Blockade of the RhoA-JNK-c-Jun-MMP2 Cascade by Atorvastatin Reduces Osteosarcoma Cell Invasion*

Osteosarcoma is characterized by a high malignant and metastatic potential, which points to the need for new therapeutic strategies to prevent cell metastasis. In this study, we show that statin-induced HMG-CoA reductase inhibition reduces cell migration and invasion in human and murine osteosarcoma cells, independently of the genotype. The statin-induced reduction of cell migration and invasion was independent of induction of apoptosis and was geranylgeranylpyrophosphate-dependent. The statin reduced matrix metalloproteinase (MMP) 2, 9, and 14 and TIMP2 expression or activity in invading cells. Forced expression of MMP2 and MMP14 overcame the inhibitory effect of the statin on cell invasion, suggesting a role for these MMPs in invasive potential. We also investigated the mechanisms involved in the reduced MMP2 activity and cell invasion. Inhibition of JNK, but not ERK1/2 signaling, reduced MMP2 activity. Pharmacological or constitutive activation of JNK overcame the reduced MMP2 activity and cell invasion induced by the statin. The statin decreased JNK phosphorylation and c-Jun nuclear translocation, suggesting that HMG-CoA reductase inhibition targets the JNK-c-Jun signaling pathway. We showed that mevalonate or geranylgeranylpyrophosphate treatment prevented the statin-induced reduction in JNK phosphorylation, MMP2 activity, and cell invasion. Forced expression of a constitutively active form of RhoA increased JNK phosphorylation and overcame the inhibitory effect of atorvastatin on MMP2 activity and cell invasion. The data establish a link between RhoA, JNK, c-Jun, and MMP2 activity that is functionally involved in the reduction in osteosarcoma cell invasion by the statin. This suggests a novel strategy targeting RhoA-JNK-c-Jun signaling to reduce osteosarcoma cell tumorigenesis.

Osteosarcoma is characterized by a high malignant and metastatic potential, which points to the need for new therapeutic strategies to prevent cell metastasis. In this study, we show that statin-induced HMG-CoA reductase inhibition reduces cell migration and invasion in human and murine osteosarcoma cells, independently of the genotype. The statin-induced reduction of cell migration and invasion was independent of induction of apoptosis and was geranylgeranylpyrophosphatedependent. The statin reduced matrix metalloproteinase (MMP) 2, 9, and 14 and TIMP2 expression or activity in invading cells. Forced expression of MMP2 and MMP14 overcame the inhibitory effect of the statin on cell invasion, suggesting a role for these MMPs in invasive potential. We also investigated the mechanisms involved in the reduced MMP2 activity and cell invasion. Inhibition of JNK, but not ERK1/2 signaling, reduced MMP2 activity. Pharmacological or constitutive activation of JNK overcame the reduced MMP2 activity and cell invasion induced by the statin. The statin decreased JNK phosphorylation and c-Jun nuclear translocation, suggesting that HMG-CoA reductase inhibition targets the JNK-c-Jun signaling pathway. We showed that mevalonate or geranylgeranylpyrophosphate treatment prevented the statin-induced reduction in JNK phosphorylation, MMP2 activity, and cell invasion. Forced expression of a constitutively active form of RhoA increased JNK phosphorylation and overcame the inhibitory effect of atorvastatin on MMP2 activity and cell invasion. The data establish a link between RhoA, JNK, c-Jun, and MMP2 activity that is functionally involved in the reduction in osteosarcoma cell invasion by the statin. This suggests a novel strategy targeting RhoA-JNK-c-Jun signaling to reduce osteosarcoma cell tumorigenesis.
Osteosarcoma is the most common primary tumor of bone in children and young adults (1). These tumors are characterized by a highly malignant and metastatic potential (2). The metastasis process is a complex process that is dependent on cancer cell capacity to invade, migrate, and proliferate (3,4). Consistently, cell invasion and migration are highly related to the activity of matrix metalloproteinases (MMPs) 2 that regulate many processes involved in tumor evolution, such as cell growth, migration, and extracellular matrix degradation (5). Notably, MMP2 (gelatinase A) and MMP9 (gelatinase B) have been implicated in invasion and metastatic processes in several cancers (6,7).
The cholesterol-lowering agents statins act as inhibitors of 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase, which catalyzes the conversion of HMG-CoA into mevalonate. Mevalonate is converted into farnesylpyrophosphate (FPP) or geranylgeranylpyrophosphate (GGPP) that can be anchored onto intracellular proteins through prenylation, thereby ensuring relocalization of target proteins to cell membranes (8,9). Inhibition of HMG-CoA reductase results in alteration of prenylation of small G proteins such as RhoA (10), which regulates cell cycle and survival by modulating downstream signaling pathways (11,12). Accordingly, inhibition of HMG-CoA reductase by statins was found to trigger apoptosis in several cancer cells (10 -13). We recently showed that lipophilic statins dose-dependently induce apoptosis in human osteosarcoma cells through inhibition of RhoA prenylation and reduced ERK1/2-Bcl2 survival pathway (14). In addition to affect cell survival, inactivation of RhoA may also inhibit human cancer cell migration and invasion (15)(16)(17). One possible mechanism is that inactivation of RhoA may down-regulate the expression of MMP2 or MMP9 (18,19). However, the mechanisms by which RhoA inactivation may control MMPs and cell migration and invasion are not fully understood. In this study, we show that statin-induced inhibition of HMG-CoA reductase decreases cell migration and invasion in human and murine osteosarcoma cells through a mechanism involving inhibition of RhoA-JNK-c-Jun signaling and MMP2 activity. Our results thus implicate a role for JNK signaling in statin-induced inhibition of human cancer cell migration and invasion. aggressive CAL72 cells (21). Murine K7M2 (ATCC) derived from mouse osteosarcoma are highly tumorigenic in vivo (22). Normal fetal human osteoblastic cells (FHSO6) were derived from the bone marrow stroma (23). The cells were routinely cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% heat-inactivated fetal calf serum, 1% L-glutamine, and penicillin/streptomycin (10,000 units/ml and 10,000 g/ml, respectively) at 37°C in humidified atmosphere containing 5% CO 2 in air. The culture media were changed every 2 or 3 days. Human umbilical vein endothelial cells (HUVEC) (provided by Dr. Briquet-Laugier, Paris, France) were isolated by collagenase digestion. The umbilical cord was collected in sterile conditions, and the umbilical vein was washed with Ca 2ϩ -containing phosphate-buffered saline and incubated for 10 min with 1 mg/ml type I collagenase. Cells were collected by flushing the vein with Ca 2ϩ -containing phosphate-buffered saline, centrifuged, plated onto gelatin-coated dishes, and cultured in endothelial cell medium (Promocell, Heidelberg, Germany), supplemented with penicillin/streptomycin, at 37°C in a humidified atmosphere containing 5% CO 2 .
Cell Migration and Invasion Assays-Cell migration was determined in the modified Boyden chamber assay, using cell culture inserts (8-m pore; BD Biosciences, Le Pont de Claix, France) as previously described with minor modifications (24). Briefly, the cells (50,000 cells/insert) were preincubated 2 h with or without statin and/or z-VAD-fmk before seeding in inserts and incubated for further 22 h. The cells that did not migrate through the filter were removed from the upper surface of the membrane using cotton-tipped swabs. The cells migrated to the underside were fixed in 3.7% paraformaldehyde in phosphate-buffered saline at 4°C and stained with toluidine blue. The membranes were then cut from the insert and observed under microscope. Five fields were randomly selected and counted for each assay performed in duplicate.
For in vitro cell invasion assay, similar experiments were performed using inserts coated with a basement membrane, Matrigel (BD Biosciences). Trans-endothelial migration was meas-ured using inserts in which HUVEC were plated (150,000 cells/ insert) and cultured up to confluence. Five fields were randomly selected and counted for each assay performed in duplicate, taking in account the migration of HUVEC, which were counted separately from migrating osteosarcoma cells.
Cell Viability-Cell viability was determined as previously described (14). Briefly, cell viability was measured by the colorimetric microassay described by Mosmann (25) Tetrazolium salt (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) was added to the culture medium for the last hour of incubation. The culture media were then removed, and the cells were lysed in Me 2 SO. Microplates were read using a multiwell scanning spectrophotometer (Titertek Elisa reader) at 540 nm. We used six replicates for each condition, and the experiments were repeated at least three times.
Matrix Metalloprotease 14 (MMP14) Activity Assay-The cell lysates were prepared using the Biotrak assay kit and analyzed according to the manufacturer's recommendations (GE Healthcare).
Cell Transfection-Transfections were performed using Exgen500 (Euromedex, Mundolsheim, France) according to the manufacturer's recommendations. Briefly, the cells seeded at 25,000/cm 2 were cultured for 24 h in DMEM supplemented with 10% fetal calf serum and rinsed with serum-free DMEM and 1 ml of DMEM containing 7 equivalents of Exgen500, and 2 g of the indicated construct was added. The cells were incubated for 1 h at 37°C, and 1 ml of DMEM supplemented with 20% fetal calf serum was added to the medium. After 24 h, the transfection medium was replaced with fresh medium.
Western Blot Analysis-The cell lysates were prepared as previously described (14). Briefly, the proteins (30 g) were resolved on 12% SDS-PAGE and transferred onto polyvinylidene difluoride nitrocellulose membranes (Millipore). The filters were incubated at room temperature for 2 h in 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.1% (v/v) Tween 20, 0.5% (w/v) bovine serum albumin (TBST/BSA) and then overnight at 4°C on a shaker with specific primary antibodies (1/1000 in TBST/BSA). The membranes were washed twice with TBST and incubated for 2 h with appropriate horseradish peroxidaseconjugated secondary antibody (1/20,000 in TBST/BSA). After final washes, the signals were visualized with enhanced chemiluminescence Western blotting detection reagent (ECL; Amersham Biosciences) and autoradiographic film (X-Omat AR; Eastman Kodak Co.). Densitometric analysis using Image-Quant software was performed following digital scanning (Agfa).
Statistics-Differences between the mean values were analyzed using the statistical package super-analysis of variance (Macintosh, Abacus Concepts Inc., Berkeley, CA) with a minimal significance of p Ͻ 0.05.

RESULTS
Atorvastatin Reduces Osteosarcoma Cell Migration Independently of Its Pro-apoptotic Effect-We first compared the migratory potential of osteosarcoma cell lines with various genotypes to normal osteoblastic cells (FHSO6 cells) derived from fetal human bone marrow (23). In vitro cell migration was evaluated using modified Boyden's chambers. All of the cells were able to migrate although with different potential (Fig. 1A). SaOS2 cells showed the highest capacity to migrate, whereas MG63 cells showed the lowest migrating capacity. The normal cell line FHSO6 exhibited an intermediate potential.
On this basis, we tested the effects of atorvastatin on SaOS2 cell migra-tion potential. We found that atorvastatin reduced cell migration in a dose-dependent manner (-40% at 10 M) (Fig. 1B). We then tested the effect of atorvastatin at the dose of 10 M on cell migration. We found that the statin reduced cell migration in all human or murine osteosarcoma cell lines but did not modulate FHSO6 cell migration potential (Fig. 1C). Similar results were obtained with simvastatin (data not shown). We previously showed that atorvastatin reduces osteosarcoma cell viability by induction of caspases-dependent apoptosis (14). We thus tested whether the reduction in cell migration by the statin was related to apoptosis induction using co-treatment with a broad spectrum caspase inhibitor (z-VAD-fmk). The addition of 25 M z-VAD-fmk fully prevented the reduction in cell viability induced by the statin (Fig. 1D) but did not significantly prevent the statin-induced inhibition of cell migration (Fig. 1C). These results indicate that atorvastatin reduces the in vitro cell migration potential of human and murine osteosarcoma cells, independently of the genotype or of its pro-apoptotic effect.
Atorvastatin Reduces Osteosarcoma Cell Invasion Independently of Its Pro-apoptotic Effect-We next determined the effect of atorvastatin on in vitro cell invasion. To this goal, we used basement membrane Matrigel-coated transwells to mimic the extracellular matrix or transwells containing confluent HUVEC to mimic the endothelial barrier. As observed for in vitro cell migration, the osteosarcoma cell lines exhibited different invasion potential in basal condition. U2OS cells were the most invasive cells compared with other cells (Fig. 2A). In contrast, MG63 cells showed no invasive potential neither through Matrigel basement membrane nor through HUVEC barrier (Fig. 2, A-C). Treatment with 10 M atorvastatin reduced Matrigel cell invasion in all invading osteosarcoma cell lines (Ϫ55 to Ϫ85%, p Ͻ 0.05 versus untreated cells) but had no effect on basal invasive potential of normal human osteoblastic cells (Fig. 2B). The addition of 25 M z-VAD-fmk did not prevent the effect of atorvastatin on cell invasion (Fig. 2B), indicating that the inhibitory effect of atorvastatin was independent of apoptosis. Atorvastatin had no effect on HUVEC cell viability ( Fig. 2D) but reduced transendothelial cell invasion of all invading osteosarcoma cells (Fig.  2C). Again, as observed for migration and Matrigel invasion tests, no effect was observed in normal FHSO6 osteoblastic cells. These data show that atorvastatin reduces the number of osteosarcoma cell invading extracellular matrix (mimicked by Matrigel) or endothelial barrier (mimicked by HUVEC monolayer), whereas nontumoral osteoblastic cells were not sensitive to atorvastatin effects in terms of in vitro cell invasion potential.
Atorvastatin Reduces the Expression and Activity of Several MMPs Family Members-We investigated the role of MMP2 and MMP9, which are believed to play an important role in the invasive process (26). MMP2 is known to be activated in complexes including TIMP2 and MT1-MMP (MMP14) (27,28). Using quantitative reverse transcription-PCR analysis, we investigated the expression of these molecules in osteosarcoma cells under atorvastatin treatment. We found that atorvastatin reduced MMP2, MMP9, MMP14, and TIMP2 mRNA levels in U2OS cells (Fig.  3A). Similar results were found in other invading osteosarcoma cell lines (SaOS2, CAL72, and K7M2 cells; data not shown). Furthermore, atorvastatin reduced gela-   activity, evaluated by zymography (Fig. 3B), and reduced MMP14 activity, evaluated by Biotrak system (Fig. 3C). Atorvastatin also reduced MMP2, MMP9, MMP14, and TIMP2 protein levels in U2OS (Fig. 3D) and in other invading osteosarcoma cell lines (SaOS2, CAL72, and K7M2 cells; data not shown), as evaluated by Western blot analysis. The reduction in MMP2 activity by the statin was dose-dependent in U2OS cells, whereas MG63 cells, which are noninvading cells, and FHSO6 cells, which are normal osteoblastic cells and poorly invading cells, were not affected (Fig. 3E). To further demonstrate the role of MMP2 in cell invasion, U2OS cells were transfected with expression vectors encoding MMP2 or MMP14. Forced expression of MMP2 and/or MMP14 abolished the inhibition of cell invasion induced by atorvastatin in these cells (Fig. 3F). These data indicate that atorvastatin reduces gelatinases expression and activity, notably MMP2, resulting in reduction in the invasive capacity in murine and human osteosarcoma cells.
Reduced RhoA-GTPase Activity Is Involved in the Inhibition of Osteosarcoma Cell Invasion Potential by Atorvastatin-We previously showed that atorvastatin inhibits prenylation and consequently activity of RhoA-GTPase, triggering an apoptotic process (14). We thus evaluated the effect of geranylgeranylpyrophosphate (GGPP) repletion on the statin-induced inhibition of U2OS cell invasion. We found that the addition of 10 M GGPP prevented the inhibition of in vitro cell invasion induced by the statin (Fig. 4A). To fully confirm the implication of RhoA-GTPase activity in the inhibition of cell invasiveness by atorvastatin, we transfected U2OS cells with a vector encoding a constitutively active form of RhoA (CA-RhoA). This vector warded off the reduction in GTP-bound (active) RhoA level (Fig. 4B) by statin, independently of the prenylation level or membranous localization. Forced expression of CA-RhoA also abolished the reduction in U2OS cell invasion induced by atorvastatin (Fig. 4C). These results indicate that atorvastatin inhibits osteosarcoma cell invasion through inhibition of RhoA-GTPase prenylation and activity.
The Reduction in MMP2 Activity by Atorvastatin Is Dependent on RhoA-GTPase Activity-We then sought to determine the mechanisms underlying the reduction of MMP2 activity by atorvastatin in osteosarcoma cells. We first evaluated the role of RhoA-GTPase in the inhibitory effect of atorvastatin on MMP2 activity using GGPP (10 M), FPP (10 M), or mevalonate (10 mM) supplementation (14). We found that mevalonate or GGPP supplementation restored the basal MMP2 activity level in atorvastatintreated U2OS cells, whereas FPP supplementation had no effect (Fig. 4D). Additionally, forced expression of CA-RhoA prevented the inhibition of MMP2 activity induced by the statin (Fig. 4E), further documenting the link between RhoA-GTPase and MMP2 activity. These data indicate that GGPP depletion and inhibition of RhoA-GTPase activity by atorvastatin mediates the reduced MMP2 activity in osteosarcoma cells.
The Reduction in MMP2 Activity Is Mediated by Decreased JNK Signaling-We then investigated the signaling pathway involved in the reduction of MMP2 activity induced by the statin in osteosarcoma cells. We previously showed that reduced phosphoERK1/2 level is responsible for the pro-apoptotic effect of atorvastatin in osteosarcoma cells (14). In this study, we used U0126, a MEK inhibitor, and SP600125, an inhibitor of JNK, which efficiently reduced phospho-ERK1/2 (Fig. 5A) and phospho-JNK (Fig. 5B) protein levels, respectively. As shown in Fig.  5C, MMP2 activity was reduced by the JNK inhibitor (15 M) but not by the MEK inhibitor (10 M) in U2OS cells in basal conditions. Furthermore, pharmacological activation of JNK using anisomycin (0.01 g/ml for the last 4 h) prevented the reduction in MMP2 activity by atorvastatin in U2OS cells (Fig.  5D). Similar effects were found in the other osteosarcoma cell lines (data not shown). These data further suggest a critical role for JNK signaling in statin-induced reduction in MMP2 activity. Consistently, transfection of a vector encoding both MKK7 and JNK, which triggered JNK phosphorylation (Fig. 5E), prevented the reduction in MMP2 activity induced by atorvastatin ( Fig. 5F). To further confirm that JNK is a downstream target of HMG-CoA reductase inhibition in osteosarcoma cells, we investigated the change in translocation of c-Jun from the cytosol to the nucleus. As shown in Fig. 5G, atorvastatin markedly reduced c-Jun nuclear content. Additionally, the transcriptional activity through AP1 complex was dose-dependently reduced in U2OS cells incubated in the presence of atorvastatin (Fig. 5H). Taken together, these data indicate that the statin reduces MMP2 activity through inhibition of JNK phosphorylation and establish JNK-c-Jun as a downstream target of HMG-CoA reductase inhibition in osteosarcoma cells.
Inhibition of MMP2 Activity and Cell Invasion by Atorvastatin Results from Inactivation of RhoA-JNK Signaling-Finally, to evaluate the link between RhoA-GTPase and JNK signaling, we used mevalonate or downstream metabolites GGPP and FPP repletion to modulate the effect of atorvastatin on RhoA-GTPase prenylation. We found that supplementation with GGPP (10 M) or mevalonate (10 mM) restored phospho-JNK protein level in atorvastatin-treated U2OS cells, whereas FPP (10 M) repletion had no effect (Fig. 6A). This indicates that JNK activation is dependent on geranylgeranylation in osteosarcoma cells. This is supported by the finding that transfection with the vector encoding CA-RhoA increased the phospho-JNK level (Fig. 6B). To determine the functional link between RhoA-GTPase and JNK, U2OS cells were transfected with the vector encoding both MKK7 and JNK, which increased the phospho-JNK level (Fig. 6B). Forced expression of MKK7-JNK fully prevented the reduced invasion induced by atorvastatin (Fig. 6C). These data indicate that HMG-CoA reductase inhibition by the statin decreases the number of invading cells through a mechanism involving inactivation of the RhoA-JNKc-Jun signaling pathway.
Taken together, our results support a model in which HMG-CoA reductase inhibition by the statin reduces RhoA-GTPase activity through inhibition of geranylgeranylation, leading to reduction in JNK phosphorylation and c-Jun nuclear translocation, which results in reduced MMPs mRNA and protein expression and inhibition of MMP2 activity, which provides a critical role for JNK signaling in statin-induced reduction in MMP2 activity and invasion process (Fig. 6D).

DISCUSSION
In this study, we show that statin-induced HMG-CoA reductase inhibition reduces cell migration and invasion in osteosarcoma cells, and we identify a molecular mechanism involving JNK-c-Jun-MMP2 activity in these effects. These data point to a potential antitumoral effect mediated by inhibition of HMG-CoA reductase and c-Jun-MMP2 activity in osteosarcoma cells.
We first show that atorvastatin reduces in vitro cell migration and invasion in a panel of osteosarcoma cell lines of human and murine origin, suggesting that this inhibitory effect is independent of the cell genotype. The inhibitory effect of atorvastatin on cell migration was not due to decreased cell viability because blocking caspases activity did not overcome inhibition of cell migration induced by the statin. This suggests a direct effect of the statin on tumor cell invasion. We hypothesized that the reduced osteosarcoma cell invasion induced by the statin may result from reduced matrix metalloproteinases (MMPs) activity. Indeed, abnormal expression of MMPs is believed to play an important role in tumor cell invasion and progression in several cancers (2, 6 -8) including osteosarcoma (29). MMP2 and MMP9 are cell surface-associated type I collagen-degrading MMPs that are highly expressed and secreted in invasive human osteosarcoma cells, suggesting that MMPs are key factors for cancer cell invasion (29). Notably, Runx2-regulated MMP9 levels are functionally related to the invasion properties of cancer cells (30), and MMP9 expression can be altered by statins in nonskeletal cells (31)(32)(33)(34) and osteosarcoma cells (19). In the present study, we show that atorvastatin dose-dependently reduced MMP2 activity in invading osteosarcoma cells. Interestingly, the statin did not alter MMP2 activity in noninvading (MG63 cells) or normal osteoblastic cells, suggesting that the statin may preferentially target MMPs in invading osteosarcoma cells. A crucial pathway that leads to the activation of the proenzyme proMMP2 is its activation by a complex including TIMP2 and membrane-type metalloprotease-1 (MT1-MMP/MMP14) (27). Importantly, we found that atorvastatin reduced the expression of MMP2, MMP14 as well as TIMP2, indicating that the statin reduced the expression of all members of the MMP2 activating complex. Our finding that forced expression of MMP2 and/or MMP14 overcame the inhibitory effect of atorvastatin on cell invasion strongly indicates that these MMPs are involved in osteosarcoma invasive potential. This also indicates that the reduced osteosarcoma cell invasion induced by the statin results in large part from the inhibition of MMP2 expression and function. The inhibitory effect of statin on MMP2 activity may be relevant to the metastatic potential because the MMP2 activation state correlates with osteosarcoma cell invasion (35).
Having established that osteosarcoma cell migration and invasion can be reduced by the statin, we investigated the signaling mechanisms involved in the reduced MMP2 activity and cell invasion. Mitogen-activated protein kinases (MAPKs) are major signaling pathways controlling MMPs (36). However, both ERK1/2 and JNK were found to modulate the production of MMP2 in osteosarcoma cells (37,38). In our study, we found that JNK, but not ERK1/2 inhibition, reduced MMP2 activity, which points to a role for JNK in the regulation of MMP2 activity in osteosarcoma cells. This is supported by our finding that JNK activation by MKK7 overexpression or by treatment with anisomycin overcame atorvastatin-reduced MMP2 activity and prevented the decrease in cell invasion induced by the statin. This identifies for the first time a link between HMG-CoA reductase, JNK, and MMP2 activity and suggests the implication of JNK in the osteosarcoma invasion process.
The transcription factor c-Jun is an important mediator of several cellular processes (39 -41). Phosphorylation of c-Jun by JNK results in translocation of the Fos-Jun complex to the nucleus and activation of promoters that contains AP-1-binding sites, including the MMP2 promoter (42). Our finding that statin reduced c-Jun translocation to the nucleus suggests that HMG-CoA reductase inhibition targets the JNK-c-Jun signaling pathway to modulate MMP2 activity and osteosarcoma cell invasion. These data support a functional link between JNK, MMP2, and osteosarcoma cell invasion and identify JNK as a novel potential target for anti-invasive action.
We then sought to determine how the statin may lead to the inhibition of JNK-MMP2 in osteosarcoma cells. Inhibition of HMG-CoA reductase is known to lead to deprivation in mevalonate and downstream metabolites such as GGPP or FPP. We found that mevalonate or GGPP repletion reversed the statininduced reduction in MMP2 activity and osteosarcoma cell invasion, implying a role for GGPP rather than FPP in the downstream events. This is supported by our finding that mevalonate or GGPP supplementation restored phospho-JNK levels in statin-treated osteosarcoma cells. These data point to a predominant role of geranylgeranylation in the regulation of the JNK-MMP2-cell invasion cascade. One possible downstream target is the GTPase RhoA, a molecule that undergoes geranylgeranylation. Our finding that overexpression of a constitutively active form of RhoA increased phospho-JNK level and overcame the inhibitory effect of atorvastatin on both MMP2 activity and cell invasion clearly indicates that RhoA-GTPase plays a major role in the anti-invasive effect of statin in osteosarcoma cells.
Overall, our data reveal a molecular mechanism by which inhibition of HMG-CoA reductase by the statin and subsequent decreased RhoA-GTPase prenylation results in reduced JNK-c-Jun signaling, decreased MMP2 activity, and cell invasion in osteosarcoma cells (Fig. 6D). These in vitro findings may be relevant to in vivo tumorigenesis because increased expression of MMP2 correlates with poor prognosis in human osteosarcoma (43), and tumor cells often show elevated HMG-CoA reductase activity (44) and increased JNK signaling (11, 13,  [45][46][47]. Notably, aggressive osteosarcoma tumors show increased Jun expression (45,46,48,49), which may contribute to tumor development and invasion. Recent data indicate that activating JNK and subsequent MMP2 expression and secretion correlate with cancer cell migration and invasion (50,51). The finding that HMG-CoA reductase inhibition reduces cell invasion and migration in osteosarcoma cells by reducing RhoA/JNK signaling and MMP2 activity suggests a novel strategy to target RhoA-JNK-c-Jun to reduce osteosarcoma cell invasiveness and tumorigenesis.