Mevalonate Cascade Inhibition by Simvastatin Induces the Intrinsic Apoptosis Pathway via Depletion of Isoprenoids in Tumor Cells

The mevalonate (MEV) cascade is responsible for cholesterol biosynthesis and the formation of the intermediate metabolites geranylgeranylpyrophosphate (GGPP) and farnesylpyrophosphate (FPP) used in the prenylation of proteins. Here we show that the MEV cascade inhibitor simvastatin induced significant cell death in a wide range of human tumor cell lines, including glioblastoma, astrocytoma, neuroblastoma, lung adenocarcinoma, and breast cancer. Simvastatin induced apoptotic cell death via the intrinsic apoptotic pathway. In all cancer cell types tested, simvastatin-induced cell death was not rescued by cholesterol, but was dependent on GGPP- and FPP-depletion. We confirmed that simvastatin caused the translocation of the small Rho GTPases RhoA, Cdc42, and Rac1/2/3 from cell membranes to the cytosol in U251 (glioblastoma), A549 (lung adenocarcinoma) and MDA-MB-231(breast cancer). Simvastatin-induced Rho-GTP loading significantly increased in U251 cells which were reversed with MEV, FPP, GGPP. In contrast, simvastatin did not change Rho-GTP loading in A549 and MDA-MB-231. Inhibition of geranylgeranyltransferase I by GGTi-298, but not farnesyltransferase by FTi-277, induced significant cell death in U251, A549, and MDA-MB-231. These results indicate that MEV cascade inhibition by simvastatin induced the intrinsic apoptosis pathway via inhibition of Rho family prenylation and depletion of GGPP, in a variety of different human cancer cell lines.


Results
Simvastatin induces concentration-and time-dependent cell death in human brain, lung, and breast cancer cell lines. We Fig. 2A-F). As shown in Fig. 2D, an apoptotic cell population was identified as sub-G 1 population in flow cytometry as had been described by us previously 27 . Simvastatin (10 μ M, 48 h) failed to activate caspase-8 but did activate caspase-9 and caspase-3/-7 in all cell lines tested at 36 h ( Fig. 2E-G) ( Supplementary Fig. 2G-L). These findings indicated that MEV cascade inhibition induced the intrinsic apoptotic pathways in all tumor cell lines tested. The involvement of the intrinsic apoptosis pathway was further confirmed by measuring mitochondrial membrane potential in simvastatin-treated (10 μ M, 36 h) cells (U87, A549, MDA-MB231) (Fig. 2L-O). Simvastatin significantly reduced mitochondrial membrane potential in these tumor cells (U87, P < 0.01, A549, P < 0.05, and MDA-MB231, P < 0.001). We also showed that shorter treatment of simvastatin (10 mΜ , 18 h) did not change mitochondrial membrane potential in these cells Simvastatin-induced cell death in human cancer cells is not rescued by cholesterol but is dependent on depletion of FPP and GGPP. Statin-induced apoptosis is due to a loss of cell membrane cholesterol and/or depletion of the polyisoprene cholesterol precursors, FPP and GGPP, which are essential lipid anchors for active small GTPase proteins in cells 5,34,35 . We used different approaches to identify the mechanism by which MEV cascade inhibition caused apoptosis. First, we performed treatment of tumor cells (U87, U251 Our results also showed that simvastatin significantly decreased total cholesterol (P < 0.01) and de novo cholesterol biosynthesis (P < 0.001) in U251 cells (Fig. 3G,H). Simvastatin did not significantly change total cholesterol or de novo cholesterol biosynthesis (P > 0.05) in A549 (Fig. 3I,J) and MDA-MB231 cells (Fig. 3K,L).
Simvastatin blocks membrane translocation of Rho family small GTPases. We investigated the ability of simvastatin to affect the translocation of small Rho GTPases from the plasma membrane to the cytosol in U251 (Fig. 5A), A549 (Fig. 5C), and MDA-MB231 (Fig. 5E). Simvastatin decreased membrane localization of RhoA, Cdc42, and Rac/1/2/3 and their cytosolic cellular fraction suggesting simvastatin-mediated block in prenylation of small GTPases and poor membrane translocation. We measured GTP-bound Rho protein and showed that simvastatin significantly (P < 0.05) increased GTP-bound Rho in U251 (Fig. 5B), whereas mevalonate, FPP, and GGPP co-treatment decreased GTP-bound Rho to untreated levels. Simvastatin did not significantly change   Results are expressed as mean ± SD of 9 replicate in 3 independent experiments (*P < 0.05, **P < 0.01, and ***P < 0.001). U87, A549, and MDA-MB231 cells were treated with simvastatin (10 μ M) and after 36 hrs total and de novo cholesterol content in cells were measured. Both total and de novo cholesterol were significantly decreased in U251 cells (G,H) after simvastatin treatment. However, there was no significant change in the amount of total and de novo cholesterol for A549 (I,J) and MDA-MB231 cells (K,L). For each experiment control cells were treated with DMSO. Results are expressed as mean ± SD of 3 replicate 3 independent experiments (*P < 0.05, **P < 0.01, and ***P < 0.001). induced significant (P > 0.001) cell death in U87, A549, and MDA-MB231 cells (36 and 60 h, Fig. 6G-L), indicating that gernylgernalytion was the leading mechanism in simvastatin-induced cell death in these cancer cells.

Discussion
In the present study, we found that MEV cascade inhibition by the HMGCR inhibitor simvastatin induced intrinsic apoptosis cell death and decreased mitochondrial membrane potential in a broad range of human tumor cell lines, including glioma (U87, U251), neuroblastoma (SH-SY5Y), lung (A549, H460, H1650, H1975), and breast cancer cells (MCF7, MDA-MB231). Inhibition of the MEV cascade prevented the membrane translocation of the small Rho GTPases RhoA, Cdc42 and Rac1/2/3 and this simvastatin effect was independent of cellular cholesterol but partially rescued by FPP and GGPP. Inhibition of FTase had no effect on cell death and FPP supplementation caused partial reversal of the simvastatin-mediated effects. FPP can also be converted into GGPP and the addition of FPP can restore farnesylation and geranylgeranylation. We also showed that, at least in U251 brain tumor cell line, the membrane localization of prenylated Rho GTPases was important for GTPase activity. Our findings are summarized in Fig. 7.
Simvastatin also induces pro-apoptotic Bcl2 family members Bcl2 and Bax in MCF-7 and MDA-MB231 cells in a time-and dose-dependent manner 38,39 . These effects were also caspase-dependent since simvastatin increased caspase-3 and -9 activity in MDA-MB231 cells 39 . Similarly, fluvastatin and atorvastatin, were also reported to induce dose-and time-dependent apoptosis in MCF-7 and MDA-MB-231 breast cancer cell lines 40,41 . Several mechanisms have been proposed for statin-induced cell death in breast cancer cells including an increase of  reactive oxygen species (ROS) 42 , the downregulation of survivin expression 43 , increased nitric oxide synthase activity (iNOS or NOS II) via geranylgeranylation 44 , and G1/S cell cycle arrest due to an increase in p21(Waf1/ Cip1) 45 .
Simvastatin was shown to induce caspase-dependent apoptosis in A549 cells which is regulated by the Bcl2 family proteins and ROS and can be reversed by treatment with N-acetylcysteine 46,47 . Other investigators have reported that simvastatin can induce apoptotic pathways in A549 and H460 cells by blocking the

Figure 7. Summary of the mechanism involved in statin-induced cell death in cancer cells. MEV cascade inhibitors induce the intrinsic apoptotic pathway which is regulated by gernaylgenralyation of small Rho GTPAse protein.
cell cycle and down-regulating cyclin D1 and cyclin dependent kinases (CDKs) expression 48,49 . In A549 cells, simvastatin-induced cell death was associated with decreased expression of survivin 50 , like in breast cancer cells.
Our previous investigations in human airway smooth muscle and fibroblasts and in human atrial fibroblasts showed that simvastatin induces the intrinsic apoptosis pathways in a small GTPase prenylation dependent way. This statin effect was not reversible with cholesterol co-treatment 27,28,51 . It has been also shown that GGTI-298, but not FTI-277, can mimic the cytotoxic effects of statins, indicating that statins induce cell death by inactivating Rho/Rac GTPase activities 52 . These results were also confirmed by translocation of RhoA, Rac1/2/3, and Cdc42 to cytosol in simvastatin-treated cells. In eukaryotic cells, prenylation is carried out by three different prenyl transferases: farnesyl transferase (FT), geranylgeranyl transferase I (GGTI) and Rab geranylgeranyl transferase (Rab GGT or GGTII) 53 . FT is responsible for prenylation of proteins such as Ras and lamins. The GGTI catalyses the geranylgeranylation of proteins in the Rho and Rac family, whereas the Rab GGT is responsible for the geranylgeranylation of the Rab protein family 54 . Simvastatin increases Rac GTP loading in THP-1 monocyte cells while decreasing prenylation of Rac in the presence of amyloid beta stimulation, and decreasing the induction of inflammation in these cells 55 . In addition, T-cell function is not affected by Rho GTP loading, whereas geranylgeranylation of these small Rho GTPases is the determining step that affects their function 56,57 . Our results showed that statin-induced Rho protein GTP loading is dependent to specific cell type and Rho protein localization and geranylgeranylation is the determining step in regulation of their function (Figs 5, 6 and 7).
One of the signaling proteins involved in transmitting extracellular stimuli to intracellular components is the Ras (Rat sarcoma) superfamily of small GTPases 5,58 . In normal cellular and biological conditions these proteins play essential functions in the regulation of pathways critical to cytoskeletal reorganization, cell survival and proliferation, transformation, and vesicular trafficking 59,60 . One of the major subgroups in Ras superfamily of small GTPases is Rho (Ras homologous) family GTPases 61 . The members of small Rho GTPase (Rho, Rac, and Cdc42) are well known for their key functions in regulating actin cytoskeleton controlling actin stress fibers [62][63][64] .
The function of small GTPases is very tightly regulated by different molecular switches, including prenylation and guanosine triphosphate (GTP) binding. When bound to GTP and prenylated, small GTPases not only translocate to the membrane but also undergo a conformational change to engage effectors that promote downstream signaling pathways 5,65 .
In addition to their role in normal physiological and developmental processes, Rho GTPases can contribute to pathological processes including cancer cell migration, invasion, metastasis, and inflammation 66,67 . Activating mutations in Ras proteins (such as K-Ras, N-Ras, and H-Ras) are found in 15 to 30% of human tumors 68 80 , and testicular carcinomas 76,77,[79][80][81] . Cdc42 is overexpressed in breast 69 and testicular cancers 76 . Therefore, inhibition of these over-expressed Rho small GTPases might improve current therapeutic approaches in the treatment these cancers.
There is increasing interest in repurposing statins for use in the treatment of human cancers. An epidemiologic study has shown that statin use in patients with cancer was associated with reduced cancer-related mortality 82 . Our results showed that the HMGCR inhibitor simvastatin induces intrinsic apoptotic cell death in different cancer cell models via a unique small Rho GTPase-dependent pathway which prevents small Rho GTPase prenylation and inhibits subsequent translocation to the membrane, thus, effectively deactivating Rho GTPases. Interestingly, our current investigation suggests that cholesterol depletion is not involved in simvastatin-induced apoptosis in glioblastoma, neuroblastoma, non-small lung cancer cells, and breast cancer cell lines.
In our forthcoming studies, we are studying how the combination of simvastatin with different chemotherapeutic agents may synergize to enhance cancer cell killing. We will explore mechanisms of cell fate that could mediate this beneficial statin effect, thereby supplementing and enhancing the therapeutic effects of cancer medications.
Cell Viability Assay. We measured the viability of different cancer cells under various treatment conditions, as described previously, using MTT assay 28,[83][84][85]

Measurement of Apoptosis by Flow Cytometry.
The Nicoletti method was used to measure cellular apoptosis 86,87 . Briefly, cells cultured in 12 well plates were treated with simvastatin (10 μ M, 48 h). Cells were detached using EDTA buffer and harvested by centrifugation at 1500 g for 5 min at 4 °C. Cells were washed once in PBS before resuspending in a hypotonic PI lysis buffer (0.1% Triton X-100, 1% sodium citrate, 0.5 mg/ml RNase A, 40 μ g/ml propidium iodide). Cell nuclei were incubated at 37 °C for 30 min and analyzed by flow cytometry. Apoptotic nuclei were located on the left side of the G1 peak and contained hypo-diploid DNA.
Luminometric Caspase Assay. For the proteolytic activity of caspase-8 (IETD-ase), -3/-7 (DEVD-ase), -9 (LEHDase), Caspase-Glo ® -3/-7, -8, and -9 (Promega) were determined in luminometric assays according to the manufacturer's instructions and our previous report 51,87 . Briefly, cells were grown in 96-well plate (15,000 cells/well) and treated with simvastatin (10 μ M, 36 h). Fresh caspase reagents were prepared containing z-LETD-Luciferin, z-DEVD-Luciferin or z-LEHD-Luciferin and whole protein cell lysate extract buffer. Cells treated only with medium and reagent blank (negative controls) were included in each experiment. Plates were shaken gently at 300-500 rpm for 30 sec and then incubated for 90 min at room temperature. The solution was transferred to a white-well plate and then the luminescence was measured for each sample and compared to negative control values 51,87 . TMRM Staining for Mitochondrial Membrane Potential Measurement. U87, A549, and MDA-MB231 cells were cultured in 6 well plates and were treated with simvastatin (10 μ M, 36 h) and then were stained with (Tetramethylrhodamine, Methyl Ester, Perchlorate TMRM (100 nM) and Hoechst (10 μ M) nuclear stain at 37 °C for 30 minutes. After washing, the cells they were imaged using florescence microscope and the fluorescence was quantified on ImageJ software in at least 50 individual cells in different views (NIH, Bethesda, MD, USA) 88 .

Measurement of Rho GTPase activity.
Rho-GTP bound was measured in snap-frozen cell lysates harvested from cells cultured in medium without FBS and treated with simvastatin (10 μ M), mevalonate (2.5 mM), FPP, and GGPP (15 μ M), Simva. + mevalonate, simvastatin + FPP, and simvastatin + GGPP for 36 h. We used a luminometric-based G-LISA Rho -GTP bound assay (Cytoskeleton, Inc, Denver, Colo) for U251, A549, and MDA-MB231 cells Briefly, cell lysates were subjected to Rho binding domain in a Rho-GTP affinity 96-well plate (Cytoskeleton, Inc, Denver, Colo). Rho-GTP was detected with specific primary antibody, followed by horseradish peroxidase-conjugated secondary antibody detection and development with a chemiluminescent reagent. A constitutively active Rho-GTP provided in the kit was used as positive control in all experiments 90 .
Scientific RepoRts | 7:44841 | DOI: 10.1038/srep44841 Cholesterol mass measurement. U87, A549, and MDA-MB231 cells were treated with simvastatin (10 μ M, 36 h) and subsequently cholesterol content of cells was determined by colorimetric reaction using the Amplex Red Cholesterol assay kit (Invitrogen) as per the manufacturer's instructions. Lipids were extracted and cholesterol isolated on thin layer chromatography plates prior to analysis as described 27,[91][92][93] . All isolates were measured immediately after drying down under N 2 .
Lipids were extracted and cholesterol isolated on thin layer chromatography plates as above. Spots corresponding to cholesterol standards were removed, and radioactivity incorporated into cholesterol determined by liquid scintillation counting as described 92 . Immunoblotting. Western blotting was used to detect Cdc42, Rac1/2/3 and RhoA in membrane and cytosolic fractions as described previously 51,94 . Pan-cadherin and GAPDH were used to confirm membrane and cytosolic fraction purity, respectively. Briefly, cells protein extracts were prepared in lysis buffer (20 mM Tris-HCl (pH 7.5), 0.5 mM PMSF, 0.5% Nonidet P-40, 100 μ M β -glycerol 3-phosphate and 0.5% protease inhibitor cocktail). Supernatant protein content was measured by Lowry protein assay after centrifugation at 13,000 g for 10 min. Proteins were separated by SDS-PAGE and transferred to nylon membranes under reducing conditions. Membranes were blocked with non-fat dried milk and Tween 20 followed by overnight incubation with the primary antibodies at 4 °C, followed by incubation with HRP-conjugated secondary antibody for 1 h at room temperature. Blots were developed by enhanced chemiluminescence (ECL) detection (Amersham-Pharmacia Biotech).
Statistical Analysis. The results were expressed as means ± SEM and statistical differences were evaluated by one-way or two-way ANOVA followed by Tukey's or Bonferroni's post hoc testing, using Graph Pad Prism 7.0. A p-value < 0.05 was considered statistically significant. For all experiments data were collected in five replicates and in three separate experiments.