Regulation of Matrix Metalloproteinase Genes by E2F transcription factors: Rb-Raf-1 interaction as a novel target for metastatic disease

Abstract

The Rb-E2F transcriptional regulatory pathway plays a major role in cell cycle regulation, but its role in invasion and metastasis is less understood. We find that many genes involved in the invasion of cancer cells, such as matrix metalloproteinases, have potential E2F binding sites in their promoters. E2F binding sites were predicted on all 23 human MMP gene promoters, many of which harbored multiple E2F binding sites. Studies presented here show that MMP genes such as MMP9, MMP14, and MMP15 which are overexpressed in non-small cell lung cancer (NSCLC) have multiple E2F binding sites and are regulated by the Rb-E2F pathway. Chromatin immunoprecipitation assays showed the association of E2F1 with the MMP9, MMP14, and MMP15 promoters and transient transfection experiments showed that these promoters are E2F responsive. Correspondingly, depletion of E2F family members by RNAi techniques reduced the expression of these genes with a corresponding reduction in collagen degradation activity. Further, activating Rb by inhibiting the interaction of Raf-1 with Rb using the Rb-Raf-1 disruptor RRD-251 was sufficient to inhibit MMP transcription. This led to reduced invasion and migration of cancer cells in vitro and metastatic foci development in a tail vein lung metastasis model in mice. These results suggest that E2F transcription factors may play a role in promoting metastasis through regulation of MMP genes, and that targeting the Rb-Raf-1 interaction is a promising approach for the treatment of metastatic disease.

Introduction

The retinoblastoma tumor suppressor protein, Rb, together with the E2F transcription factors are the main regulators of the mammalian cell cycle (1). Rb physically interacts with E2Fs 1-3 via their transcriptional activation domain, repressing their transcriptional activity (2). In response to mitogenic signaling, Rb is inactivated in the G1 phase of the cell cycle in multiple waves of phosphorylation by cyclin dependent kinases 2, 4 and 6, leading to its dissociation from E2Fs 1-3 (3). This facilitates the expression of various genes that are necessary for DNA synthesis and cell cycle progression, including cyclin E, dihydrofolate reductase, DNA polymerase α etc. (4). Not surprisingly, oncogenic mutations target the Rb-E2F pathway to promote cell proliferation (5). The Rb gene itself is mutated in a variety of cancers, while mutations in signaling molecules like K-Ras, p16INK4 and PTEN that affect Rb function are prevalent in almost all cancers (4-7). This indicates a major role for the Rb-E2F pathway in cell cycle progression and oncogenesis. Further, E2Fs are known to be important for proper execution of development, differentiation, apoptosis, and DNA damage repair programs (8, 9), establishing a larger role for E2Fs in the biology of normal mammalian cells and their transformation into cancer cells.

Our earlier studies had shown that the kinase Raf-1 physically interacts with Rb early in the cell cycle, facilitating Rb phosphorylation (10, 11). Disruption of the interaction of Raf-1 with Rb using the small molecule disruptor RRD-251 prevented Rb phosphorylation, cell cycle progression, angiogenesis and tumor growth in mouse models (12-14). It was found that RRD-251 could inhibit the expression of E2F-regulated proliferative promoters like Cdc25A and TS. Interestingly, recent studies from our lab have demonstrated that E2F1 could induce VEGF receptors, FLT-1 and KDR, indicating a role for E2F1 in tumor angiogenesis as well (15). Given this background, attempts were made to assess whether E2Fs can also affect the expression of genes involved in cell invasion and cancer metastasis. Towards this purpose, we used Genomatix MatInspector software to analyze the promoters of matrix metalloproteinase genes, which remodel the extracellular matrix and facilitate cell invasion and metastasis (16). We find that most MMP promoters have multiple E2F binding sites; data presented here show that three MMPs that are overexpressed in NSCLC, namely MMP9, MMP14, and MMP15 are in fact E2F regulated. Supporting this contention, the Rb-Raf-1 disruptor, RRD-251, which prevents Rb phosphorylation and inhibits E2F1-mediated transcription, could inhibit the transcription of MMP genes. In addition, RRD-251 could prevent invasion in vitro, and decrease colonization of the lung in an in vivo tail vein metastasis model. These results suggest that the Rb-E2F pathway contributes to the expression of MMP genes and that targeting this pathway might be a potential avenue to combat metastatic disease.

Materials and Methods

Cell Lines and Reagents

A549 NSCLC cells were cultured in F12K medium with 10% serum (Cellgro). MDA-MB-231 and MDA-MB-435 human breast cancer cells were cultured in DMEM with 10% serum. H1650 human NSCLC cells were grown in RPMI with 10% serum. A549 cells stably expressing the firefly luciferase gene (A549-luc) were obtained from Caliper and grown in RPMI with neomycin (200 ng/mL). For treatment with RRD-251, cells were rendered quiescent by serum starvation for 18 hours, and then grown in 10% serum-containing in F12K medium with RRD-251. The Rb-Raf disruptor RRD-251 was prepared as described and was >99% pure as analyzed by HPLC (12).

Cloning of MMP promoters

DNA was extracted from primary aortic endothelial cells using standard protocols (10). Primers spanning 2kb of the MMP9 and MMP15 promoter were used to PCR amplify the fragment with Hotmaster Taq (5-Prime). Primer sequences were:

• 5’-TACGGTGCTTGACACAGTAAATC-3’ (MMP9 forward),

• 5’-CTGACTGCAGCTGCTGTTGTGG-3’ (MMP9 reverse),

• 5’-GCTACTTTCCTTCACTGAACAGG-3’ (MMP15 forward),

• 5’-CGAGTGAAGTGCGACAGTGCGGCC-3’ (MMP15 reverse).

The fragments were then subcloned into pCR2.1 using TA cloning (Invitrogen). The plasmids were digested with Kpn1 and Xho1 and ligated into pGL3-basic luciferase vector (Promega). The MMP14 promoter was a kind gift from Dr. Jouko Lohi at The University of Helsinski (17).

Transient transfections and Luciferase Assays

A549 cells were transfected with 0.5 μg of MMP reporters along with 1 μg E2F1, 2 μg of Rb-Large Pocket or full length, and 2 μg Raf-1 full length expression vector using Fugene HD reagent in a ratio of 4 μl Fugene to 2 μg plasmid (Roche). Cotransfection with 0.5 μg of pRL construct containing Renilla reniformis luciferase gene used as normalizing control. Luciferase assays were performed using Dual Luciferase Assay System (Promega)(15). Relative luciferase activity was defined as the ratio of firefly luciferase activity to Renilla luciferase activity. Error bars represent standard deviation of three experiments.

Gelatin Zymography

Media was concentrated using 7 kD molecular weight cut off protein concentrators at 4°C (Pierce) and subjected to electrophoresis on 8% polyacrylamide gels containing 2 mg/mL bovine skin gelatin (Sigma). Gels were washed twice with 2.5% Triton-X100, and then incubated for 24 hours at 37° C in Tris-HCl buffer (150 mM NaCl, 10 mM CaCl₂, 50 mM Tris-HCl pH7.6 and 0.05% NaN₃). Gels were stained with 0.2% Coomassie Brilliant Blue and destained (30% methanol, 10% glacial acetic acid, and 60% H₂0) until gelatinolytic bands could be detected. Gelatinolytic signals were quantified by densitometry.

Chromatin immunoprecipitation (ChIP) assays

Chromatin immunoprecipitation assays were performed on asynchronous A549 cells as previously described (18). Immunoprecipitations were done using polyclonal antibodies for E2F1-5 and Rb (Santa Cruz Biotechnology); a Rabbit anti-mouse secondary antibody (Pierce) was used as the negative control. The interaction with specific promoters was detected using PCR with primer sequences detailed in Supplemental Table 1.

siRNA transfections and Real-time PCR

For siRNA transfections 100 pmol of siRNAs (Santa Cruz) with Oligofectamine were added to cells. For real-time PCR, total RNA was isolated using RNeasy miniprep kit (QIAGEN) following manufacturer’s protocol, followed by first-strand cDNA synthesis using iScript cDNA synthesis kit (Bio-Rad). Data was analyzed by ΔΔCT method, where gene of interested was normalized to 18s RNA, then compared to the non-targeting siRNA control sample. Error bars represent the standard deviation of three independent experiments.

Invasion Assays

Boyden Chamber assays were used to assess the invasive ability of A549, and MDA-MB-231 cells as described previously (19, 20). Briefly, the upper surface of the 6.5 mm filters (Corning) were coated with collagen (100 μg/filter) and Matrigel (BD Bioscience) (50 μg/filter). Twenty thousand cells were plated in the upper chamber with 0.1% bovine serum albumin (Sigma). Media containing 20% fetal bovine serum was placed in the lower well as chemoattractant. The cells which invaded through the filters were quantified by counting three fields under 40× objective magnifications.

Would healing assays

One hundred thousand A549 cells were plated in a 6-well plate (Falcon). The cells were scratched with a sterile 200 μl pipet tip in three separate places in each well, stimulated with serum, and the same area was examined after 24 hours using phase contrast microscopy (21).

Collagen Degradation Assays

Collagen Degradation Assays were carried out as previously described (22). First, 1 mL of type IV collagen was mixed with 7 mL of 13 mM HCl, then neutralized with a buffer containing 0.2 M NaP_i, 16.6 mL 5M NaCl, 80 mL 0.1 N NaOH. Seven hundred μl of this solution was added to 12 well tissue culture plates to obtain a final concentration of 1 μg/mL of collagen. Plates were incubated at 37°C for 2 hours to polymerize. Thirty-five thousand CCL-210 cells were placed in a 40 μl button and were left to attach 5 hours at 37°C. Complete media was added, and after 4 days, cells were trypsinized, and the remaining collagen was stained with Coomassie Brilliant Blue for 15 minutes, then destained (30% methanol, 10% glacial acetic acid, 60% H₂0). Images were taken using Epson Perfection V700 Photo Scanner.

Proliferation Assays

Bromodeoxyuridine (BrdUrd) labeling kits were obtained from Roche. Cells were plated in poly-D-lysine coated chamber slides at 10,000 cells/well and serum starved for 24 hours. Cells were then stimulated with serum in the presence or absence of 10 μg/mL Mitomycin C for 3 hours, then incubated in complete media. S-phase cells were visualized by microscopy and quantified by counting three fields of 100 in quadruplicate.

In Vivo Metastasis Assay

Five million A549 cells stably expressing firefly luciferase (A549-Luc-C8) (Caliper) were injected into the lateral tail vein of 5-week-old female SCID-beige mice under an IACUC approved protocol. Mice were given DPBS: DMSO vehicle control or RRD-251 diluted with DPBS: DMSO once/day. For bioluminescence imaging, mice were anesthetized and 30 mg/Kg of D-luciferin in PBS was administered by intraperitoneal (i.p.) injection. Ten minutes after injection, bioluminescence was imaged with a charge-coupled device camera (Caliper). Bioluminescence images were obtained with a 15 cm field of view, binning (resolution) factor of 8, 1/f stop, open filter, and an imaging time of 30 s to 2 min. Bioluminescence from relative optical intensity was defined manually, and data were expressed as photon flux (photons·sec⁻¹·cm⁻²·steradian⁻¹) and were normalized to background photon flux over a mouse that was not given an injection of luciferin.

Tissue Processing and Immunohistochemical Staining

Lungs were fixed in 10% neutral buffered formalin after necropsy, before processing into paraffin blocks. Paraffin sections (5-Am thick) were rehydrated and processed using hematoxylin and eosin staining using standard techniques.

Statistical analysis

Statistical analysis was performed using one-tailed Student’s t test. Values were considered significant when the P value was <0.05.

Results

MMP9, MMP14, and MMP15 promoters are responsive to E2F1 and Rb

Microarray studies had suggested that MMP genes may be E2F responsive (23, 24) and to explore this possibility, we examined the promoter region 2 Kb upstream of the transcription start site of 23 MMP genes using MatInspector (Genomatix) program. Putative E2F binding sites were observed on the promoters of all 23 MMP genes examined (Supplementary Table 2). Since MMP9, MMP14, and MMP15 are overexpressed in a variety of metastatic tumors including non-small cell lung cancer (NSCLC), these promoters were studied further. MMP9, MMP14, and MMP15 promoters had three, five, and four E2F binding sites respectively upstream of the TSS within the 2 Kb regions. In addition, the MMP14 promoter had two E2F binding sites downstream of TSS (Fig. 1A).

Figure 1.

(A) Schematic representation of MMP9, MMP14, and MMP15 promoters showing potential E2F binding sites as diamond/circle symbols. The arrows represent the position of primers spanning E2F binding sites tested in ChIP assays. (B) ChIP assays conducted on asynchronously growing A549 cells using the indicated antibodies. Sonicated genomic DNA is used for input. The numbers indicate the position in the promoter, in respect to TSS, where a putative E2F binding site was identified. C-Fos was used as a negative control, whereas DHFR, Cdc6, and Cdc25a are used as positive controls. Irrelevant antibody was used (Ir Ab) as a negative control for IP. (C) Transient transfection experiments in A549 and H1650 cells showed that E2F1 could significantly (***, P<0.001; **, P<0.005; * P<0.05) induce MMP9, 14 and 15 promoters, and this was repressed by Rb large pocket or full length Rb; co-transfection of Raf-1 could reverse Rb-mediated repression. Control lanes had the reporter with empty vector. (D). Western blot shows overexpression of E2F1 upon transfection, compared to empty vector transfected A549 and H1650 cells. (E) A549 and H1650 cells transfected with E2F1 have increased levels of MMP9 (A549 P=0.019*; H1650 P<0.001**) gelatinase activity as seen by coomassie stained zymography (inverted image).

ChIP assays were conducted on asynchronously growing A549 cells to assess whether E2F1 and Rb associate with these promoters. The location of primers used is shown in Figure 1A as arrows. As shown in Fig. 1B, there was a significant amount of E2F1 bound to MMP9, MMP14, and MMP15 promoters, and at least two E2F binding sites recruited E2F1 on each promoter. As in the case of E2F-regulated proliferative promoters, Rb could also be detected on most MMP promoters. E2F1 was present on three positive control promoters, DHFR, Cdc6, and Cdc25a. There was no Rb or E2F1 present on the unrelated c-Fos promoter, which was the negative control. There was no DNA associated with an IP done with an irrelevant antibody, further establishing the specificity of the assay. This experiment suggests that the E2F sites present on these MMP promoters can recruit E2F1 and Rb.

Experiments were done to assess whether these E2F binding sites were functional. Towards this purpose, A549 cells were transiently transfected with luciferase reporter constructs driven by MMP9, MMP14, and MMP15 promoters. It was found that co-transfection of E2F1 led to a significant induction of all the three promoters (Figure 1C); further, co-transfection of the large pocket region of Rb (Rb-LP) or the full length Rb (Rb-FL) could repress the E2F1-mediated induction. Consistent with previous studies on proliferative E2F-target genes (10, 11), over-expression of Raf-1 could relieve the repression mediated by Rb. Taken together, these results suggest that the Rb-E2F pathway might regulate MMP9, MMP14, and MMP15 expression.

Gelatin zymography was used to determine whether overexpression of E2F1 enhances MMP9 gelatinase activity. Consistent with the transfection data, MMP9 activity was increased 1.78 fold in A549 cells and 2.54 fold in H1650s overexpressing E2F1 (Figure 1D-E). This suggests that the endogenous MMP9 promoter is responsive to E2F1 overexpression, leading to MMP9 secretion in cell lines.

Previous studies had demonstrated that proliferative promoters are induced mainly by the transcriptionally active family members, E2Fs1-3. To determine if MMP genes are also regulated exclusively by E2F1-3, ChIP assays were performed on asynchronous A549 cells. While the proliferative dihydrofolate reductase (DHFR) promoter recruited only E2Fs 1 and 3, E2Fs1-5 were recruited to the promoters of both MMP9 and MMP15 (Figure 2A); E2Fs1-4 were recruited to the MMP14 promoter. Consistent with the ChIP assay data, transient transfection experiments on A549 cells showed that MMP promoters are significantly induced by E2F1-5 whereas DHFR is significantly induced by E2F1-3 (Fig 2B, C). This data suggests that MMPs may be a new class of E2F target genes, which can positively respond to E2Fs 1-5.

Figure 2.

(A) MMP9, MMP14, and MMP15 are responsive to E2F1-5. ChIP assays were carried out on asynchronously growing A549 cells. One binding site in each promoter was analyzed: -1920 to -1904 in MMP9; -1667 to -1532 in MMP14; and -1625 to -1609 in MMP15. (B-C) Transient transfection experiments in A549 cells showed that DHFR is significantly induced by E2F1-3 (***, P<0.001; **, P<0.005; * P<0.05) though no significant difference with E2F4 (P=0.18) or E2F5 (P=0.47), whereas E2F1-5 could significantly induce all MMP promoters.

E2F1 and E2F3 are required for MMP gene expression and collagen degradation

Given that E2F1-5 could induce MMP9, MMP14 and MMP15 promoters in transient transfections, attempts were made to assess whether E2Fs regulate the expression of their endogenous promoters in NSCLC cells. Towards this purpose, A549 and H1650 cells were transfected with 100 pMol of siRNAs to E2F1-5, or a non-targeting control siRNA. Transfection with siRNA targeting E2F1, E2F3, or E2F5 significantly reduced the expression of MMP9 and MMP14 mRNA as seen by quantitative RT-PCR, whereas E2F2 and E2F4 had no effect (Fig. 3A). Since E2F5 was not detected on the MMP14 promoter in the chip assay of site -1657, it is possible that E2F5 is affecting the promoter through a different E2F binding site. Surprisingly, MMP15 mRNA levels were not changed when E2Fs were depleted in either A549 cells or H1650 cells (Fig 3B). DHFR mRNA levels were reduced when E2F1 or E2F3 were depleted, correlating with their binding to DHFR promoter in ChIP assays (Fig. 2A); depletion of E2F2 had no effect, consistent with its lack of binding in the ChIP assay; nevertheless, it could induce this promoter when overexpressed. This suggests that E2F1, E2F3, and E2F5 are involved in transcriptional induction of MMP9 and MMP14 genes in NSCLC cells, but they may play a lesser role in regulating the endogenous MMP15 promoter. To determine whether MMP activity could be rescued by an alternate E2F family member when E2F1 is depleted, A549 cells were transiently transfected with siRNA targeting E2F1, then transfected with MMP-luc constructs, with or without E2F3 expression vector. E2F1 depletion lead to reduced MMP9 and MMP14-luc activity, and overexpression of E2F3 could rescue MMP9 and MMP14-luciferase activity. Similar results were obtained when E2F3 was depleted, followed by overexpression of E2F1. MMP15-luciferase activity was not affected by E2F1 or E2F3 depletion (Fig 3D).

Figure 3.

(A-B) Transiently transfecting 100pmol of E2F1, E2F3, and E2F5 siRNA reduced the expression of MMP9 and MMP14 mRNA in A549 and H1650, and there was no significant difference with E2F2 or E2F4 siRNA. DHFR mRNA levels were significantly reduced by E2F1 or E2F3 siRNA (P<0.05). MMP15 mRNA levels were not affected significantly. (C) CCL-210 lung fibroblast cells depleted of E2F1-5 by siRNA show less collagen degradation compared to control siRNA. The results of three independent experiments are shown. The total cell number on top of collagen was trypsinized, and counted at the termination of the experiment. Depletion of E2F1-5 had no effect on growth of CCL-210 cells when plated as a confluent monolayer. Images are one representative field of CCL-210 cells atop collagen, taken at 100x total magnification with phase contrast microscopy. (D) MMP9 and MMP14 luciferase activity is reduced by transiently transfecting E2F1 or E2F3 siRNA, followed by transfection of MMP-luc reporters. Co-transfection with the alternate family member (E2F1 overexpression in E2F3si cells, E2F3 overexpression in E2F1si cells) could rescue MMP-luciferase activity. Basal levels of MMP15 were not affected, though both E2F1 and E2F3 overexpression could induce MMP15-luciferase activity (*P<0.05).

Recent studies suggest that in certain tumor milieus, the ability of fibroblasts to actively degrade extracellular collagen is a climacteric step that allows cancer cells to escape the primary tumor site (25). Since we found that NSCLC cells depleted of E2F1, E2F3, or E2F5 had less MMP9 and MMP14, we next examined whether CCL-210 lung fibroblasts had an impaired ability to degrade type IV collagen, when depleted of E2F1-5. To this end, CCL-210 cells were transfected with siRNA to E2F1-5 or a non-targeting control RNA and plated on type IV collagen. After four days, CCL-210 cells with depleted E2F1-5 had less collagen degradation as indicated by Coomassie staining of the residual collagen, though the depletion of E2F1 or E2F3 had the most pronounced effect (Fig. 3C). To determine whether siRNA had any effect of proliferation of CCL-210 cells, cells were counted after being trypsinized off the collagen. There was no significant difference in cell number with any siRNA suggesting that CCL-210 cells grown to confluency are not dependent of proliferation for collagen degradation. Taken together, these results suggest that depletion of E2Fs in lung cells significantly diminishes MMP gene transcription, and hinders resultant biological processes such as collagen degradation.

Depletion of E2F1 or E2F3 reduces cell migration and invasion

There is evidence that cell migration is accomplished in part through cleavage of adherens junctions by MMPs (26). To determine whether E2F-mediated modulation of MMP genes might affect migration of A549 cells, wound healing assays were conducted in vitro. To ensure that changes in migration were independent of cell proliferation, asynchronous cells were pretreated with 10 μg/mL Mitomycin C, which arrests the cell cycle (Figure 4A-B). There was a comparable amount of migration in Mitomycin C treated and untreated cells, indicating that the observed migration was a direct result of motility into the empty space and independent of proliferation (Figure 4A). Next, to determine whether E2F depletion affected migration, cells were transfected with siRNA to E2F1, E2F3, a combination thereof, or a non-targeting control siRNA. Serum induced migration of cells transfected with the control, non-targeting siRNA into the wound; but migration was significantly reduced in cells transfected with E2F1 and E2F3 siRNA (Figure 4B). This suggests that E2F1 and E2F3 contribute to the migration of cells. This agrees with studies showing that E2F1(-/-) mice have abnormal epidermal repair upon injury, and impaired cutaneous wound healing (27).

Figure 4.

(A) A549 cells that have been treated with 10 μg/mL Mitomycin C have significantly reduced BrdU incorporation compared to untreated A549 cells (*P<0.05). (B) Serum stimulated A549 cells treated with 10 μg/mL Mitomycin C have similar migratory capacity as untreated A549 cells (**P<0.005)100x total magnification. (C) A549 cells depleted of E2F1, E2F3, or both had significantly reduced invasive properties, as seen in a Boyden Chamber assay (***P<0.001). (D) Depletion of E2F1, E2F3, or E2F1 and E2F3 combined significantly hinders the ability to A549 cells to migrate in response to serum (**P<0.005; ***P<0.001).

We next examined whether invasion was affected by depletion of E2F1 or E2F3 using a Boyden Chamber assay. A549 cells were transfected with siRNA targeting E2F1, E2F3 or a combination thereof. As shown in Figure 4C-D, cells which were depleted of E2F1 or E2F3 had completely lost the ability to invade through collagen and matrigel coated transwell filters, while cells transfected with a non-targeting control siRNA showed 1.8 ± 0.4 fold invasion in serum stimulated cells. This suggests that E2F1 or E2F3 are required for degradation of the ECM components, through the modulation of genes involved in their degradation.

Rb-Raf-1 disruptor, RRD-251, represses MMP transcription and inhibits invasion and migration in vitro

Previous work in our lab has demonstrated that the Raf-1 kinase interacts with Rb, and phosphorylates Rb early in the cell cycle (10). A small molecule Rb-Raf-1 disruptor, RRD-251, inhibited Rb phosphorylation, thereby keeping Rb associated with E2F1, preventing cell proliferation and tumor growth (12, 14). We hypothesized that RRD-251 would likely inhibit the migration of cancer cells as well, given that depletion of E2Fs inhibited migration. Wound healing assays conducted on MDA-MB-231, MDA-MB-435, A549, and H1650 cells showed that treatment with RRD-251 significantly reduced the migration of cells (Figure 5A). Similarly, collagen degradation was also inhibited after CCL-210 cells were treated with RRD-251 (Figure 5B).

Figure 5.

(A) MDA-MB-231, MDA-MB-435, A549, and H1650 cells treated with RRD-251 show reduced migration compared to serum. (B) CCL-210 cells treated with RRD-251 show reduced degradation of type IV collagen. Images show three independent experiments. (C) A549 cells treated with RRD-251 have significantly reduced MMP9, MMP14, and MMP15 mRNA (*P<0.05). (D-E) A549-luc-C8 and MDA-MB-231 cells have significantly more invasion when stimulated with serum (**P<0.005). This effect is abrogated when treated with RRD-251 (A549-luc, ***P<0.001; MDA-MB-231, **P<0.005).

To determine if RRD-251 could prevent E2F mediated transcription of MMP genes, quiescent A549 cells were serum stimulated in the presence or absence of 20 μM RRD-251. MMP mRNA levels decreased after treatment with RRD-251 (Figure 5C), comparable to the reduction in expression when E2Fs were depleted. The ability of RRD-251 to inhibit invasion of cancer cells was next examined. A549-luc-C8 cells were rendered quiescent by serum starvation for 24 hours, and then stimulated with either serum alone, or serum and RRD-251 for 18 hours and invasion was measured by a Boyden Chamber assay. It was found that RRD-251 could significantly abrogate the invasive capacity of A549-luc and MDA-MB-231 cells (Fig. 5D and 5E). Collectively, these results suggest that RRD-251 inhibits the invasion and migration of cells, and this correlates with the repression of MMP genes.

RRD-251 inhibits metastatic lung colonization in vivo

Given that RRD-251 could modulate MMP levels and inhibit invasion and migration in vitro, we next investigated if RRD-251 could inhibit metastasis in vivo. We injected A549-luc-C8 cells (5×10⁶) into the lateral tail vein of 5-week-old female, SCID-beige mice. Mice were then randomized into either the DMSO vehicle group, or the RRD-251 group, which received i.p. Injection of 50mg/kg everyday for four weeks. Colonization of lungs was monitored using the Caliper-IVIS 200 system after administration of luciferin. Mice treated with RRD-251 had significantly less metastasis to the lung and surrounding tissues (Figure 6A-B). Photon Flux in vehicle treated mice was 3.9 ± 0.6 fold higher than mice treated with RRD-251. To confirm these observations seen in vivo, lung bioluminescence was examined ex vivo. Mice treated with RRD-251 had 80% less lung bioluminescence (Fig 6C-D), indicating less metastasis. H&E staining indicates that few tumors were able to seed in the lungs of mice treated with RRD-251 (Fig. 6E-F). These results suggest that RRD-251 can be used to inhibit metastatic growth of tumor cells in vivo.

Figure 6.

(A) A549-luc-C8 cells were injected into the lateral tail vein of SCID-beige mice, and animals were imaged weekly for 5 weeks. Daily administration of 50mg/kg RRD-251significantly reduced lung colonization (*P<0.05). Representative images are shown in (B). (C and D) At the completion of the experiment, lungs were analyzed ex vivo and extent of colonization quantified. Mice treated with RRD-251 had significantly less tumor burden (*P=0.015). (E and F). H & E staining confirms the presence of numerous distinct metastatic foci in the lungs of vehicle treated mice, shown by arrows, but few in mice treated with RRD-251.

Discussion

It is well established that cyclin dependent kinases phosphorylate and inactivate Rb in the G1 phase of the cell cycle, releasing E2F transcription factors from Rb (8, 28). It was initially believed that the predominant function of E2Fs was to activate genes required for the progression of the cell cycle through S-phase (3, 29, 30). Later studies showed that E2F transcription factors could regulate a diverse number of biological processes including cell differentiation, development, apoptosis, DNA damage repair, and more recently, angiogenesis (15, 31-34). As mentioned earlier, our lab had shown that the signaling kinase Raf-1 directly interacts with Rb early in the cell cycle; further, Raf-1 could phosphorylate Rb (10). This phosphorylation of Rb by Raf-1 was necessary for the subsequent complete inactivation of Rb by cyclin dependent kinases. Disrupting the Rb-Raf-1 interaction using RRD-251 could inhibit cell proliferation, adherence independent growth, angiogenesis and prevent the growth of lung cancer and melanomas in xenograft models (13). Studies presented here show that disrupting the Rb-Raf-1 interaction could be a fruitful way of combating metastatic colonization of tissues by cancer cells.

A considerable amount of research has been dedicated to identifying novel E2F regulated genes by gene profiling arrays and chromatin immunoprecipitation arrays (ChIP on chip) (35-39). In these arrays, various proteins and enzymes involved in the metastatic spread of tumor cells were initial hits, though validation studies have been lacking. Arguably the most crucial process for cancer cell invasion is the physical degradation of the ECM (26) but a role for E2F transcription factors in this process had not been identified. It is intriguing that, at least in lung cancer cell lines, E2Fs function as transcriptional activators of MMP9, MMP14, and MMP15. Many MMP gene promoters have multiple GC boxes which can bind to Sp1 and Sp3, including MMP9, MMP14, MMP15, and others (40). It is well established that Sp1 proteins can work coordinately with E2F transcription factors to regulate gene expression (41).

Cells utilize enzymes including serine-, thiol-, proteinases, heparanases, and metalloproteinases to free them from the primary tumor locale (16). Though the activity of matrix metalloproteinases and other metzincin family proteins are important for metastasis, the most prognostically valuable are the Matrix Metalloproteinase family (MMPs) (42). A relevant MMP signature is MMP2, MMP9, and MMP14, which have been shown to correlate with advanced stage breast cancer morbidity and late relapse in breast cancer patients (43, 44). MMP14 and MMP2 have also been detected at high levels in NSCLC patient samples, whereas MMP14 and MMP15 RNA levels have been shown to correlate with human glioma grade (45, 46). Therefore, it is a possibility that E2Fs might indirectly regulate tumor metastasis as a consequence of transcriptionally activating MMPs.

In addition to the crucial role for MMPs in degrading the extracellular matrix during invasion, they also play a role in angiogenesis. Pro-angiogenic factors like VEGF and bFGF are normally localized to the matrix, and cannot engage receptors until freed through MMP9 cleavage (47-49). Since we have previously shown that VEGF receptors, FLT-1 and KDR, are also E2F regulated genes, it is likely that the role E2F has in angiogenesis is multi-faceted.

These observations raise the possibility that mutations that initiate the oncogenic process by activating the E2F transcriptional regulatory pathway might also contribute to subsequent steps of tumor progression and metastasis. There is evidence that the Rb-E2F pathway might affect EMT as well, and this requires additional investigation (50). Taken together, these studies link the Rb-E2F cell cycle regulatory pathway to advanced stages of cancer development and metastasis. The finding that disrupting the Rb-Raf-1 interaction could prevent cell proliferation, angiogenesis, tumor growth and now metastatic colonization of organs suggest that targeting the Rb-R2F pathway might be a fruitful avenue to combat metastatic disease.