Small, cell-permeable and target-specific chemical ligands are particularly useful in systematic genomic approaches to study biological questions. They can rapidly penetrate into cells, bind to their target proteins and create a loss-of-function (or gain-of-function) phenotypes.1 Also, chemical genomics is an emerging field where several chemistry and genomics tools and technologies are combined to enhance the drug discovery process.2
Here, we used farnesyl transferase inhibitors to understand the H-Ras function. Farnesylation of Ras is the mandatory step for the activation of Ras. On the basis of the theoretical assumption that preventing Ras farnesylation could result in the inhibition of Ras function, a range of farnesyl transferase inhibitors have been developed. Ras proteins undergo a series of posttranslational modifications. The posttranslational modification of Ras by farnesyl is essential for its membrane association and transforming activity. The first and obligatory step, the addition of a farnesyl group to its carboxy-terminal “CAAX” motif, is catalyzed by FPT.3, 4 Inhibition of Ras farnesylation is a promising approach for developing a new generation of mechanism-based anticancer drugs.4, 5, 6
H-Ras has essential roles for tumor maintenance and angiogenesis. The expression of activated H-Ras is required for the maintenance of experimental melanoma tumor growth.7 Addition of activated H-Ras results in cells capable of forming rapidly proliferating tumors in vivo. Introduction of activated H-Ras into immortalized endothelial cells makes it capable for activating the angiogenic switch. Ras converted cells to an angiogenic phenotype and formed angiosarcomas in mice.8 Activation of H-Ras not only appears to be the critical event in tumor initiation but also a main player in the mouse tumor angiogenic response9 through up-regulation of VEGF and COX-2 expression, which is regulated by activation of both the Ras/Raf/MEK/ERK pathway as well as the Ras/PI3K/PDK/PKB pathway.10 Recent reports showed that H-RasV12 is more potent in transformation than K-RasV12 or N-RasV12 through phosphatidylinositol 3-Kinase (PI3K) activation in fibroblast and epithelial cells.11
Angiogenesis, the formation of new vasculature from preexisting blood vessels, involves tissue remodeling.12 This process requires proteolytic enzymes to degrade extracellular matrix. The most dominant proteins in the vascular tissue milieu are the interstitial collagens. Proteolytic remodeling of the collagen structure in vascular tissue can be carried out by several soluble matrix metalloproteinases (MMPs), including MMP-1 (collagenase-1), MMP-8 (collagenase-2) and MMP-13 (collagenase-3). In addition, MMP-9 has been reported to have a major role during the progression of angiogenesis from mouse model assay.13, 14
Although the suggestion has recently been made that inhibition of Ras might be a target for effective angiogenic therapy,15, 16 a direct role for Ras-mediated signal transduction and intracellular mechanism of farnesyl transferase inhibitors (FTIs) to inhibit tumor angiogenesis is unclear. In this study, we tried to determine the target molecules regulated by H-Ras and related to the antitumor effects of FTIs using the chemical genomic approach. Being used, the inhibitors having quite different structures are optimized through the chemical modification. Arteminolide-Rd (AP-Rd) is prepared by reduction of the α-methylene-γ-lactone group of arteminolide E (APE), which was isolated from the aerial parts of Artemisia sylvatica.17 APE is a sesquiterepne lactone. LB42908 is a drug-like and nonpeptidic compound with a pyrrole and imidazole ring.18 SCH66336 is prepared through structure–activity relationship (SAR) study based on the SCH44342 compound, which is the benzocycloheptapyridine tricyclic compound.19 We profiled H-Ras and the inhibitors that are regulating the genes by using cDNA microarray. From this, we identified MMP-13 as another H-Ras effector molecule for angiogenesis.
Material and methods
Arteminolide-Rd (AP-Rd) was prepared by reduction of the α-methylene-γ-lactone group of APE.17 SCH66336 and LB42908 were synthesized by reported methods.19 We purchased the MAK inhibitor PD98059, NF-kB inhibitors kamebakaurin and MMP-9/13 inhibitor (Calbiochem), and PI3K inhibitor Wortmannin was purchased from Sigma. The H-Ras, phospho-ERK and ERK antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The phospho-AKT, AKT and PARP antibodies were purchased from Cell Signaling (Bevery, MA). The MMP-13 antibody was obtained from Oncogene research Products (San Diego, CA). HRP-linked antirabbit or antimouse IgG antibodies were obtained from Cell Signaling (Bevery, MA).
Rat2 fibroblast cells and stable H-Ras-transformed Rat2 cells were grown in DMEM (Invitrogen) medium supplemented with 10% fetal bovine serum (FBS, Invitrogen) at 37°C in a humidified incubator with 5% CO2. Geneticin (Invitrogen) was used for selection and maintenance of stable transformants. SVR cells (over-expressed H-Ras cells) were derived from primary murine endothelial cells and were maintained with 5% FBS.20 In experiments, cells were seeded at a density of 3 × 105 per 10-cm dish. After incubation for 24 hr, cells were treated with each inhibitor.
cDNA microarray hybridization and data analysis
Total RNA was extracted using the TRIzol reagent (Invitrogen). One hundred micrograms of total RNA were used to prepare direct Cy3- and Cy5-dUTP (fluorescent dye)-labeled first-strand cDNA probes using a single-base anchored oligo-dT17 primer (Amersham Pharmacia Biotech) and SuperScript reverse transcriptase (Invitrogen). mRNA from reference cells (Rat2 or H-Ras cells) was used to prepare cDNA labeled with Cy3-dUTP, and mRNA harvested from H-Ras cell lines or FTIs treated H-Ras cells was used to prepare cDNA labeled with Cy5-dUTP. The probes were mixed and hybridized to a human 10K cDNA chip (from KRIBB, Taejon, Korea) at 42°C overnight. The information and list of the human 10K chip can be found at http://kugi.kribb.re.kr. Fluorescent images of hybridized microarrays were captured using the GenePix 4000 (Axon) and analyzed using ImaGene 5.5. The background substraction, data normalization, cluster analysis and expression value (signal) were calculated using GeneSight 4.0 software (BioDiscovery). Optimal normalized data should be horizontal and centered at zero. This data were used to generate fold change. Housekeeping controls, β-actin and GAPDH genes, served as endogenous controls and were useful for monitoring the quality of the target.
The cDNA was synthesized from a total of 5 μg of RNA using Superscript II reverse transcriptase (Invitrogen). Two microliters were used in a total of 20 μl reaction volume as a template for PCR amplification. PCR was performed under standard conditions in 20 μl; 10 mM Tris, pH 8.3, 40 mM KCl, 1.5 mM MgCl2, 250 μM dNTP, 10 pM each primer (sense and antisense) and 1 U Taq DNA polymerase (Bioneer, Taejon, Korea). Primer sequences were designed using the primer software at the web site http://frodo.wi. mit.edu/primer3/primer3_code.html. Quantity analysis is based on the intensity of the RCR product using the Quantity ONE software (Bio-Rad).
Northern blot analysis
RNA samples (10 μg) were fractionated on a 1% agarose–formaldehyde gel and then transferred onto a positively charged nylon membrane (Roche). A DNA template was prepared from total RNA using RT-PCR and PCR. PCR was performed with specially designed primers, including the sequence of the appropriate T7 RNA polymerase using the Expand High Fidelity PCR System (Roche). RNA was labeled with digoxigenin-11-UTP using the DIG Northern Starter Kit (Roche). Hybridization was performed in DIG Easy Hyb solution at 68°C overnight. Membranes were washed twice in 2× SSC, 0.1% SDS at 15–20°C for 10 min, twice in 0.1× SSC, 0.1% SDS at 68°C for 30 min, and performed the immunological detection with CDP-Star (Roche) before exposure to X-ray film for 5–25 min at 15–25°C.
Gelatin zymography for MMPs activity assay was performed as previously described.21 Proteins in the conditioned media were separated without prior boiling by electrophoresis through SDS/polyacrylamide (10%) gels containing 1 mg/ml gelatin (Sigma) under nonreducing conditions. After electrophoresis, the gel was rinsed with renaturing buffer (2.5% Triton X-100) for 30 min and then incubated overnight at 37°C in developing buffer (50 mM Tris, pH 7.5, 5 mM CaCl2, 1 μM ZnCl2). Coomassie blue R250-stained zymograms were scanned with a personal densitometer.
Western blot analysis
A 20 μg protein was resolved by 7.5 or 12% SDS-PAGE and transferred to the PVDF membrane (Roche, Germany). The membrane was blocked with 5% nonfat dry milk in TBS-T (50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 0.1% Tween 20). The antibodies were used at dilutions recommended by the manufacturers. The membrane was incubated with each antibody for 1–2 hr at room temperature, washed 5 times with TBS-T and visualized with Chemiluminence POD reagents (Roche, Germany).
Colony forming assays were carried out in 6-well plates previously lined with 0.6% agar in DMEM (10% FBS) containing 0.1% DMSO. Rat2 and H-Ras-transformed Rat2 cells were seeded at a density of 1 × 105 cells/well in DMEM (2.5% FBS) containing 0.3% agar. Photomicrographs were taken 3 weeks after plating.
Capillary formation in 3D collagen
3D collagen gels were made from collagen solution according to the manufacturer's recommendation (Chemicon). The collagen solution was prepared on ice by mixing together the following stock solutions: collagen solution, 5× DMEM medium and neutralization solution in a ratio of 8:2:0.25 by volume. Suspensions of H-Ras-transformed endothelial cells (SVR cells) containing FTIs or the MMP-inhibitor were mixed with the chilled collagen solution in a ratio of 1:9 (by volume). One hundred microliters of the mixture were transferred to a 96-well tissue culture plate, and the resulting cell suspension contained 6,000 cells per well, and then immediately transferred to a 37°C incubator for 60 min to initiate polymerization of the collagen. After formation of the collagen gel, the collagen gel was covered with culture media and cultured for 2 weeks.
Zebrafish angiogenesis assay
The effect of FTIs or the MMP9/13 inhibitor was assayed in vivo using the zebrafish angiogenesis model.22 Embryos were generated by natural pair-wise mating, as described by Serbedzija et al.22 Through 20 hr of development, the fertilized embryos were plated into individual wells of 96 well plates at 28°C. After 24 hr, the FTIs or the MMP9/13 inhibitor were added directly to the well. After 48 hr, embryos were fixed in 4% paraformaldehyde for 2 hr at room temperature. For staining, embryos were equilibrated in NTMT buffer (0.1 M Tris-HCl pH 9.5; 50 mM MgCl; 0.1 M NaCl; 0.1% Tween 20) at room temperature. Once the embryos equilibrated in NTMT, 4.5 μl of 75 mg/ml NBT and 3.5 μl of 50 mg/ml X-phosphate were added. After staining for 10 min, all the blood vessels in the fish embryo were labeled and then photographed with a Nikon microscope and CCD camera.
Farnesyl transferase inhibitors suppress phenotypic transformation induced by H-Ras
Oncogenic H-Ras has been known to stimulate cell growth and change cell morphology into spindle-like foci. Activation of ERK is one of the well-known signaling events lead by H-Ras. As shown in Figure 1, expression of oncogenic H-Ras activates ERK1/2 kinase and makes cell morphology into needle-like shapes. In addition, H-Ras enhanced colony formation in the soft agar assay.
We used the chemical genomic approach to understand H-Ras function in cancer cells. First, we examined whether FTIs with each different structures, AP-Rd, SCH66336 and LB42908 (Fig. 2), can inhibit the Ras signaling pathway in H-RasV12-transformed Rat2 cells. APE was isolated from the Artemisia sylvatica,17 and AP-Rd was prepared by reduction of the α-methylene-γ-lactone group of APE. Both APE and AP-Rd blocked H-Ras farnesylation, but APE has a sequiterpene lactone with the α-methylene-γ-lactone, which exhibits a wide range of biological activities23 and has cytotoxicity in normal cells. Therefore, we decided to use AP-Rd as a FTI with SCH66336 and LB42908. As shown in Figure 3, 3 FTIs completely blocked Ras farnesylation and phosphorylation of ERK but did not inhibit phosphorylation of Akt. ERK and Akt activation was assayed with the phosphorylated form specific to the Erk and Akt antibody under our conditions. In addition, by treatment with FTIs, spindle-like, the foci-forming morphology of H-Ras Rat2 cells were changed to a flattened morphology, which is similar to the Rat2 cells. In the presence of 10 μM U0126 (MAK1/2 inhibitor), spindle-like morphology of H-Ras Rat2 cells were again changed to a flattened morphology. However, cells treated with the PI3-K inhibitor (10 μM, Wortmannin) did not show morphological changes (data not shown). These results suggested that the inhibitors mainly blocked the ERK pathway, and H-Ras-mediated morphological change may be mediated through the ERK signaling pathway but not through the PI3K pathway.
Profile of H-Ras target genes sensitive to FTIs
To identify H-Ras target genes, expression profiles of 10,000 genes in Rat2 cells vs H-Ras-transformed Rat2 cells were compared using microarray analysis. Among the 10,000 genes, expression of 80 genes was shown to be altered by H-Ras transformation. We presented data that were up- or down-regulated at least 1.7-fold in the H-Ras-transformed Rat2 fibroblast cells (Supplementary data). Among the 80 genes, 32 were up-regulated, and 48 were down-regulated. Interestingly, many of the genes encoding extracellular protein and cytoskeletal components proteins were down-regulated, whereas genes encoding metabolic enzyme proteins and transcription factors were up-regulated.
As shown in Figure 3, FTIs are potent chemicals to block functions of H-RasV12. We searched H-Ras target genes that are sensitive to FTIs by using human cDNA microarray. Among the 80 genes that were regulated by H-Ras, only 14 genes were sensitive to the inhibitors. The identified genes were similarly regulated by 3 different FTIs, AP-Rd, SCH6636, LB42907, supporting specific FTI targets except the gene regulated by structural differences (Fig. 4 and supplementary data). MMP 10, 13 and stress-induced-phosphoprotein 1 were down-regulated by FTIs, whereas collagen type1 α1, protein tyrosine phosphatase receptor type A and vimentin were up-regulated. Because MMPs are reported to play an important role in matrix degradation during tumor growth, tumor invasion and tumor-induced angiogenesis, we focused our efforts to understand the H-Ras dependent MMPs expression and activation. RT-PCR analysis of the MMPs expression in Rat2 cells or inhibitor-treated H-Ras-transformed cells demonstrated that MMP-3, MMP-9, MMP-10, MMP-12 and MMP-13 were stimulated by H-Ras. Among them, MMP-9 and MMP-13 were much more sensitive to FTIs. However, expression of MMP-2 was decreased by H-Ras (Fig. 5).
MMP-13 is regulated by H-Ras and ERK signaling pathway
Recently, MMP-13 has been reported that its expression is correlated with the invasion capacity of the tumors.24, 25, 26, 27, 28 However, the regulatory mechanisms responsible for the MMP-13 expression are poorly understood at present. As shown in Figure 6, the MMP13 mRNA and protein levels were increased in H-RasV12-transformed Rat2 cells compared with Rat2 cells. When H-RasV12-transformed Rat2 cells were treated with each FTI or ERK inhibitor (10 μM, U0126), the MMP13 mRNA and protein levels decreased. However, the MMP-13 transcription level slightly increased by the PI3K inhibitor (10 μM, Wortmannin) (Fig. 6a). In the gelatin zymograph assay, the FTIs and ERK inhibitor (10 μM, U0126) inhibited the activity of MMP-9 and MMP-13. However, the PI3K inhibitor (10 μM, Wortmannin) did not inhibit the activity of them (Fig. 6b). When the NF-kB inhibitor (0.5 μg/ml, kamebakaurin), p38 inhibitor (SB203580) and JNK inhibitor (SP600125) were treated in cells, the expression of MMP-13 did not change (Fig. 6c and data not shown). This result demonstrated that the induction of the MMP-13 gene by H-RasV12 is mediated by the ERK signaling pathway.
Effects of H-Ras and MMP-13 on angiogenesis
Although the FTIs completely blocked H-Ras farnesylation and MMP-13 expression, interestingly, they could neither inhibit cell proliferation nor induce of apoptosis on the 2D cell culture dish (data not shown). Therefore, we assessed the effects of inhibition of H-Ras and MMP-13 activities on angiogenesis by using an in vitro and in vivo model. To examine the effects of the inhibitor H-RasV12 or MMP-9/13 on angiogenesis in vitro, we employed a tube formation assay with H-Ras-transformed endothelial cells. When cultured in a 3D collagen gel matrix, these cells displayed marked morphological changes (Fig. 7a). As shown in Figure 7a, H-Ras-transformed cells strongly stimulated cell elongation and formed a capillary structure, while H-Ras cells treated with FTIs or the MMP-13 inhibitor inhibited the migration. To investigate whether FTIs and the MMP-inhibitor could inhibit angiogenesis in vivo, we used the zebrafish as a model animal. Blood vessel patterning is highly characteristic in the developing zebrafish embryo, and the subintestinal vessels (SIVs) can be stained and visualized microscopically.22 When we treated the embryos with FTI or the MMP-9/13 inhibitor, they blocked angiogenesis in vivo. Treatment with 20 μM of SCH 66336 completely inhibited SIV formation (Fig. 7b). LB 42908 and AP-Rd also inhibited SIV formation at 20 and 30 μM, respectively (data not shown). This finding supported that activation of H-Ras is important to induce a transformed phenotype and angiogenesis. Similarly, when the MMP-9/13 inhibitor was used, SIV formation in the zebrafish was inhibited (Fig. 7b) too. These results supported that H-Ras-mediated MMP-13 expression and activation is involved in angiogenesis and down-regulation of MMP-13 by FTIs, which may be one of anticancer mechanism of FTI.
FTIs were originally developed on the premise that farnesyl transferase inhibition would prevent Ras farnesylation and, therefore, block transduction of the proliferative signal.3, 4, 5 Several reports showed that FTIs induce apoptosis through the release of cytochrome c from the mitochondria resulting in caspase-3 activation.29, 30 However, we could not detect any apoptosis under conditions that the farnesylation of H-Ras was completely inhibited by AP-Rd, SCH66336 or LB42908. Only APE having α-methylene-γ-lactone group induced PARP degradation. Our results suggest that apoptosis induced by APE may be caused by nonspecific inhibition of unidentified target proteins, and blocking of the H-Ras farnesylation is not sufficient to induce apoptosis.
Although FTI was originally conceived to target mutant or aberrant Ras functions in cancer, recent studies suggest that the antitumor activity of FTIs are not due to the inhibition of Ras proteins exclusively, but may also involve inhibition of RhoB, a G-protein that regulates receptor trafficking,31 the centromere-binding proteins CENP-E and CENP-F32 and other proteins, which are yet to be identified.33 So, further studies are needed to identify the crucial target proteins of FTI except H-Ras and to understand the antitumor activity and antiangiogenesis activity of FTI.
In this report, we found that the H-Ras target genes were sensitive to FTIs by the chemical genomic approach using cDNA microarray analysis. First of all, we identified 80 genes being induced or repressed by H-RasV12. Many of the genes encoding extracellular protein and cytoskeletal component proteins were down-regulated by H-RasV12, whereas genes encoding the growth of cells were up-regulated. These expression profiles are similar to the profiling for H-Ras (G12V)-transformed rat embryonic fibroblast cells by using subtractive suppression hybridization (SSH), a PCR-based cDNA subtraction technique,34 and mouse embryonic fibroblasts transformed by H-RasV12-mutated protein and the EIA oncogene by using DNA microarry analysis.35 The phenotypic changes by H-Ras transformation are results from alterations in the organization of the actin cytoskeleton and adhesive interactions. Epithelial cells transformed by oncogenic Ras acquire a mesenchymal phenotype, which is associated with a decrease in cell–cell adhesion and an increase in focal adhesions and stress fibers.36, 37 Our results may explain, at least in part, the H-Ras-mediated morphological and growth transformation of fibroblast too.
Next, we determined target genes to FTIs in H-Ras-transformed cells. From this analysis, we found that cytoskeletal components, adhesion molecules and extracellular proteins, such as MMP-10, MMP-13 and collagen type1 α were distinct H-Ras targets and were sensitive to FTIs. Several studies suggested that H-Ras has an essential role for tumor maintenance angiogenesis initiation and progression of several types of tumors.7, 8, 9 MMPs are thought to play an important role in matrix degradation during the tumor growth, tumor invasion and tumor-induced angiogenesis.38 In this report, we identified that MMP-9 and MMP-13 are more sensitive to FTIs than other H-Ras-dependent MMPs such as MMP-3, 10 and 12. It was reported that the cell lines immortalized following Myc expression were found to up-regulate MMP-7, 11 and 13, but the expression of MMP-9 was up-regulated by H-Ras.39 Several studies also reported that MMP-2 and/or MMP-9 is the essential invasive phenotype of cancer cells. In addition, MMP-9 up-regulation has been most commonly associated with H-Ras transformation,39, 40, 41, 42 and SCH66336 down-regulates secretion of matrix proteinases 9 and inhibits carcinoma cell migration.43 In the present study, we found for the first time that MMP-13, as well as MMP-9, is intimately involved in H-Ras using the chemical genomic approach.
H-Ras have been reported to activate MEK-ERK and PI3K pathways. However, FTI inhibited phosphorylation of ERK, but not AKT, under our conditions. In addition, FTIs and U0126 (an inhibitor of MEK1/2 activation) inhibited MMP-13, but Wortmannin (PI3K inhibitor) did not. The induction of MMP-13 has been demonstrated to be mediated through mitogen-activated protein kinases (MAPKs), which regulate cell growth, differentiation, survival and death.44, 45, 46 It has been reported that serotonin-induced MMP-13 production is mediated by phospholipase C, protein kinase C and ERK1/2 in rat uterine smooth muscle cells,46 and the macrophage migration inhibitory factor (MIF) also up-regulated MMP-13 through the ERK1/2 and AP-1-dependent pathway.45 However, in cases of transformed human epidermal keratinocytes, MMP-13 was induced by TNF-α and TGF-β is dependent on the activity of the p38 pathway.44 The induction of the MMP-13 gene in chondrocytes by IL-1 required the transcription factor NF-kB, as well as p38 MAPK and JNK, but not ERK MAPK.47 JNK was recently shown to be required for IL-1 induction of MMP-13 expression in mouse fibroblasts and a rheumatoid arthritis model.48 Additionally, Lechuga reported TGF-β1 modulates MMP-13 expression in hepatic stellate cells by complex mechanisms involving p38MAPK, PI3-Kinase, AKT and p70S6k.49 In this study, we suggested that the induction of MMP-13 by H-Ras was critically regulated through ERK1/2 but not PI3-kinase as well as NF-kB, p38 MAPK and JNK. However, when both p38 and ERK pathways were blocked by their inhibitors, MMP-13 expression was strongly decreased.
The MMP-13 plays important role in bone remodeling during growth, development and wound healing. MMP-13 is especially known to be expressed primarily by cancer cells at the invading edge of the tumor, and its expression is correlated with the invasion and metastasis capacity of the tumors.24, 25, 26, 27, 28 Therefore, we tested the effect of FTI and the inhibitor of MMP9/13 on angiogenesis. As expected, both FTI and the inhibitor of MMP9/13 showed significant antiangiogenesis activity. Several studies reported that MMPs are associated with invasion phenotype in tumor cells or Ras-transformed cells.42, 50 We demonstrated that inhibition of MMPs directly block angiogenesis. Inhibition of MMP-9 reduced angiogenesis in human microvascular endothelial cells.51 Chicken MMP-13 (chMMP-13) was the only enzyme whose induction and expression coincided with the onset of angiogenesis and blood vessel formation in the chorioallantoic membrane (CAM) of the chick embryos.52 Our data suggested that MMP-9 and MMP-13 were directly involved in the H-Ras-induced angiogenesis phenotype. These results suggested that FTI regulated activity and expression of MMP9/13 via inhibition of the Raf/MEK/ERK pathway.