Amel Dudakovic, Andrew J. Wiemer, Kimberly M. Lamb, Laura A. Vonnahme, Sara E. Dietz, and Raymond J. Hohl

Transcription

1 /08/ $20.00 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 324, No. 3 Copyright 2008 by The American Society for Pharmacology and Experimental Therapeutics / JPET 324: , 2008 Printed in U.S.A. Inhibition of Geranylgeranyl Diphosphate Synthase Induces Apoptosis through Multiple Mechanisms and Displays Synergy with Inhibition of Other Isoprenoid Biosynthetic Enzymes Amel Dudakovic, Andrew J. Wiemer, Kimberly M. Lamb, Laura A. Vonnahme, Sara E. Dietz, and Raymond J. Hohl Department of Pharmacology (A.D., R.J.H.), Interdisciplinary Program in Molecular and Cellular Biology (A.J.W., R.J.H.), and Department of Internal Medicine (K.M.L., L.A.V., S.E.D., R.J.H.), University of Iowa, Iowa City, Iowa Received September 27, 2007; accepted December 13, 2007 ABSTRACT Inhibitors of isoprenoid synthesis are widely used for treatment of human diseases, including hypercholesterolemia and osteoporosis, and they have the potential to be useful for treatment of cancer. Statin drugs inhibit the enzyme HMG-CoA reductase, whereas nitrogenous bisphosphonates have more recently been shown to inhibit farnesyl disphosphate synthase. In addition, our laboratory has recently developed several potent and specific bisphosphonate inhibitors of geranylgeranyl diphosphate synthase, including digeranyl bisphosphonate. Because all three enzymes fall in the same biosynthetic pathway and many of the biological effects are due to depletion of downstream products, we hypothesized that simultaneous inhibition of these enzymes The mevalonate pathway (Fig. 1) is one of the most targeted biochemical pathways in human disease with millions of people currently taking statins or bisphosphonates. The statins, including lovastatin, have been used for several years to treat hypercholesterolemia (Grundy, 1988). The statins inhibit HMG-CoA reductase (HMGCR), depleting cells of downstream isoprenoids, including farnesyl diphosphate (FPP) and geranylgeranyl diphosphate (GGPP) (Brown and Goldstein, 1980). FPP is a key intermediate in synthesis of cholesterol, and its depletion results in decreased de novo cholesterol synthesis (Andersen and Dietschy, 1978) leading This work was supported by the Roy J. Carver Charitable Trust as a Research Program of Excellence, the Roland W. Holden Family Program for Experimental Cancer Therapeutics, and the Iowa Center for Research by Undergraduates. A.D. and A.J.W. contributed equally to this article. Article, publication date, and citation information can be found at doi: /jpet would result in synergistic growth inhibition. In this study, we show that inhibition of geranylgeranyl diphosphate synthase induces apoptosis in K562 leukemia cells. This induction of apoptosis is in part dependent upon both geranylgeranyl diphosphate depletion and accumulation of farnesyl diphosphate. Combinations of either lovastatin or zoledronate with digeranyl bisphosphonate synergistically inhibited growth and induced apoptosis. These combinations also potently inhibited cellular geranylgeranylation. These results support the potential for combinations of multiple inhibitors of isoprene biosynthesis to inhibit cancer cell growth or metastasis at clinically achievable concentrations. to increased low-density lipoprotein receptor expression and decreased serum low-density lipoprotein cholesterol in vivo (Goldstein and Brown, 1990). The isoprene moieties from FPP and GGPP are also post-translationally incorporated into several proteins, including many members of the Ras family of small GTPases, which control cell growth and proliferation (Takai et al., 2001), and the Rho family of GTPases, which are important mediators of cell migration (Burridge and Doughman, 2006). Because prenylation is necessary for the activation of small GTPases, statins have been investigated as potential cytotoxic agents for use in cancer chemotherapy (Swanson and Hohl, 2006). The nitrogenous bisphosphonates (NBPs), including zoledronate are diphosphate analogs used clinically to treat bone disorders, including osteoporosis and metastatic bone disease (Licata, 2005). These drugs have also been shown in the laboratory to induce apoptosis, and they have direct growth inhibitory effects on malignant cells (Shipman et al., ABBREVIATIONS: HMGCR, HMG-CoA reductase; FPP, farnesyl diphosphate; GGPP, geranylgeranyl diphosphate; NBP, nitrogenous bisphosphonate; FDPS, farnesyl diphosphate synthase; GGDPS, geranylgeranyl diphosphate synthase; DGBP, digeranyl bisphosphonate; TUNEL, terminal deoxynucleotidyl transferase dutp nick-end labeling; PI, propidium iodide; PCR, polymerase chain reaction; PARP, poly(adp-ribose) polymerase; RT, room temperature; PBS, phosphate-buffered saline; CI, combination index; GGOH, geranylgeraniol; GGTase, geranylgeranyl transferase; FOH, farnesol; PC, phosphatidylcholine; SQS, squalene synthase; Lov, lovastatin; GGTI-298, N-4-[2(R)-amino-3-mercaptopropyl]amino-2-naphthylbenzoyl-(L)-leucine methyl ester trifluoroacetate. 1028

2 GGDPS Apoptosis and Synergy 1029 Fig. 1. Biosynthesis of geranylgeranyl diphosphate and relevant inhibitors of isoprenoid biosynthesis. HMGCR is the rate-limiting enzyme in the biosynthesis of cholesterol, and it is inhibited by statin drugs, including lovastatin. Zoledronate and other NBPs inhibit FDPS. Both zoledronate and lovastatin deplete cells of FPP, GGPP, and other downstream products. DGBP inhibits GGDPS and depletes GGPP. 1997). Like statins, the clinically used NBPs deplete cells of isoprenoid diphosphates, resulting in the inhibition of posttranslational protein prenylation (Reszka and Rodan, 2004). In particular, the NBPs inhibit FPP synthase (FDPS) (van Beek et al., 1999b; Dunford et al., 2001). However, their cellular effects, including induction of apoptosis, may be largely a result of downstream GGPP depletion (Fisher et al., 1999; van Beek et al., 1999a). Based on this rationale, we developed a series of isoprenoid-containing bisphosphonates that specifically inhibit geranylgeranyl diphosphate synthase (GGDPS) (Shull et al., 2006; Maalouf et al., 2007; Wiemer et al., 2007). We have shown that several of these compounds, including digeranyl bisphosphonate (DGBP), can potently inhibit GGDPS in vitro (Wiemer et al., 2007). We have also shown that inhibition of cellular GGDPS leads to depletion of intracellular GGPP levels and not depletion of FPP (Wiemer et al., 2007). Both statins and NBPs have limited potential as singleagent treatments of malignancies. Inhibition of protein prenylation by the statins does not occur at standard concentrations for treatment of hypercholesterolemia (Lewis et al., 2005). NBPs are limited as anticancer agents because of high bone affinity and low plasma concentrations (Green and Clezardin, 2002). Therefore, if these agents are to be useful for treatment of malignant disease, it is possible that they will need to be used in combination with other agents. This strategy has led to the identification of an increasing number of synergistic interactions between inhibitors of isoprenoid biosynthesis and chemotherapeutic agents (Santini et al., 2006). For example, statins are synergistic with cytosine arabinoside (Holstein and Hohl, 2001a) and paclitaxel (Holstein and Hohl, 2001b). NBPs are synergistic with farnesyl transferase inhibitors (Caraglia et al., 2004), paclitaxel (Jagdev et al., 2001), and imatinib (Kuroda et al., 2003). It is noteworthy that the combination of statins and NBPs is synergistic (Schmidmaier et al., 2006). Although HMGCR and FDPS have now been extensively studied as therapeutic targets, GGDPS has not. Our recent advances in the design of potent, specific, and cell-permeable GGDPS inhibitors allow the consequences of GGDPS inhibi-

3 1030 Dudakovic et al. tion to be studied. In this study, we show for the first time that inhibition of GGDPS inhibits growth and induces apoptosis through at least two mechanisms, GGPP depletion and more surprisingly, the accumulation of FPP. It is noteworthy that inhibition of GGDPS by digeranyl bisphosphonate is synergistic with either HMG-CoA reductase inhibition by lovastatin or FDPS inhibition by zoledronate. Materials and Methods Cell Culture and Incubations. K562 leukemia cells were obtained from American Type Culture Collection (Manassas, VA), and they were cultured accordingly. For all experiments, cells were incubated in fresh RPMI 1640 medium containing 10% fetal bovine serum at a concentration of cells/ml. For Western blots and terminal deoxynucleotidyl transferase dutp nick-end labeling (TUNEL) assays, total cells were incubated for times and concentrations indicated. Annexin V and propidium iodide (PI) analysis required total cells. Real-time PCR experiments required total cells. DNA Synthesis Assays. K562 cells were incubated in 96-well plates and treated with compounds as described previously (Shull et al., 2006). After 22 h, 20 l of[ 3 H]thymidine ( TBq/mmol; 3.75 Ci/mmol in media) was added to each well. At 24 h, cells were filtered through glass microfiber paper using a Brandel (Gaithersburg, MD) cell harvester. [ 3 H]Thymidine incorporated into cellular DNA was quantified by scintillation counting. Annexin V and PI Analysis. Annexin V and PI analysis was adapted from the technical data sheet for fluorescein isothiocyanate annexin V (556419; BD Biosciences PharMingen, San Diego, CA), with slight modifications. Treated cells were transferred to 1.5-ml microcentrifuge tubes, they were centrifuged at 1000 rpm for 10 min, and then the supernatant was aspirated. Cells were resuspended in 500 l of buffer (10 mm HEPES, 150 mm NaCl, 1 mm MgCl 2,5mM KCl, and 1.8 mm CaCl 2, ph 7.4), and they were transferred to polystyrene test tubes. Five microliters of FITC annexin V was added, and the cells were vortexed and then incubated 15 min at room temperature (RT). Ten microliters of 50 g/ml PI solution (Sigma-Aldrich, St. Louis, MO) was added to the cell suspension, and the suspension was vortexed. Samples were analyzed using FACScan (BD Biosciences, Franklin Lakes, NJ). Western Blot Analysis. Protein concentrations were determined using the bicinchoninic acid method (Pierce Chemical, Rockford, IL). All proteins except poly(adp-ribose) polymerase (PARP) were resolved by electrophoresis on a 12% gel, and they were transferred to a polyvinylidene difluoride membrane. PARP was resolved using 7.5% gels. Primary and secondary antibodies were added sequentially for 45 min, and proteins were visualized using an enhanced chemiluminescence detection kit (GE Healthcare, Little Chalfont, Buckinghamshire, UK). Anti pan-ras was obtained from InterBiotechnology (Tokyo, Japan). Rap1a (sc-1482), Rab6 (sc-310), PARP (sc-7150), and -tubulin (sc-8035) were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Caspase-3 (9662) was obtained from Cell Signaling Technology Inc. (Danvers, MA). Horseradish peroxidase-conjugated anti-mouse secondary was obtained from GE Healthcare, and Horseradish peroxidase-conjugated anti-goat was obtained from Santa Cruz Biotechnology, Inc. After visualization, proteins were quantified using ImageJ software ( nih.gov/ij/). TUNEL Assay. Promega (Madison, WI) TUNEL assay kit (G3250) was performed according to the manufacturer s protocol. Cells were washed twice in PBS, resuspended in 0.5 ml of PBS, and fixed by the addition of 5 ml of methanol-free formaldehyde for 20 min on ice. Cells were washed twice, 5 ml of 70% ice-cold ethanol was added, and cells were stored overnight at 20 C. Washing in PBS was repeated, and cells were resuspended in 1 ml of PBS. The contents were centrifuged, and the cell pellet was resuspended in 80 l of equilibrium buffer and incubated at RT for 5 min. The contents were centrifuged, and cell pellet was resuspended with 50 l of recombinant terminal deoxynucleotidyl transferase incubation buffer, incubated at 37 C for 1 h in the dark. Reaction was terminated by addition of 1 ml of 20 mm EDTA. Cells were washed twice in 1 ml of 0.1% Triton X-100 solution in PBS containing 5 mg/ml bovine serum albumin, resuspended in 0.5 ml of 5 g/ml PI solution containing 250 g of DNase-free RNase A, and then they were incubated at RT for 30 min, and finally analyzed by flow cytometry. FPP and GGPP Quantification. FPP and GGPP levels were determined as described previously (Tong et al., 2005). In brief, isoprenoid diphosphates were extracted, and they were used as substrates for incorporation into fluorescent CAAX peptides by farnesyltransferase or geranylgeranyl transferase. Prenylated fluorescent peptides were quantified by fluorescence detection. Quantification of mrna. Treated cells were lysed with Invitrogen (Carlsbad, CA) TRIzol reagent, and total RNA were extracted accordingly. An ABI reaction kit (Applied Biosystems, Foster City, CA) was used to synthesize the cdna for each condition by reverse transcription. Primers for specific genes were mixed with SYBR Green intercalating dye, and they were added to cdna. PCR was allowed to proceed for 40 cycles. Data were analyzed using ABI SDS 2.3 software (Applied Biosystems), normalized to 18S ribosomal RNA, and quantities were determined using the relative standard curve method as described by the manufacturer. Primers were designed using PrimerQuest (Integrated DNA Technologies, Inc., Coralville, IA). The PCR product was designed to be 200 to 400 base pairs, and primers were checked for specificity with Basic Local Alignment and Search Tool. The following primers were used: HMGCR (5 -ACAGGCTTGAATGAAGCTTTGCCC-3, 5 -GACAT- GCAGCCAAAGCAGCACATA-3 ), FDPS (5 -CTTCCTGCAGAGTT- CCTATCAGAC-3, 5 -TCTCCAGCAGGATCTTCTTGGCAT-3 ), and SQS (5 -ACTTCCCAACGATCTCCCTTGAGT-3, 5 -TCCAAACCTC- TTGAGGCCAGAACT-3 ). Isobologram and Combination Index Analysis. Isobolograms were generated using Calcusyn software (Biosoft, Cambridge, UK). Combination index (CI) values were calculated according to the method of Chou and Talalay (1984) as described in the software manual. For each drug, 48-h IC 50 values were determined by thymidine incorporation. Concentration-response curves were generated for each drug and combination using IC 50 values as the highest concentration and four 80% dilutions. Statistical Analysis. Unpaired two-tailed t tests were used to calculate statistical significance. Unless otherwise indicated, comparisons were done relative to the control. All columns in bar graphs represent the mean of the indicated number of replicates. Error bars on graphs represent S.E. An level of 0.05 was set as the level of significance. Results Inhibition of Geranylgeranyl Diphosphate Synthase Inhibits Growth and Induces Apoptosis in Human Chronic Myelogenous Leukemia Cells. To determine whether GGDPS inhibition impairs growth of K562 human chronic myelogenous leukemia cells, cells were treated with DGBP for 48 h. The amount of cellular DNA synthesis was assessed by measuring levels of [ 3 H]thymidine incorporation (Fig. 2A). DGBP inhibited growth of these cells, with an IC 50 value of 55 M. To determine whether this effect was due to cell death, cells were treated with DGBP, and total cell death was measured using annexin V and PI staining. DGBP induced a concentration- and time-dependent increase in the number of apoptotic and necrotic cells (Fig. 2B). This increase was observed in both the early apoptosis fraction (annexin V /PI ) and the late apoptosis and necrosis fractions (annexin V /PI ) (data not shown).

4 Fig. 2. Inhibition of GGDPS induces growth inhibition and apoptosis in K562 chronic myelogenous leukemia cells. Cells were seeded and treated with different concentrations of DGBP as described under Materials and Methods. A, growth as a percentage of untreated control cells at 48 h was evaluated by [ 3 H]thymidine incorporation (mean S.E.; n 4). B, total apoptosis at varying times was determined by annexin V and PI staining (mean S.E.; n 2). C, induction of PARP and caspase-3 cleavage at varying times was determined by Western blot analysis. One representative Western blot is shown (n 3). Inhibition of Geranylgeranyl Diphosphate Synthase Leads to Induction of the Caspase Cascade in K562 Cells. To confirm the flow cytometric results, which showed that DGBP induces apoptosis, cells were treated with DGBP, and they were assessed for cleavage of PARP and caspase-3 by Western blot analysis. PARP is a DNA repair enzyme that is cleaved during apoptosis by executioner caspases such as activated caspase-3 (Schreiber et al., 2006). Caspases are cysteinyl aspartate-specific proteases that are activated by cleavage during early apoptosis (Thornberry and Lazebnik, 1998). Etoposide (50 M), which is a well-characterized activator of apoptosis, was used as a positive control. DGBP, like etoposide, induced time-dependent cleavage of PARP as demonstrated by the appearance of an 85-kDa band (Fig. 2C). PARP cleavage was first observed at 24 h, and the effect was more pronounced after 72 h, analogous to what was observed with etoposide. DGBP treatment also leads to the cleavage GGDPS Apoptosis and Synergy 1031 and activation of caspase-3 as indicated by the appearance of a 17-kDa cleavage product. As expected, etoposide treatment resulted in caspase-3 cleavage. These results indicated that DGBP activates the apoptosis machinery, leading to caspase-3 activation, PARP cleavage, and eventual cell death. Geranylgeraniol Supplementation Prevents Digeranyl Bisphosphonate-Induced Inhibition of Protein Geranylgeranylation. To determine whether product supplementation could prevent DGBP-mediated inhibition of protein geranylgeranylation, cells were treated with GGOH, which is phosphorylated to form GGPP (Crick et al., 1997). Cells were treated for 48 h with DGBP in the presence or absence of GGOH, and then they were analyzed for inhibition of Rap1a and Rab6 geranylgeranylation by Western blot analysis (Fig. 3). To maximize inhibition of prenylation and to limit apoptotic effects, DGBP was used at a concentration of 40 M for this experiment. Both Rap1a and Rab6 are exclusively geranylgeranylated; Rap1a is a substrate for geranylgeranyl transferase (GGTase) I, whereas Rab6 is a substrate for GGTase II. It should be noted that the antibody for Rap1a is for the C terminus, and as such, it detects only the unmodified form. The antibodies for Rab6 and Ras detect both modified and unmodified forms. As positive controls, cells were treated with lovastatin and zoledronate, which deplete cells of FPP and GGPP, and therefore limit farnesylation and geranylgeranylation. Cells were also treated with FOH (Fig. 3, lane 2) and GGOH (Fig. 3, lane 3), neither of which limits prenylation. Like GGOH, FOH is phosphorylated to form FPP. As expected, 20 M lovastatin limited Ras farnesylation and Rap1a and Rab6 geranylgeranylation (lane 5). Lovastatin-induced limitation of Ras farnesylation was prevented by addition of 1 mm mevalonate (Fig. 3, lane 6) or 10 M FOH (Fig. 3, lane 7), but not 10 M GGOH (Fig. 3, lane 8). Lovastatin-induced limitation of geranylgeranylation was prevented by addition of mevalonate (Fig. 3, lane 6) or GGOH (Fig. 3, lane 8), but not FOH (Fig. 3, lane 7). Mevalonate alone did not affect prenylation (data not shown). Concentrations of 120 M zoledronate limited farnesylation and geranylgeranylation (Fig. 3, lane 10), and the effect on each was prevented by addition of Fig. 3. Inhibition of geranylgeranylation by DGBP is prevented by addition of exogenous GGOH. Cells were seeded and treated with DGBP or other inhibitors of isoprenoid synthesis alone or with selective add backs of the reaction product(s). Inhibition of protein prenylation is shown for K562 cells treated with DGBP or other inhibitors of isoprenoid synthesis for 48 h as determined by Western blot analysis. One representative Western blot is shown (n 3).

5 1032 Dudakovic et al. 10 M FOH and 10 M GGOH, respectively (Fig. 3, lanes 11 and 12). Treatment with 40 M DGBP did not affect farnesylation, but it inhibited geranylgeranylation of both the GGTase I substrate Rap1a and GGTase II substrate Rab6 (Fig. 3, lane 14), an effect that was completely reversed by addition of 10 M GGOH (Fig. 3, lane 15). Unlike HMG-CoA Reductase Inhibition, Apoptosis Induced by Geranylgeranyl Diphosphate Synthase Inhibition Is Not Fully Reversed by Product Supplementation. To determine whether product supplementation could prevent the apoptosis induced by inhibition of GGDPS as indicated by PARP and caspase-3 cleavage, cells were treated with 100 M DGBP for 72 h (Fig. 4A). As controls, cells were treated with either lovastatin, zoledronate, or GGTI-298, a peptidomimetic inhibitor of geranylgeranyl transferase I (Lerner et al., 1995). We were surprised to find that the addition of 10 M exogenous GGOH did not prevent the ability of 100 M DGBP to induce apoptotic effects (Fig. 4A, lanes 5 and 9). In fact, addition of 10 M GGOH to 100 M DGBP enhanced the cleavage of both PARP and caspase-3, whereas GGOH had no effects on PARP and caspase-3 on its own (Fig. 4A, lane 7). This contrasts to 50 M lovastatin-induced PARP and caspase-3 cleavage, which was completely prevented by the addition of 1 mm mevalonate (Fig. 4A, lanes 3 and 8). Treatment with 250 M zoledronate (Fig. 4A, lane 4) displayed minor amounts of cleavage, whereas 20 M GGTI-298 did not induce any PARP and caspase-3 cleavage (Fig. 4A, lane 6). We performed a TUNEL assay to detect DNA fragmentation that results from the apoptotic signaling cascade, using the same conditions (Fig. 4B). As with PARP and caspase-3 cleavage, 50 M lovastatin induced DNA fragmentation (Fig. 4B, column 3) that was prevented by the addition of 1 mm mevalonate (Fig. 4B, column 8), whereas 10 M GGOH (Fig. 4B, column 9) was not able to prevent DNA fragmentation resulting from 100 M DGBP (Fig. 4B, column 5). GGOH (10 M) (Fig. 4B, column 7) and 20 M GGTI (Fig. 4B, column 6) did not lead to DNA fragmentation. Zoledronate (250 M) (Fig. 4B, column 4) lead to DNA fragmentation but not as well as DGBP or lovastatin. Mevalonate alone had no effect on cleavage of PARP or caspase-3 and DNA fragmentation (data not shown). Inhibition of Geranylgeranyl Diphosphate Synthase Leads to Apoptosis through Two Different Mechanisms. To further explore the effects seen by combinations of DGBP and GGOH, cells were treated with low and high concentrations of DGBP, and levels of intracellular FPP and GGPP were measured (Fig. 5). Treatment with DGBP concentration-dependently decreased GGPP, whereas it increased FPP. GGOH partially reversed both the increase in FPP and decrease in GGPP. Previous studies have shown that farnesol and geranylgeraniol induce apoptosis (Edwards and Ericsson, 1999). Isoprenoid alcohols, but not diphosphates, inhibit choline phosphotransferase leading to depletion of phosphatidylcholine (PC) (Voziyan et al., 1993; Miquel et al., 1998). Cells were treated with low (25 M) or high (100 M) DGBP in the Fig. 4. Apoptosis induced by high concentrations of DGBP is not prevented by exogenous GGOH. K562 cells were seeded and treated with DGBP or other inhibitors of isoprenoid synthesis alone or with selective add backs of the reaction product for 72 h. A, induction of PARP and caspase-3 cleavage as determined by Western blot analysis. One representative Western blot is shown (n 3). B, DNA fragmentation as determined by TUNEL assay (mean S.E.; n 2;, p 0.05 relative to control). Fig. 5. Inhibition of GGDPS depletes cells of GGPP and leads to accumulation of FPP. K562 cells were treated for 48 h with varying concentrations of DGBP in the presence or absence of exogenous GGOH as described under Materials and Methods. Total amounts of intracellular FPP and GGPP were measured as described in text. Amounts were expressed relative to the number of cells at the end of the incubation (mean S.E.; n 2;, p 0.05 relative to control levels of each isoprenoid; a, p 0.05 relative to respective DGBP treatment).

6 presence or absence of GGOH (10 M) or PC (100 M) (Fig. 6). GGOH (10 M) prevented apoptotic effects caused by low (25 M) DGBP (Fig. 6, A and B). Although 10 M GGOH completely prevented staining, 100 M PC was not able to significantly impair, annexin V and PI staining (Fig. 6A, columns 4 6) and PARP cleavage (Fig. 6B, lanes 4 6) caused by 25 M DGBP. However, PC was able to partially impair apoptotic effects caused by high (100 M) DGBP (Fig. 6, A and C). Addition of 100 M PC prevented some annexin V and PI staining (Fig. 6A, columns 7 and 9) and most PARP cleavage (Fig. 6C, lanes 4 and 6) caused by 100 M DGBP. In contrast, the addition of 10 M GGOH to 100 M DGBP lead to enhanced annexin V and PI staining (Fig. 6A, columns 7 Fig. 6. The mechanism of DGBP-induced apoptosis is concentrationdependent. K562 cells were treated with different concentrations of DGBP in the presence or absence of exogenous GGOH or PC for 72 h. A, total apoptosis as determined by annexin V and PI staining (mean S.E.; n 4;, p 0.05 relative to control; a, p 0.58 relative to 50 M etoposide, b, p 0.05 relative to 25 M DGBP, c, p 0.37 relative to 25 M DGBP, and d, p 0.05 relative to 100 M DGBP). Total apoptosis is expressed as the sum of early apoptosis (annexin V /PI ) and late apoptosis/necrosis (annexin V /PI ) fractions. B, Western blot analysis in cells treated with low (25 M) DGBP concentration (n 3). C, Western blot analysis in cells treated with high (100 M) DGBP concentration (n 3). and 8) and PARP cleavage (Fig. 6C, lanes 4 and 5). Etoposideinduced apoptotic effects were not significantly altered by the addition of PC (Fig. 6, A C, lanes/columns 2 and 3). Annexin V and PI staining and PARP cleavage were not affected by 10 M GGOH or 100 M PC (data not shown). Inhibition of Geranylgeranyl Diphosphate Synthase Alters Steady-State mrna Levels of Isoprenoid Biosynthetic Enzymes. Because some of the effects of DGBP seemed to result from accumulation of FPP, whose downstream sterol products are known to regulate the isoprenoid biosynthetic machinery through transcriptional events (Brown and Goldstein, 1997), we hypothesized treatment with DGBP would result in decreased mrna expression of HMGCR, FDPS, and SQS. Cells were treated with DGBP, and then they were tested for altered expression of enzymes involved in isoprenoid biosynthesis via quantitative real-time PCR (Table 1). As characterized previously, treatment with either lovastatin or zaragozic acid resulted in increased mrna expression of HMGCR, FDPS, and SQS. Supporting our hypothesis, treatment with DGBP resulted in decreased expression of HMGCR, FDPS, and SQS mrna. Digeranyl Bisphosphonate-Induced Inhibition of Geranylgeranyl Diphosphate Synthase Is Synergistic with Lovastatin and Zoledronate. To further examine the consequences of GGDPS inhibition in the context of other protein geranylgeranylation inhibitors, cells were tested for synergistic interactions between DGBP and lovastatin, zoledronate, or GGTI-298. For these studies, concentration-response curves for each compound were constructed using [ 3 H]thymidine incorporation at the 48-h time point. All drugs were tested alone or in combination with constant ratios of DGBP according to the method of Chou and Talalay (1984). The data were analyzed in two ways. First, isobologram analyses were conducted (Fig. 7). A general isobologram with a line of additivity and regions of synergy and antagonism is shown (Fig. 7A). Lovastatin and DGBP were strongly synergistic (CI ) for inhibition of K562 cell growth (Fig. 7B). Zoledronate was also synergistic with DGBP (CI ) (Fig. 7C), but not to the extent of the synergy with lovastatin. GGTI-298 was antagonistic to DGBP (CI ) (Fig. 7D). Although isobologram analysis can determine whether a combination of compounds is synergistic and approximate the magnitude of the synergistic interaction, it does not quantify this at the experimental concentrations per se. Therefore, the CI values for each experimental combination of DGBP with lovastatin or zoledronate were calculated (Table 2). At all of the experimental concentrations tested, lovastatin and DGBP displayed stronger synergy than the combination of zoledronate and lovastatin (CI ). This result has been observed in RPMI-8226 and U937 cells as well (data not shown). To further support the synergy observed in [ 3 H]thymidine TABLE 1 Differential effects of lovastatin, zaragozic acid, and DGBP on expression of HMGCR, FDPS, and SQS mrna in K562 cells after 48 h determined by real-time PCR Data are mean S.E. (n 4). 20 M Lovastatin GGDPS Apoptosis and Synergy M Zaragozic Acid 10 M DGBP HMGCR FDPS SQS * p 0.05 relative to control for all samples except * p 0.16.

7 1034 Dudakovic et al. M DGBP (Fig. 8, lanes 4, 8, and 12). It is noteworthy that the combination of lovastatin and DGBP had an antagonistic effect on inhibition of Ras farnesylation. Although lovastatin caused the accumulation of an unmodified upper Ras band (Fig. 8, lanes 4 and 5), DGBP did not (Fig. 8, lanes 8 and 9). When combined, DGBP prevented the inhibition of Ras farnesylation induced by lovastatin (Fig. 8, lanes 12 and 13). Quantification of the Rap1a bands and Rab6 upper bands further supports the synergistic interaction of lovastatin and DGBP toward geranylgeranylation inhibition. Fig. 7. Inhibition of GGDPS is synergetic with inhibition of HMGCR and FDPS, but it is antagonistic to inhibition of GGTase. K562 cells were treated with varying concentrations of inhibitors alone or in indicated combinations at constant ratios for 48 h. Cell growth was assessed by [ 3 H]thymidine incorporation. Isobologram analysis was conducted as described. Isobolograms shown were generated from cells treated with five different concentrations of each drug or combination (n 4). A, idealized isobologram analysis displays line of additivity, and regions of synergy and antagonism. B, experimental isobologram analysis of DGBP in combination with lovastatin. C, isobologram of DGBP and zoledronate. D, isobologram of DGBP and GGTI-298. incorporation assays, CI analysis was performed after annexin V and PI staining on cells treated with combinations of lovastatin and DGBP (Table 3). Similar to the [ 3 H]thymidine incorporation assays, the annexin V and PI staining assays revealed a synergistic interaction toward total cell death (CI ). More specifically, these assays also demonstrated a synergistic interaction occurs both for cells in early (CI ) and late apoptosis (CI ). Digeranyl Bisphosphonate Potentiates Lovastatin- Induced Inhibition of Protein Geranylgeranylation. Because both lovastatin and DGBP inhibit protein geranylgeranylation through different but overlapping mechanisms, combinations of the two compounds were tested for the ability to inhibit protein geranylgeranylation (Fig. 8). Cells were treated with lovastatin and DGBP alone or in combination. To observe synergistic effects, it was necessary to use lower concentrations of lovastatin and DGBP. Both lovastatin and DGBP exhibited submaximal inhibition of geranylgeranylation at concentrations below 10 and 20 M, respectively. Concentrations of 1.1 M lovastatin and 2.2 M DGBP had minimal effects, whereas the combination of these concentrations was sufficient to impair geranylgeranylation (Fig. 8, lanes 3, 7, and 11). This effect was also pronounced at concentrations of 3.3 M lovastatin and 6.6 Discussion We have now shown for the first time that inhibition of GGDPS leads to growth inhibition and induces apoptosis. This demonstrates that like many of its counterparts in the mevalonate pathway, including HMG-CoA reductase (Pérez- Sala and Mollinedo, 1994), FDPS (Shipman et al., 1997), and protein prenyltransferases (Lebowitz et al., 1997), inhibition of GGDPS may be a viable therapeutic strategy. Inhibition of GGDPS has two consequences, depletion of cellular GGPP and an accumulation of FPP. At any concentration of DGBP, the inhibition of protein geranylgeranylation can be prevented by addition of exogenous GGOH. At low amounts of GGDPS inhibition, apoptotic consequences are able to be prevented by exogenous GGOH. This agrees with previous studies that showed GGOH reverses apoptosis induced by the clinical NBPs (van Beek et al., 1999a). At greater GGDPS inhibition, apoptotic effects are not prevented by exogenous GGOH, rather they are enhanced. It is possible that increased cellular FPP (and FOH) levels caused by GGDPS inhibition in the presence of exogenous GGOH are high enough to inhibit cellular cholinephosphotransferase (Miquel et al., 1998). Our data support this conclusion because apoptosis induced by high concentrations of DGBP can be partially prevented by addition of exogenous PC. It is possible that PC cannot fully prevent apoptosis because it does not prevent the inhibition of protein geranylgeranylation caused by GGDPS inhibition. In addition to inhibition of PC synthesis, there are at least three other well characterized consequences of FPP or FOH accumulation that are relevant to regulation of isoprenoid biosynthesis. First, FPP is converted to squalene and eventually cholesterol, which when elevated decreases transcription of the enzymes responsible for its synthesis (Brown and Goldstein, 1997). Our results clearly show decreased mrna levels of three of these enzymes, HMGCR, FDPS, and SQS associated with FPP accumulation caused by GGDPS inhibition. The fact that these changes were seen in media containing serum is consistent with the requirement of endogenous TABLE 2 Experimental combination index values for K562 cells treated with DGBP in combination with various inhibitors of isoprenoid biosynthesis for 48 h as determined by 3 H thymidine incorporation (n 4) Inhibitor data are given in micromolar concentration. CI for Experimental Values Lov DGBP CI DGBP GGTI CI DGBP Zol CI Lov Zol CI Zol, zoledronate.

8 TABLE 3 Experimental combination index values for K562 cells treated with indicated combinations of DGBP and lovastatin for 72 h as determined by annexin V and PI staining (n 2) Lov and DGBP values are micromolar concentration. CI for Experimental Values Lov DGBP Early Apoptosis Late Apoptosis Total Cell Death Fig. 8. Combinations of DGBP and lovastatin synergistically inhibit protein geranylgeranylation. K562 cells were seeded and treated with DGBP or lovastatin at various submaximal concentrations alone or in combination for 48 h. Inhibition of protein prenylation is shown as determined by Western blot analysis (n 2). Ras is exclusively farnesylated in these cells, Rap1a is geranylgeranylated by GGTase I, and Rab6 is geranylgeranylated by GGTase II (see text for full explanation). Unmodified bands were quantified using ImageJ as described under Materials and Methods. Rap1a and Rab6 inhibition values were expressed as a percentage of the maximal obtained inhibition (% max). sterol synthesis to regulate gene expression. However, further studies of GGDPS in serum-free media may reveal more dramatic changes. Second, FPP and FOH have been shown conclusively to increase HMGCR degradation in pulse-chase experiments (Meigs and Simoni, 1997). Third, both FPP and FOH have been shown to inhibit the mevalonate kinase reaction (Dorsey and Porter, 1968). We have also shown for that first time that inhibition of GGDPS in combination with inhibition of either HMGCR or FDPS results in synergistic inhibition of cell growth and induction of apoptosis. It is possible that both the FPP accumulation and GGPP depletion caused by GGDPS inhibition serve to enhance lovastatin-induced apoptosis. In contrast, lovastatin and zoledronate both similarly deplete FPP and GGPP and lack the apoptotic effects caused by FPP accumulation. This line of reasoning also explains why the combination of zoledronate and DGBP is more synergistic than the combination of zoledronate and lovastatin. The data presented herein suggest that inhibition of GGDPS may be a good therapeutic strategy either alone on in combination with other isoprenoid pathway inhibitors. DGBP may be useful alone or in combination with statins for inhibition of cancerous cell growth or other cellular processes that are dependent upon geranylgeranylation. Ongoing studies will further evaluate effects of GGDPS inhibition on cancer and bone disorders. GGDPS Apoptosis and Synergy 1035 Acknowledgments We thank Larry W. Shull and Dr. David F. Wiemer for synthesis of digeranyl bisphosphonate, Dr. Huaxiang Tong for assistance with FPP and GGPP measurements, Justin Fishbaugh and the University of Iowa Flow Cytometry Facility for assistance with FACScan analysis, and the University of Iowa DNA Facility for assistance with PCR. References Andersen JM and Dietschy JM (1978) Relative importance of high and low-density lipoproteins in the regulation of cholesterol-synthesis in the adrenal-gland, ovary, and testis of the rat. J Biol Chem 253: Brown MS and Goldstein JL (1980) Multivalent feedback regulation of HMG CoA reductase, a control mechanism coordinating isoprenoid synthesis and cell growth. J Lipid Res 21: Brown MS and Goldstein JL (1997) The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell 89: Burridge K and Doughman R (2006) Front and back by Rho and Rac. Nat Cell Biol 8: Caraglia M, D Alessandro AM, Marra M, Giuberti G, Vitale G, Viscomi C, Colao A, Del Prete S, Tagliaferra P, Tassone P, et al. (2004) The farnesyl transferase inhibitor R (Zarnestra) synergistically enhances growth inhibition and apoptosis induced on epidermoid cancer cells by zoledronic acid (Zometa) and pamidronate. Oncogene 23: Chou TC and Talalay P (1984) Quantitative analysis of dose-effect relationships: the combined effects of multiple drugs or enzyme inhibitors. Adv Enzyme Regul 22: Crick DC, Andres DA, and Waechter CJ (1997) Novel salvage pathway utilizing farnesol and geranylgeraniol for protein isoprenylation. Biochem Biophys Res Commun 237: Dorsey JK and Porter JW (1968) The inhibition of mevalonic kinase by geranyl and farnesyl pyrophosphates. J Biol Chem 243: Dunford JE, Thompson K, Coxon FP, Luckman SP, Hahn FM, Poulter CD, Ebetino FH, and Rogers MJ (2001) Structure-activity relationships for inhibition of farnesyl diphosphate synthase in vitro and inhibition of bone resorption in vivo by nitrogen-containing bisphosphonates. J Pharmacol Exp Ther 296: Edwards PA and Ericsson J (1999) Sterols and isoprenoids: signaling molecules derived from the cholesterol biosynthetic pathway. Annu Rev Biochem 68: Fisher JE, Rogers MJ, Halasy JM, Luckman SP, Hughes DE, Masarachia PJ, Wesolowski G, Russell RG, Rodan GA, and Reszka AA (1999) Alendronate mechanism of action: geranylgeraniol, an intermediate in the mevalonate pathway, prevents inhibition of osteoclast formation, bone resorption, and kinase activation in vitro. Proc Natl Acad Sci USA96: Goldstein JL and Brown MS (1990) Regulation of the mevalonate pathway. Nature 343: Green JR and Clezardin P (2002) Mechanisms of bisphosphonate effects on osteoclasts, tumor cell growth, and metastasis. Am J Clin Oncol 25:S3 S9. Grundy SM (1988) HMG-CoA reductase inhibitors for treatment of hypercholesterolemia. N Engl J Med 319: Holstein SA and Hohl RJ (2001a) Interaction of cytosine arabinoside and lovastatin in human leukemia cells. Leuk Res 25: Holstein SA and Hohl RJ (2001b) Synergistic interaction of lovastatin and paclitaxel in human cancer cells. Mol Cancer Ther 1: Jagdev SP, Coleman RE, Shipman CM, Rostami HA, and Croucher PI (2001) The bisphosphonate, zoledronic acid, induces apoptosis of breast cancer cells: evidence for synergy with paclitaxel. Br J Cancer 84: Kuroda J, Kimura S, Segawa H, Kobayashi Y, Yoshikawa T, Urasaki Y, Ueda T, Enjo F, Tokuda H, Ottmann OG, et al. (2003) The third-generation bisphosphonate zoledronate synergistically augments the anti-ph leukemia activity of imatinib mesylate. Blood 102: Lebowitz PF, Sakamuro D, and Prendergast GC (1997) Farnesyl transferase inhibitors induce apoptosis of Ras-transformed cells denied substratum attachment. Cancer Res 57: Lerner EC, Qian Y, Hamilton AD, and Sebti SM (1995) Disruption of oncogenic K-Ras4B processing and signaling by a potent geranylgeranyltransferase I inhibitor. J Biol Chem 270: Lewis KA, Holstein SA, and Hohl RJ (2005) Lovastatin alters the isoprenoid biosynthetic pathway in acute myelogenous leukemia cells in vivo. Leuk Res 29: Licata AA (2005) Discovery, clinical development, and therapeutic uses of bisphosphonates. Ann Pharmacother 39: Maalouf MA, Wiemer AJ, Kuder CH, Hohl RJ, and Wiemer DF (2007) Synthesis of fluorescently tagged isoprenoid bisphosphonates that inhibit protein geranylgeranylation. Bioorg Med Chem 15: Meigs TE and Simoni RD (1997) Farnesol as a regulator of HMG-CoA reductase degradation: characterization and role of farnesyl pyrophosphatase. Arch Biochem Biophys 345:1 9. Miquel K, Pradines A, Terce F, Selmi S, and Favre G (1998) Competitive inhibition of choline phosphotransferase by geranylgeraniol and farnesol inhibits phosphatidylcholine synthesis and induces apoptosis in human lung adenocarcinoma A549 cells. J Biol Chem 273: Pérez-Sala D and Mollinedo F (1994) Inhibition of isoprenoid biosynthesis induces apoptosis in human promyelocytic HL-60 cells. Biochem Biophys Res Commun 199: Reszka AA and Rodan GA (2004) Nitrogen-containing bisphosphonate mechanism of action. Mini Rev Med Chem 4: Santini D, Caraglia M, Vincenzi B, Holen I, Scarpa S, Budillon A, and Tonini G (2006) Mechanisms of disease: preclinical reports of antineoplastic synergistic action of bisphosphonates. Nat Clin Pract Oncol 3:

9 1036 Dudakovic et al. Schmidmaier R, Simsek M, Baumann P, Emmerich B, and Meinhardt G (2006) Synergistic antimyeloma effects of zoledronate and simvastatin. Anticancer Drugs 17: Schreiber V, Dantzer F, Ame JC, and de Murcia G (2006) Poly(ADP-ribose): novel functions for an old molecule. Nat Rev Mol Cell Biol 7: Shipman CM, Rogers MJ, Apperley JF, Graham R, Russell G, and Croucher PI (1997) Bisphosphonates induce apoptosis in human myeloma cell lines: a novel anti-tumour activity. Br J Haematol 98: Shull LW, Wiemer AJ, Hohl RJ, and Wiemer DF (2006) Synthesis and biological activity of isoprenoid bisphosphonates. Bioorg Med Chem 14: Swanson KM and Hohl RJ (2006) Anti-cancer therapy: targeting the mevalonate pathway. Curr Cancer Drug Targets 6: Takai Y, Sasaki T, and Matozaki T (2001) Small GTP-binding proteins. Physiol Rev 81: Thornberry NA and Lazebnik Y (1998) Caspases: enemies within. Science 281: Tong H, Holstein SA, and Hohl RJ (2005) Simultaneous determination of farnesyl and geranylgeranyl pyrophosphate levels in cultured cells. Anal Biochem 336: van Beek E, Lowik C, van der Pluijm G, and Papapoulos S (1999a) The role of geranylgeranylation in bone resorption and its suppression by bisphosphonates in fetal bone explants in vitro: a clue to the mechanism of action of nitrogencontaining bisphosphonates. J Bone Miner Res 14: van Beek E, Pieterman E, Cohen L, Lowik C, and Papapoulos S (1999b) Nitrogencontaining bisphosphonates inhibit isopentenyl pyrophosphate isomerase/farnesyl pyrophosphate synthase activity with relative potencies corresponding to their antiresorptive potencies in vitro and in vivo. Biochem Biophys Res Commun 255: Voziyan PA, Goldner CM, and Melnykovych G (1993) Farnesol inhibits phosphatidylcholine biosynthesis in cultured-cells by decreasing cholinephosphotransferase activity. Biochem J 295: Wiemer AJ, Tong HX, Swanson KM, and Hohl RJ (2007) Digeranyl bisphosphonate inhibits geranylgeranyl pyrophosphate synthase. Biochem Biophys Res Commun 353: Address correspondence to: Dr. Raymond J. Hohl, University of Iowa, SE 313 GH, Iowa City, IA raymond-hohl@uiowa.edu