Studies on Multifunctional Effect of All-Trans Retinoic Acid (ATRA) on Matrix Metalloproteinase-2 (MMP-2) and Its Regulatory Molecules in Human Breast Cancer Cells (MCF-7)

Background. Vitamin A derivative all-trans retinoic acid (ATRA) is considered as a potent chemotherapeutic drug for its capability of regulating cell growth and differentiation. We studied the effect of ATRA on MMP-2 in MCF-7, human breast cancer cells, and the probable signaling pathways which are affected by ATRA on regulating pro-MMP-2 activity and expression. Methods. Gelatin zymography, RT-PCR, ELISA, Western blot, Immunoprecipitation, and Cell adhesion assay are used. Results. Gelatin zymography showed that ATRA caused a dose-dependent inhibition of pro-MMP-2 activity. ATRA treatment downregulates the expression of MT1-MMP, EMMPRIN, FAK, NF-kB, and p-ERK. However, expression of E-cadherin, RAR, and CRABP increased upon ATRA treatment. Binding of cells to extra cellular matrix (ECM) protein fibronectin reduced significantly after ATRA treatment. Conclusions. The experimental findings clearly showed the inhibition of MMP-2 activity upon ATRA treatment. This inhibitory effect of ATRA on MMP-2 activity in human breast cancer cells (MCF-7) may result due to its inhibitory effect on MT1-MMP, EMMPRIN, and upregulation of TIMP-2. This study is focused on the effect of ATRA on MMP, MMP-integrin-E-cadherin interrelationship, and also the effect of the drug on different signaling molecules which may involve in the progression of malignant tumor development.


Introduction
Retinoic acids, including vitamin A and its analogues, regulate the growth and differentiation. Retinoic acids suppress tumor formation in animals and have shown to be effective chemotherapeutic agents in many cancers. Predominantly all-trans retinoic acid (ATRA) and its stereoisomer 9-cis retinoic acid were found to be very potent metabolites of retinol exerting pleiotropic effects on many different biological processes. In contrast to all-trans and 13-cis, 9-cis retinoic acid has a greater effect on cell morphology. It has been reported that 13-cis retinoic acid decreases the development of primary tumors in patients with head and neck cancer, whereas ATRA provides a specific therapy for acute promyelocytic leukemia [1,2]. The antitumor effect of retinoids is most often attributed to the induction of differentiation, but these compounds were also shown to stop the growth of tumor cells by inducing apoptosis or accelerated senescence [3]. The ability of retinoic acid to induce differentiation in cancer cells suggests their potential role as a cancer chemotherapeutic agent.
Retinoids exert their effects by modulation of gene expression by two distinct classes of nuclear receptors: retinoic acid receptors (RAR α, β, γ) and rexinoid receptors (RXR α, β, γ). The receptors belong to the steroids or thyroid hormone super-family [4]. These receptors bear six domains: A/B for ligand-independent transactivation, C with two zinc fingers for DNA binding, D is a hinge region and E is responsible for dimerization with RXR, ligand binding, ligand-dependent transactivation, and association with corepressor (CoR) or coactivators (CoA) complex. The function of F is still not known [5]. Upon ligand binding, receptors bind to retinoic acid response elements (RAREs), specific DNA sequence, as receptor dimer. ATRA binds with high affinity (kd 0.2-0.4 nM) to RARs, but not to RXRs, whereas 9-cis retinoic acid binds to RXR with high 2 Journal of Oncology affinity [4]. RARs may exist as homodimer and can form heterodimer with RXRs as well and interact with RARE [6,7]. The receptor dimer complex remains silenced by the binding of corepressors. Retinoid ligand binding to the receptor dimer complex induces a conformational change that removes the corepressors and recruits coactivators, thus, facilitating initiation of transcription [8]. Two types of cellular retinoic acid binding protein CRABP I and II can act as transporter of retinoic acid to facilitate the transport of cytoplasmic retinoic acid into the nucleus leading to ligandreceptor interaction [4].
For tumor invasion and migration, the tumor cells and host coordinately regulate matrix degradation, cellcell attachments, and cell matrix attachment. Aberrant expression and activity of MMPs are associated with malignant development of tumor cells [9]. Most of the MMPs are secreted from cells in their inactive pro forms. The activation mechanism usually involves cleavage of the prodomain in the amino side of the conserved cystein residue followed by opening of the active site through the breakage of zinc-cystein bond [10]. The gelatinases are distinguished by the presence of an FN-like region, inserted in the catalytic domain. This insertion is important in substrate recognition [11]. Among MMPs gelatinases demonstrate specific cell-MMP interaction. For example, MMP-2 binds to integrin α v β 3 and MMP-9 to CD44 [12]. Expression of integrin α v β 3 on the cell surface supports maturation and activation of MMP-2 [13]. The membrane bound MMPs (MT-MMPs) are characterized by a c-terminal transmembrane portion and a short furin-like sequence between the pro and the catalytic domain. The furin site provides an alternative cleavage site for MMP activation [14].
MMP activity can induce tumor growth or survival, invasion, angiogenesis, and cell migration [15]. The expression of MMPs in tumor cells is regulated in a paracrine manner. Tumor cells produce factors, such as chemokines, cytokines, and the extra-cellular matrix metalloproteinase inducer (EMMPRIN), which in turn upregulate MMPs [16]. MMPs are the main group of proteolytic enzymes that facilitate tumor cell migration by degrading the basement membrane and other components of extra-cellular matrix (ECM) [17]. Beside degradation of ECM, tumor cell invasion and migration rely on the receptor-dependent attachment and release from the matrix and other cells. MMP activity directly modulates the cell-cell and cell-matrix attachment and regulates the process of cell invasion and migration [18]. Integrins play central role in anchorage-dependent growth, apoptosis, differentiation, and migration. Integrins differentially regulate the production of MMPs. In human melanoma cell lines, α5β1 integrin and α v β 3 integrin modulates release of MMP-2 and subsequent invasive behaviour [19]. Reduction of cell adhesion is of major importance in tumor metastasis and can be achieved by a variety of mechanisms affecting the E-cadherin-catenin complex. These include reduction or loss of E-cadherin expression, mutation of the genes of constituent molecules, redistribution of E-cadherin within the cell and shedding of E-cadherin, and so forth, [20].
In this communication, we report multifunctional effect of ATRA on (i) gelatinase-A expression and activity, (ii) various signaling molecules, and (iii) the integrin group of cell surface receptors in human breast cancer cells, MCF-7.

Cell
Culture. A375 human melanoma cell line and MCF-7 human breast cancer cell line was obtained from National Centre for Cell Sciences (NCCS), Pune, India. MCF-7 cells were grown and maintained in MEM containing 10% FBS and A375 cells were grown in DMEM supplemented with 10% FBS in a CO 2 incubator at 37 • C.
3.2. Drug Treatment. 7.5 mg Retinoic acid (purchased from Sigma, USA) was dissolved in 5 mL DMSO to prepare 5 mM stock solution. ATRA was added to the experimental dishes at concentrations of 10, 20, 30 μM.

Cell Viability
Assay. A375 cells (300,000/mL) and MCF-7 cells (300,000/1.5 mL) were grown in serum free culture medium (SFCM) in absence (Control) and in presence of 30 μM ATRA for 24 (in case of MCF-7) and 48 (in case of A375) hours. Control and ATRA-treated cells were collected by trypsinization. 10 μL of cell suspension (Control and ATRA-treated) in PBS were taken and 10 μL Trypan blue was added into it. Cell suspensions were mixed well and kept for 3-5 minutes. A number of stained and unstained cells were counted in a haemocytometer slide.

3.4.
Zymography. MCF-7 cells (300,000/1.5 mL) were initially grown in MEM supplemented with 10% FBS in petridishes, washed with serum-free culture medium (SFCM), and treated with increasing concentrations of ATRA (10 μM, 20 μM, 30 μM) for 24 hours in SFCM. Control cells were grown without ATRA but in presence of 1% DMSO (solvent for ATRA) for 24 hours and SFCM were collected. The matrix metalloproteinases were separated from SFCM using Gelatin Sepharose 4B beads and shaking for 2hours at 4 • C. The beads were washed x3 with Tris-buffered saline with Tween-20 (TBS-T) and suspended in 50 μL of 1X sample buffer (0.075 gm Tris, 0.2 gm SDS in 10 mL water, pH 6.8). The suspension was incubated for 30 minutes at 37 • C and then centrifuged at 3000 r.p.m. for 3 minutes. The supernatant was then subjected to zymography on 7.5% SDS-PAGE copolymerized with 0.1% gelatin. Gel was washed in 2.5% Triton-X-100 for 30 minutes to remove SDS and was then incubated overnight in reaction buffer (50 mM Tris-HCl pH 7, 4.5 mM CaCl 2 , 0.2 M NaCl). After incubation, the gel was stained with 0.5% Coomassie blue in 30% methanol and 10% glacial acetic acid. The bands were visualized by destaining the gel with water.

Enzyme Linked Immunosorbent Assay (ELISA)
. MCF-7 cells (300,000/1.5 mL) were grown in serum-free culture medium (SFCM) in absence (Control) and in presence of 30 μM ATRA (Experimental) for 24 hours. The culture supernatants were collected by centrifugation at 3000 r.p.m for 3 minutes. The wells of microtitre plate were coated in triplicate with 50 μL culture SFCM and with 50 μg protein from both control and experimental set and kept at 4 • C overnight (plate was wrapped in wrap to prevent evaporation). Blank wells (with buffer in which samples are suspended) were also prepared. Next day wells were washed with blocking buffer (1% BSA in PBS) to block nonspecific binding sites and incubated for 1 hour at 37 • C. Then the wells were washed thrice with Washing Buffer (0.5% NP-40 & 0.5% BSA dissolved in PBS). Anti-TIMP-2, anti-E-cadherin, anti-RAR, and anti-CRABP primary antibody solution (1 : 1000 dilution) was added to the wells and incubated at 37 • C for 1 hour. Wells were washed thrice with Washing Buffer. Respective second antibody solution (1 : 1000 dilution buffer) was added to wells and incubated at 37 • C for 1 hour. Wells were washed six times with Washing Buffer (3-5 minutes per wash). Substrate (TMB) was added to the wells (in darkness) and kept as long as required (i.e., until color developed begins to become too intense). Then 1 M H 2 SO 4 stop solution was added and reading was taken in ELISA reader at 450 nm.

Immunoblot Assay of MT1-MMP, EMMPRIN, FAK and
NF-κB. MCF-7 cells (300,000/1.5 mL) were grown in serum free culture medium (SFCM) in absence (control) and in presence of 30 μM ATRA for 24 hours. The respective cells were collected. Cell extraction was carried out using cell extraction buffer (37.5 mM Tris, 75 mM NaCl and 0.5% Triton-X-100) and the protein content of the extracts were estimated by Lowry's method. Equal amount of protein was taken and incubated with 1X sample buffer for 30 minutes followed by 5-8 minutes incubation with 0.1 volumes βmercaptoethanol at 80-90 • C. Samples were then subjected to electrophoresis on 7.5% SDS-PAGE. The proteins were transferred on to nitrocellulose membranes by Western Blot at 300 mA for 3 hours. The membranes were blocked with 1% BSA and subsequently washed thrice with TBS-T. The immunoblots were reacted with anti-MT1-MMP, anti-EMMPRIN, anti FAK, anti-NF-κB and anti-Ig-G antibodies, respectively, (1 : 1000 dilution) for 1.5 hours at 37 • C followed by incubation with respective alkaline phosphatase coupled second antibodies. Bands were developed using NBT-BCIP as substrate.

Immunoblot Assay of ERK, Phospho-ERK and VEGF
by Immunoprecipitation. MCF-7 cells (300,000/1.5 mL) were grown in serum-free culture medium (SFCM) in absence (Control) and in presence of 30 μM ATRA for 24 hours. The cells were collected. Cell extraction was carried out using the cell extraction buffer and the protein content of the extracts was estimated by Lowry's method. Equal amount of protein was taken from each extract and ERK, VEGF was immunoprecipitated from the extracts using anti-ERK, anti-VEGF, and anti-Ig-G antibodies and protein-G agarose beads and shaking overnight at 4 • C. The resultant immunecomplex was washed thrice in PBS, suspended in 1X sample buffer and incubated at 37 • C for 30 minutes, followed by incubation with 0.1 volumes of β-mercaptoethanol for 5-8 minutes at 80-90 • C. Samples were then subjected to electrophorese on 7.5% SDS-PAGE. The proteins were transferred on to nitrocellulose membrane by Western Blot. The membranes were blocked with 1% BSA and subsequently washed thrice with TBS-T. The immunoblots  (Figure 1(c)). The culture supernatants were collected and MMPs were concentrated using Gelatin-sepharose 4B beads. Gelatinases were eluted from bead with sample buffer and subjected to zymography on 7.5% SDS-PAGE copolymerised with 0.1% gelatin. The zymogram was treated with 2.5% triton X-100 for 30 minutes followed by incubation in reaction buffer for 20 hours and stained with coomassie blue. The quantitative measurement of the zymogram (Figure 1 were reacted with anti-ERK antibody, anti-phospho-ERK antibody, anti-VEGF, and anti-Ig-G antibodies, respectively, in 1 : 1000 dilution, and kept at 37 • C for 1.5 hours.
After washing with TBS-T, membranes were incubated with alkaline phosphatase coupled respective second antibodies (1 : 1000 dilutions). Bands were visualized using NBT/BCIP as substrate.

Effect of ATRA on Cell Viability.
Cell viability in MCF-7 (Figure 1(a)) cells before and after ATRA treatment was checked using trypan blue stain (0.4%). In case of control 113 cells (average of 3 experiments) in 0.1 μL (Control) were counted as viable and in treated average of 3 experiments showed 110 cells/0.1 μL (Treated) were viable. However, T-test showed that there was no statistically significant difference (P = .085) between two groups. Therefore, no changes were observed in cell viability of control and ATRAtreated. Figure 1

Effect of ATRA on TIMP-2 Protein Expression.
When the culture supernatant from both control and 30 μM ATRAtreated MCF-7 cells were assayed for MMP-2 and tissue inhibitor of MMP-2 (TIMP-2) protein by ELISA (Figure 2), it was found that ATRA substantially increased TIMP-2 protein level. P = .000144 (P < .05) indicates data obtained are highly significant.   (Figure 3(b)).

Effect of ATRA on MMP-2 Gene Expression.
To analyze whether ATRA affects MMP-2 by regulating at transcription levels, we studied mRNA expression for MMP-2 by RT-PCR. Figure 4     Binding of cells to fibronectin decreased significantly upon ATRA treatment for 24 hours whereas binding of cells to fibronectin remained almost unaltered for 12 hours ATRAtreated cells (result not shown).

Effect of ATRA on α5 and β1 Integrin Receptor Expression.
RT-PCR analysis ( Figure 5(b)) of control and ATRA-treated cells showed significant down regulation of α5 expression in ATRA-treated cells. β1 expression was also downregulated upon 30 μM ATRA treatment for 24 hours.

Effect of ATRA on RAR and CRABP Expression.
ELISA was performed to observe the status of RAR (Figure 8(a)) and CRABP (Figure 8(b)) upon ATRA treatment and it was found that RAR and CRABP expression increased appreciably in 30 μM ATRA-treated (E) MCF-7 cells, compared to that of the control cells(C). P = .028 (P < .05) indicates that differences between control treated cells are statistically significant.

Discussion
In this communication we report the effect of all-trans retinoic acid (ATRA) on the expression and activity of matrix metalloproteinases and possible mechanism that may influence the drug to exert its effect. Treatment of MCF-7 cells with ATRA showed appreciable inhibition of pro-MMP-2 enzyme expression and activity. Treating cells with 30 μM ATRA showed significant induction of the TIMP-2 protein expression in the SFCM. The activity of MMPs in extra cellular space is specifically inhibited by tissue inhibitors of metalloproteinases (TIMPs). TIMPs bind to the highly conserved zinc binding site of active MMPs at molar equivalence. Over expression of TIMP-2 can inhibit the activity of MMP-2 [21] and it also inhibits the invasive and metastatic behaviors of cancer cells. Braunhut et al. showed that retinoic acid upregulated TIMP-2 expression in endothelial cells [22]. Other studies also showed a pronounced induction of the TIMP-1 protein by vitamin A [23,24].  [25,26]. Our results demonstrate ATRA mediated reduction of MT1-MMP protein expression which may cause reduced activation of pro-MMP-2 in ATRA-treated MCF-7 cells. Extracellular matrix metalloproteinase inducer (EMMPRIN), a 58 kD type-I trans membrane protein is known to play role in ECM remodeling through the activation of MMP production [27]. EMMPRIN exhibits potential to induce MMP-2 expression [28]. In this study downregulation of EMMPRIN expression by ATRA may cause reduced pro-MMP-2 activity in the SFCM of ATRA-treated MCF-7 cells.
The interaction of cells with adhesion protein in the ECM provides signals which affect the morphology, motility, gene expression and survival of adherent cells [29]. The relationship that exists between integrin receptor, fibronectin, and MMP expression is of particular interest. Fibronectin is a prototype cell adhesion protein. It is present as a polymeric fibrillar network in the ECM and as soluble protomer in body fluids. Two regions in each fibronectin subunit possess cell binding activity: III 9-10 and III 14-V . The RGD motif in fibronectin is located in III 10 and is the most important recognition site for most of the known integrins [30]. Integrins, a family of cell surface receptors, mediate attachment of the cells to ECM and initiate a series of signaling events that ultimately activates proteases like gelatinases. Integrins mediate both cell-cell and cellsubstratum adhesion and signaling [29,31]. Binding of integrin receptors to fibronectin can initiate organization of the cytoskeleton and focal contacts and signals which affect gene expression. Fibronectin through interaction with its corresponding α 5 β 1 integrin receptor regulate the activity and expression of MMP-2 and MMP-9 [32][33][34]. On binding to ECM molecules, integrins are thought to undergo a conformational change, which allows the intracellular domain of their β subunit to interact with focal adhesion proteins. FAK appears to localize to nascent focal adhesions. Upon activation, FAK combines with Src and FAK/Src signaling pathway and promotes cell migration [35,36].
Our study demonstrates that treatment of MCF-7 cells with ATRA downregulates the expression of α5 and β1 which reflects the result of cell adhesion assay, showing reduced tumor cell binding to ECM protein fibronectin after ATRA treatment in a time dependent manner. Our earlier study showed that ATRA treatment downregulates α 5 and β 1 subunit expressions of α 5 β 1 receptor in SiHa cells [37]. Focal adhesion kinase (FAK) was first shown to be linked to integrin signaling when it was identified as a major substrate for tyrosine phosphorylation. It has been suggested that Rho, a Ras-like GTP binding protein that regulates the formation of focal adhesions and actin stress fibers, may be involved in the activation of FAK. Thus, multiple cancer cell survival stimuli that signal through diverse pathways converge to induce the tyrosine phosphorylation of a common substrate, FAK [38]. Furthermore, FAK induces MMP-2 secretion from cell interior to its medium [39,40]. ATRA treatment caused a significant decrease in FAK expression in MCF-7 cells. It may contribute to the cell motility and decreased gelatinase activity by regulating the secretion of MMP-2 in the SFCM. Alterations in the adhesion properties of neoplastic cells may play a pivotal role in the development and progression of the malignant phenotype in a range of tumor types. Adhesion molecules are intimately involved in the control of such processes as morphological differentiation, cellular proliferation, and invasion and colonization at distant organs [41]. E-cadherin dependent cell-cell adhesion is important for the maintenance of epithelial structural integrity and the loss of E-cadherin expression has been shown to correlate with increased invasive potential of both carcinoma cell lines and human tumor samples [42]. Recent data demonstrate that tumor-associated MMPs can modulate cell-cell adhesion by cleaving E-cadherin [43]. Evidence is emerging that there may be cross-talk between the cadherins and the integrins. Anticadherin antibodies have been shown to prevent the loss of α6 and β1 integrins in terminally differentiating keratinocytes [44]. Positive expression of E-cadherin could decrease cell adhesion to fibronectin partially through transcriptional inhibition of α5β1 integrin gene [45]. Our results showed that ATRA treatment up-regulated E-cadherin expression in MCF-7 cells which could downregulate the α5 and β1 expression and possibly attributed to inhibition of cell-matrix adhesion and therefore in turn inhibit pro-MMP-2 activity. Overexpression of E-cadherin decreased MMP-2 activity in prostate cancer cells and MT1-MMP in squamous cancer cells [46]. Therefore, overexpression of E-cadherin in ATRA-treated MCF-7 cells may downregulate pro-MMP-2 activity in the SFCM of ATRA-treated cells.
In our study, we found decreased expression of NF-κB in ATRA-treated MCF7 cells. NF-κB is active in nucleus and is inhibited through its sequestration in the cytoplasm by inhibitor of κB (IκB). IκB kinase (IKK) can degrade IκB and thus release NF-κB from inhibition [47]. NF-κB mediates MT1-MMP induction and pro-MMP-2 activation. Osteopontin, a membrane protein induces NF-κB mediated pro-MMP-2 activation through IκBα/IKK phosphorylation [48]. Therefore, by suppressing the NF-κB protein expression  (Figures 7(b) and 7(c)) and with anti-VEGF & protein G Agarose bead (Figure 7(d)), keeping the samples for overnight at 4 • C with shaking. In each case the resultant immune complex was washed thrice with PBS and the respective protein bound with antibody were eluted from the agarose bead using 1X sample buffer. Samples were then incubated in β-mercaptoethanol for 10 minutes at 90 • C. Samples were subjected to electrophoresis on 7.5% SDS-PAGE. The proteins were transferred to nitrocellulose membrane by Western Blot. The membranes were incubated with anti-ERK (Figure 7(b)), anti-phospho ERK (Figure 7(c)), and anti-VEGF antibody (Figure 7(d)), respectively. The immunoblots were then incubated with alkaline phosphatase-coupled secondary antibodies and bands were visualized by NBT/BCIP substrate. Ig-G (Figure 7(e)) was used to confirm equal loading. Quantitative measurements of immunoblots were done by using Image J Launcher (version 1.4.3.67). (C) represents the expression of respective proteins in control cells whereas (E) represents expression in 30 μM ATRA treated cells.
ATRA may downregulate pro-MMP-2 activation. Furthermore, constitutive activation of NF-κB may very well play an anti-apoptotic role in the breast and prostate carcinoma cells [49]. The inhibition of constitutive NF-κB activation by reducing NF-κB expression in ATRA-treated MCF-7 cells may therefore enhance the basal apoptotic rate in these cells. Dual specificity phosphatase 6 (DUSP6) dephosphorylates activated ERK and blocks the growth stimulatory signals. DUSP6 and RGS16 (regulator of G-protein signaling 16) transcriptional upregulation inhibits tumor cell proliferation and ERK phosphorylation and serves as a negative regulatory mechanism to prevent further ERK phosphorylation. Liu et al. showed RA induced upregulation of DUSP6 & RGS16 [50]. Our study demonstrates decreased p-ERK level in ATRA-treated cells. Reduced ERK phosphorylation may result due to ATRA induced transcriptional activation of DUSP6 and RGS16. Activation of ERK plays an essential role in the expression of MT1-MMP [51,52]. Reduced activation of ERK may cause lesser expression of MT1-MMP in ATRAtreated cells.
Angiogenesis is known to be one of the most important aggravating steps in the progression of cancer. Various angiogenic factors are known to be upregulated in cancer. Some of the proangiogenic factors such as thrombin, H 2 O 2 , hepatocyte growth factors, vascular endothelial growth factor (VEGF) are known to have contribution to the activation and release of MMP-2 [53,54]. Upon ligand (VEGF) binding to its receptor VEGFR-2 tyrosine kinase signaling cascade gets stimulated. As a result of this signaling cascade, proliferation/survival factors (bFGF), migration factors (ICAMs/VCAMs/MMPs), and vessel permeability factors (eNOs) are produced [55]. Downregulation of VEGF expression in ATRA-treated MCF-7 cells may contribute to the reduced pro-MMP-2 activity in treated cells.
All-trans retinoic acid after entering into the cell, its entry from cytoplasm to nucleus is facilitated by CRABP I/II. In the nucleus it binds with its specific receptor RAR/RXR and exerts its effect by regulating various gene transcriptions. Several studies suggested that RAR/RXR and CRABP expression got increased when cells received ATRA treatment [56,57]. Our study also demonstrates up-regulation of RAR and CRABP expression upon ATRA treatment in MCF-7 cells.
In conclusion, the present study demonstrates the inhibitory effect of ATRA on pro-MMP-2 activity and the possible molecular mechanisms. Our experimental findings strongly indicate that ATRA induced inhibition of pro-MMP-2 activity may result due to increased expression of TIMP-2, and downregulation of FAK, MT1-MMP, EMM-PRIN, and NF-kB expression. ATRA may interfere with the integrin down-stream signaling altering the expression and activity of pro-MMP-2.