiency was determined using traditional RT-PCR as outlined above, to be between cycles 20 and 25. For the amplification step the initial denaturation was at 95°C for 4 minutes, the MgCl 2 concentration was 15 mM, and 30 µl were added to each slide. Amplification was performed for 20 cycles.

Detection
Post IS-RT-PCR detection was performed as reported previously [21] with the following modifications: Incubation in the NBT/BCIP/ASB solution was for 15 minutes, slides were counterstained with eosin (Australian Biostain Pty Ltd, Australia) and mounted in Ultramount (Fronine, Australia). Consistency of observed gene expression in consecutive tumour sections demonstrated the reliability of the IS-RT-PCR technique and as expected, all cells were found to express the ubiquitous β-actin gene.

Signal quantitation
The level of signal intensity was assigned a value from zero to four (0+ to 4+). Slides showing no purple precipitation and only eosin counterstain were assigned a value of zero (0+), whilst cells stained with precipitate were assigned a value between one and four (1+ to 4+). The observed signal intensity was assessed based on the observed level of signal intensity and tissue morphology using light microscopy. Visual characteristics including the overall level of precipitation, direct comparison of the entire tissue area to compensate for background, and the number and intensity of cells exhibiting purple precipitation were also considered. In addition, a coded labelling system was used to ensure that slide examination was unbiased. In vitro signal intensity following exposure to Con A and cytokines was compared to the untreated or basal level allowing an estimate of up/down-regulation to be determined. The in vivo xenograft material was histologically assessed and scored from 0+ to 4+ (with 0 = no expression; 1+ = low expression; 2+ = moderate expression; and 3+ = strong expression). We compared the relative levels of gene expression of each MMP within the xenografts analysed to controls. Following independent examination of slides and sections by three individuals the signal level intensity was averaged. The photographs presented in Figures 1, 2 and 3 are of the same area within consecutive sections and are therefore a representation of the overall observed gene expression levels. For IS-RT-PCR two negative controls were used to ensure assay validity. The "No RT" control was used as a measure of the detection system specificity. Lack of precipitation in this control ensures that a positive signal is due to expression of the gene of interest. The RNase/DNase control was used for two reasons, to control for false positive results due to insufficient genomic DNA digestion, and to ensure that there was no signal due to any residual biotin-dCTP in the tissue section. Examples of these controls are given for the MDA-MB-231 cell line ( Figure 1A and 1B), where the only colouration in the slides is due to the eosin counterstain, and purple precipitation which would be indicative of positive gene expression signal, is not observed. Primers amplifying β-actin mRNA were used as a positive control for demonstration of mRNA preservation in the tissue, and the ability to detect gene expression. We found a clear positive signal, the purple precipitate localised to all cells examined, demonstrating β-actin expression. An example of this is given in Figure 1C for the MDA-MB-231 cell line.

Induction of in vitro MMP gene expression IS-RT-PCR results following Con A treatment
Expression levels of the three MMPs were examined with respect to β-actin in the MDA-MB-231, and MCF-7 cell lines, with or without exposure to Con A. In the MDA-MB-231 cell line, MT1-MMP expression was increased from 2+ to 4+ by Con A treatment, while MMP-1 and MMP-3 remained unchanged at 2+ following exposure to Con A. MT1-MMP induction by Con A in the MDA-MB-231 cell line is shown in Figure 1E and 1F. In the MCF-7 cell line, the MMP-1, MMP-3 and β-actin expression levels also remained unaffected by Con A at 2+, however, the MT1-MMP expression was decreased by Con A from 3+ in the absence of Con A to 1+ in the presence of Con A (not shown).  Table 2).

Discussion
The central aim of investigations into molecular carcinogenesis is the identification of gene products involved in cancer progression during which tumour cells acquire the capacity for invasion with resulting metastasis [27]. Metastasis is an active process involving the altered attachment of the tumour cell to the basement membrane, localised degradation of connective tissue, and migration through stromal tissue [28,29]. Such matrix degradation is mediated by members of several proteinase families (the serine, cysteine, aspartate proteinases and MMPs), with substantial tissue destruction being carried out by members of the MMP family [3]. The expression of members of the MMP family and their inhibitors, the TIMPs, have been examined in both physiological and pathological conditions by many methods both in vitro and in vivo, including the use of total RNA from tissues and cell lines, Northern analysis, RT-PCR in situ and standard in situ  , however Northern analysis of MCF-7 cells failed to find MT1-MMP in MCF-7 cells, whereas more invasive HBC cell lines showed MT1-MMP expression in concordance with their ability to be induced by Con A to activate MMP-2 [12,34,35]. MT1-MMP has also been shown to correlate with increased invasion and to enhance migration of MCF-10A epithelial cells [6,12,36], and transfection of MT1-MMP into MCF-7 cells stimulates higher migration and invasiveness [37,38], and also stimulates VEGF production, angiogenic stimulation, and xenograft take rate in nude mice [39,40]. Presumably the IS-RT-PCR method is more sensitive than Northern analysis, as may be expected. Indeed, we detected lower levels of MT1-MMP in the MCF-7 xenografts than in the majority of other xenografts derived from more invasive HBC cells. More recently, using isolated RNA for quantitative analyses, we were unable to detect MT1-MMP in MCF-7 derived xenografts, but consistent with our observations here, increasing levels of MT1-MMP was detected in MDA-MB-231 derived xenografts as compared to the parental cells [13]. The reasons for these differences are not clear and require further experimentation, but it is important to note that the cultured MCF-7 cells would have received estrogenic stimuli from both the phenol red and foetal calf serum in the medium [41].
MT1-MMP overexpression in the mammary gland results in abnormalities including hyperplasia, fibrosis, lymphocytic infiltration and adenocarcinoma [42], suggest-ing a pivotal role for MT1-MMP in carcinogenesis. However, ISH analyses of human breast carcinomas have primarily localised MT1-MMP mRNA expression to the stromal cells [11,29,[43][44][45], although one study found localised MT1-MMP in the tumour [46]. An immunocytochemical analysis showed expression of MT1-MMP in both tumour and stromal cells [47]. More recently, one study by Bisson et al, combining ISH and IHC data has colocalised MT1-MMP to the α-smooth muscle positive myofibroblast cells in close contact with tumour cells [48]. Our studies certainly show the potential for the cell lines examined to make MT1-MMP, even the relatively well-conserved cell lines such as MCF-7. Although it is acknowledged that in vitro propagation of cell lines may alter some genetic pathways, it is also apparent that cell lines and tumour samples have distinctive gene expression patterns in common [49]. Being ER-positive, and predominantly epithelial, MCF-7 cells far better represent what one may find clinically in breast cancers than the other cell lines examined here [50]. The invasive HBC cell lines show a mesenchymal-like phenotype that may have resulted from epithelial-mesenchymal transition (EMT) in breast carcinoma. EMT is thought to have occurred in the MDA-MB-231, MDA-MB-435 and Hs578T cell lines [51]. The demonstration of MT1-MMP expression observed in the stromal component is not totally unexpected. MT1-MMPs ability to stimulate invasion and metastasis in in vitro systems along with the estrogen receptor status of the cells is well documented as described above. Also, the basal levels observed between the MCF-7 and more invasive HBC cell lines may also be of impact in respect to the higher invasive potential of the cells.
MMP-1 mRNA was detected in the tumour cells of five of the six xenografts examined, the exception, surprisingly, being the MDA-MB-231 splenic metastasis. This is unexpected since MDA-MB-231 cell overexpress MMP-1 in culture [52], and these cells have a mutation/polymorphism in the promotor region which drives strong MMP-1 expression [18]. This transcriptional repression has not been found to occur extensively in breast cancer [18]. The reasons for suppression of the MMP-1 expression in splenic metastasis are not clear, but could include a strong repression, possibly due to altered signalling from the stroma. The splenic metastasis stroma was notably also lacking all three MMPs examined, but showed a strong βactin signal in both compartments. One cannot rule out the possibility of selective aberrant loss of MMP mRNA rather than the β-actin, however, this is unprecedented. Further analysis of splenic metastases of the MDA-MB-231 cells, and primary human tumours, would be warranted. MMP-1 has been previously demonstrated to be equally expressed in vitro in both MCF-7 and MDA-MB-231 cells [53], although in a similar study using Northern analysis we found selective expression in MDA-MB-231 cells but not MCF-7 [51]. Increased production and secretion of MMP-1 has been correlated with increased metastatic potential [52,54] [55], and also in response to fibronectin fragments [56], attesting to the potential regulation of this MMP by the microenvironment. In vivo, MMP-1 expression has been demonstrated in the stroma of nine of thirty-four breast carcinomas [11] when examined by ISH, localised to stromal constituents during tumour formation [57], and to be upregulated in the stromal component of ductal carcinomas when examined by ISH and IHC [57][58][59]. Thus, again, our observation that MMP-1 expression was absent in the stroma of three out of six HBC xenografts appears to contrast the clinical situation. This may reflect the increased sensitivity of IS-RT-PCR and may also reflect the stage in tumour formation of the MCF-7 and MDA-MB-231 mesenteric xenografts that we examined.
MMP-3 mRNA was localised to the tumour component in all six xenograft samples examined. In vitro, MMP-3 mRNA has been demonstrated in the MDA-MB-231 HBC cell line [55], where it is stimulated by fibroblasts via the extracellular matrix metalloproteinase inducer (EMMPRIN) [60], and can enhance tumourigenicity and migration [54]. In vivo, MMP-3 is centrally involved in mammary gland development [61], and has been demonstrated to promote tumour initiation and formation in the tetracycline-regulated mouse mammary model [62,63]. IHC and ISH in vivo analysis has demonstrated MMP-3 in both the tumour and stroma cellular compartments of both invasive and non-invasive tumours, with the level of stromal expression increasing with tumourigenicity [59,64,65], and in the extracellular matrix adjacent to breast tumours [66]. Our observation of MMP-3 gene expression in the stroma of four out of six of the xenografts examined, and in particular in the Hs578T xenograft, further support a role for this MMP in breast cancer.
Although HBC cell lines have been demonstrated to express certain MMPs, this is less evident in vivo where, as detailed above, the surrounding stromal cells contribute to much of the MMP activity. Important differences in MMP expression between the stromal cells recruited by each cell line and the tumour cells themselves, were detected in the current study. The MCF-7 xenograft stroma demonstrated low stromal MT1-MMP, MMP-1 and MMP-3 expression. MT1-MMP expression was observed at a lower level in the stroma associated with the MDA-MB-231 mesenteric metastasis. In the Hs578T xenograft stroma, MMP-1 was absent while MMP-3 was observed at a higher level in the stroma as compared to the tumour cells. No stromal expression of the three MMPs examined was observed in either the MDA-MB-231 spleen metastasis or the MDA-MB-435 xenografts. The level of stromal gene expression will depend on signals from the different HBC lines in the primary site, and also on the different responsivities in the different host sites for the splenic and mesenteric metastases. In regard to the primary site, overall recruitment of the stroma will differ among the HBC lines; however, β-actin mRNA detection was used to normalise the data. We did detect lower levels of β-actin in the Hs578T xenograft, and indeed higher levels of β-actin in the tumour cell component of the MCF-7 xenograft. However, this does not appear to have influenced the detection of the MMPs as indicated by the moderate levels of all MMPs examined in both compartments in the MCF-7 xenograft, and the moderate level of MMP-3 in the Hs578T tumour compartment.

Conclusion
These data combined indicate communication, and in particular HBC cell-line specific communication, between the epithelial and stromal cellular compartments for MMP production and utilisation. Considerable work has been performed in vitro to examine the effects of different HBC cell lines on MMP expression by co-culture with fibroblasts and other factors known to be produced by the tumour cells which mediate induction of MMPs (eg Con A, TPA, EMMPRIN) [12,15,21,[31][32][33]53,60]. These and other studies, implicate increasing MMP levels associated with both epithelial tumour cells and surrounding stromal cells, in association with increasing tumourigenicity and metastatic potential [12,15,21,24,53,54,60,67]. Our in vivo data indicate that, the MDA-MB-231, MDA-MB-435 and Hs578T HBC cell lines are able to influence MMP gene expression in the surrounding stroma. This is reinforced by a recent study demonstrating the influence of tumour associated fibroblasts on MCF-7 tumour cells to activate MMP-2 in vitro and increase tumour size in vivo [68]. These data combined indicate IS-RT-PCR may serve as a more sensitive, one-step procedure for the examination of MMP gene expression in human breast cancers.
contributed toward the design of the in situ sample preparation. R.E.I. contributed toward data management and manuscript finalisation. M.G.I. participated in the conception and design of the study. L.R.G. participated in the conception and design of the study and its coordination. All authors read and approved the final manuscript.