Lovastatin augments sulindac-induced apoptosis in colon cancer cells and potentiates chemopreventive effects of sulindac

GASTROENTEROLOGY 1999;117:838–847

Lovastatin Augments Sulindac-Induced Apoptosis in Colon Cancer Cells and Potentiates Chemopreventive Effects of Sulindac BANKE AGARWAL,* CHINTHALAPALLY V. RAO,‡ SANJAY BHENDWAL,* WILLIAM R. RAMEY,§ HAIM SHIRIN,* BANDARU S. REDDY,‡ and PETER R. HOLT* *Division of Gastroenterology, Department of Medicine, and §Department of Surgery, St. Luke’s–Roosevelt Hospital Center, and Departments of Medicine and Surgery, College of Physicians and Surgeons, Columbia University, New York, New York; and ‡Chemoprevention Program, American Health Foundation, Valhalla, New York

Background & Aims: 3-Hydroxy-3-methylglutaryl–coenzyme A (HMG-CoA) reductase inhibitors (HRIs) were found incidentally to reduce new cases of colon cancer in 2 large clinical trials evaluating coronary events, although most patients in both treatment and control group were taking nonsteroidal anti-inflammatory drugs (NSAIDs). NSAIDs are associated with reduced colon cancer incidence, predominantly by increasing apoptosis. We showed previously that lovastatin induces apoptosis in colon cancer cells. In the present study we evaluated the potential of combining lovastatin with sulindac for colon cancer chemoprevention. Results: Lovastatin, 10–30 ␮mol/L, augmented sulindac-induced apoptosis up to 5-fold in 3 colon cancer cell lines. This was prevented by mevalonate (100 ␮mol/L) or geranylgeranylpyrophosphate (10 ␮mol/L) but not farnesylpyrophosphate (100 ␮mol/L), suggesting inhibition of geranylgeranylation of target protein(s) as the predominant mechanism. In an azoxymethane rat model of chemical-induced carcinogenesis, the total number of colonic aberrant crypt foci per animal (control, 161 ⴞ 11) and the number of foci with 4ⴙ crypts (control, 40 ⴞ 4.5) decreased to 142 ⴞ 14 (NS) and 43 ⴞ 2.9 (NS), respectively, with 50 ppm lovastatin alone; to 137 ⴞ 5.4 (P ⴝ 0.053) and 36 ⴞ 2.1 (NS) with 80 ppm sulindac alone; and to 116 ⴞ 8.1 (P ⴝ 0.004) and 28 ⴞ 3.4 (P ⴝ 0.02) when 50 ppm lovastatin and 80 ppm sulindac were combined. Conclusions: Addition of an HRI such as lovastatin may augment chemopreventive effects of NSAIDs or/and may allow lower, less toxic doses of these drugs to be used.

olorectal cancer is the second leading cause of cancer mortality in the United States. Several lines of evidence suggest that nonsteroidal anti-inflammatory drugs (NSAIDs) have cancer chemopreventive properties. In cultured tumor cells, NSAIDs have been shown to inhibit cell proliferation, cause cell cycle arrest, and induce apoptosis.1,2 NSAIDs inhibited tumor formation

C

in rodent models with carcinogen-induced cancers.3–7 The NSAID sulindac caused regression of and prevented recurrences of colorectal adenomas, which are precursor lesions of large bowel cancer in patients with familial adenomatous polyposis (FAP),8,9 as well as in MIN mice, animal model of FAP.10 Epidemiological studies also suggest that aspirin has a protective effect against colorectal cancer incidence and mortality.11 The biochemical basis for these properties of NSAIDs have been attributed to a reduction of prostaglandin levels through the inhibition of 2 isoforms of cyclooxygenase (COX),12–14 inducible COX-2 or a constitutive COX-1, enzymes that catalyze the formation of prostaglandins from arachidonic acid. Products of arachidonic acid metabolism, including prostaglandins, may influence tumor progression, and their levels are increased in adenomas and carcinomas.15,16 When aspirin inhibits colon tumor formation in experimental animals, prostaglandin E2 levels are reduced in colonic mucosa.12–14 However, not all studies of NSAID-induced inhibition of colon carcinogenesis show that prostaglandins are involved.17–20 3-Hydroxy-3-methylglutaryl–coenzyme A (HMGCoA) reductase inhibitors (HRIs) are used extensively to reduce serum cholesterol in patients with coronary artery disease. HRIs prevent formation of mevalonate from HMG-CoA by inhibiting the enzyme HMG-CoA reductase and thereby inhibit cholesterol synthesis. Two downstream intermediates in the cholesterol synthetic pathAbbreviations used in this paper: ACF, aberrant crypt foci; APC, adenomatous polyposis coli; BrdU, bromodeoxyuridine; COX, cyclooxygenase; FAP, familial adenomatous polyposis; FTI, farnesyl protein transferase inhibitor; HMG-CoA; 3-hydroxy-3-methylglutaryl–coenzyme A; HRI, HMG-CoA reductase inhibitor; MTT, 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenol tetrazolium bromide; TUNEL, terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nickend labeling. r 1999 by the American Gastroenterological Association 0016-5085/99/$10.00

October 1999

way, farnesylpyrophosphate and geranylgeranylpyrophosphate, belong to a class of compounds named isoprenoids and are bound to several cellular proteins by a posttranslational modification named isoprenylation. Isoprenylation of these cellular proteins, which include low-molecular-weight G proteins such as ras, rho, and rac, is crucial for their membrane attachment and function.21–24 By blocking the formation of isoprenoids and thereby inhibiting protein isoprenylation, HRIs exert important cellular effects, which include reducing proliferation and inducing apoptosis.25–29 In 2 large clinical trials with HRIs designed to study changes in coronary outcomes in patients with coronary artery disease, there was a 43%30 and 19%31 reduction in the number of newly diagnosed cases of colon cancer diagnosed during a 5-year follow-up period in patients receiving pravastatin and simvastatin, respectively. Narisawa et al.32,33 have shown that pravastatin and simvastatin reduce carcinogen-induced colon cancers in rodent models by 50%–65%. We showed previously that lovastatin induces apoptosis and inhibits proliferation in colon cancer cells,34 thus providing a rationale for studying HRIs for possible chemopreventive effects, particularly in colon cancer. Furthermore, 83% of patients in both the pravastatin arm and the placebo arm of the CARE (Cholesterol And Recurrent Events) trial30 were receiving daily aspirin (for coronary artery disease prophylaxis). Despite aspirin ingestion, patients taking pravastatin during the study still showed a major reduction in the number of new cases of colon cancers. This suggested a possible additive effect of HRIs with NSAIDs in colon cancer chemoprevention. This study explores the possible augmentation of the chemopreventive effect of NSAIDs by HRIs. Because the chemopreventive effect of NSAIDs is currently believed to be a result of increased apoptosis, we investigated whether HRIs could also increase apoptosis induced by NSAIDs in colon cancer cells. In the first part of the study, we show in a cell culture system that an HRI, lovastatin, augments sulindac-induced apoptosis in several colon cancer cell lines. In the second part of the study, we found that the combination of sulindac and lovastatin, used in low doses that individually do not have a significant chemopreventive effect on colon carcinogenesis, produced a 30% reduction in the appearance of colonic aberrant crypt foci (ACF), an accepted biomarker of risk for the subsequent development of carcinogeninduced colon cancer formation in rodent models.35,36 To reduce the incidence of colon cancer, combinations of chemopreventive agents are increasingly being studied in an attempt to amplify preventive effects or/and to reduce side effects of known effective agents. Our data suggest

LOVASTATIN SENSITIZES COLON CANCER CELLS TO APOPTOSIS 839

that addition of HRIs such as lovastatin may greatly increase the chemopreventive effects of sulindac or may allow for effective action by the NSAID at a lower dose.

Materials and Methods In Vitro Studies Cell culture. HCT-116, SW-480, and LoVo cells were obtained from American Type Culture Collection (Rockville, MD) and maintained in Dulbecco’s modified Eagle medium or Ham solution (GIBCO BRL, Grand Island, NY) with 10% fetal bovine serum in an atmosphere of 95% O2, 5% CO2 at 37°C without antibiotics. All studies were performed with cells at 50% confluence. Cells were treated with lovastatin at concentrations from 5 to 30 µmol/L for 48 hours. Because of different sensitivities of the cell lines to lovastatin-induced apoptosis, the following concentrations of lovastatin were used in experiments with the cell lines studied: HCT-116, 0, 10, and 30 µmol/L; SW480, 0, 5, and 10 µmol/L; and LoVo, 0, 10, and 20 µmol/L. The medium then was changed, the same doses of lovastatin were added, and cells were treated with 100, 250, or 500 µmol/L sulindac (Sigma Chemical Co., St. Louis, MO) for 48 hours before harvesting for analysis. For add-back experiments, 100–500 µmol/L mevalonate, 100–250 µmol/L farnesylpyrophosphate, or 10 µmol/L geranylgeranylpyrophosphate (Sigma) was added with lovastatin at the beginning of the experiment and again at 48 hours when sulindac was added. Flow cytometry. Flow cytometry was used to quantitate apoptotic cells by 2 methods, DNA histogram and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL) staining, performed simultaneously on the same samples. TUNEL staining was used to confirm the results observed by measurement of subdiploid cells. TUNEL staining detects cells in an earlier phase of apoptosis than by measurement of subdiploid cells, so that more cells are measured as apoptotic. Cells were analyzed on a FACSort flow cytometer (Becton Dickinson, San Jose, CA) after staining using a commercially available apo-BrdU kit (Phoenix Flow Systems, San Diego, CA). TUNEL staining. DNA strand breaks in apoptotic cells were detected by incorporation of fluorescein-labeled deoxyuridine triphosphate into fragmented DNA by terminal deoxynucleotidyl transferase using the apo-BrdU kit.37 The cells were collected and stained according to the protocol provided by the manufacturer. The data were plotted on a dot plot FL2-A vs. FL2-W, and a singlet gate was applied. These gated cells then were plotted on dot plot FL1-H(log) vs. FL2-A(lin), and cells stained with bromodeoxyuridine (BrdU) were counted as apoptotic. The data were also plotted on FL2-H histograms, and the number of sub-G1 cells was counted as apoptotic cells. All flow-cytometric studies were performed in triplicate and repeated 3 times. The data are presented as mean ⫾ SD of 3 readings from each experiment. Similar data were obtained when the experiments were repeated. The Student t test was used to calculate the statistical significance between the controls (no lovastatin) and the 2 dose levels of lovastatin used.

840 AGARWAL ET AL.

P value of ⬍0.05 was considered significant. All differences were statistically significant except those marked with an asterisk. MTT assay. Cells were grown in 96-well plates and treated with lovastatin for 48 hours. The medium then was changed, and lovastatin and sulindac were added. After an additional 48-hour incubation, 50 µg of 3-[4,5-dimethylthiazol2-yl]-2,5-diphenol tetrazolium bromide (MTT; Sigma) was added to each well (the floating cells were not removed), and the plates were incubated at 37°C for 2 hours. Next, 100 µL of MTT solubilization solution (10% Triton X-100 plus 0.1N HCl in anhydrous isopropanol) was then added, and the plates were agitated on a mechanical shaker to dissolve the crystals. Absorbance was measured spectrophotometrically at a dual wavelength of 570 and 405 nm; the mean of 6 readings was used for calculations. The data are presented as the absorbance of treated cells as a percentage of the absorbance of untreated samples. The data are the means ⫾ SD of 8 readings from each experiment. The Students t test was used to calculate statistical significance between the controls (no lovastatin) and the 2 dose levels of lovastatin used. P values of ⬍0.001 were considered significant. Expression of isoforms of COX. Exponentially growing cells were collected by scraping, washed 3 times in ice-cold phosphate-buffered saline, and resuspended in lysis buffer containing 20 mmol/L Tris-HCl (pH 7.4), 2 mmol/L EDTA, 2 mmol/L EGTA, 6 mmol/L mercaptoethanol, 1% Nonidet P-40, 0.1% sodium dodecyl sulfate (SDS), and 10 mmol/L NaF, plus the protease inhibitors leupeptin (10 µg/mL), aprotinin (10 µg/mL), and phenylmethylsulfonyl fluoride (0.1 mol/L; all from Sigma). After lysis with sonification, the resulting insoluble material was removed by centrifugation at 15,000 rpm for 15 minutes at 4°C, and the supernatant was stored at ⫺80°C. Protein concentrations were measured by the Bradford method, and 50-µg samples were mixed with 2⫻ Laemmli buffer, boiled for 5 minutes, electrophoresed in 10% SDS– polyacrylamide gel electrophoresis, and transferred to Immobilon membranes (Millipore, Bedford, MA). Western blot analyses were then performed as described previously38 using specific polyclonal antibodies to COX-2 (Oxford Biomedical Research, Oxford, MI) and COX-1 (Santa Cruz Biotechnologies, Santa Cruz, CA). Detection of antibody binding was by enhanced chemiluminescence (Amersham, Arlington Heights, IL).

In Vivo Studies Animals, diet, and carcinogen. Azoxymethane (CAS 25843-45-2) was purchased from Ash Stevens (Detroit, MI). Sulindac was provided by the National Cancer Institute (Bethesda, MD). Lovastatin was provided by Merck Laboratories (Rahway, NJ). Weaning male F344 rats were purchased from Charles River Breeding Laboratories (Kingston, NY). All ingredients of the semipurified diet were obtained from Dyets Inc. (Bethlehem, PA) and stored at 4°C until preparation of the experimental diets. Male F344 rats received at weaning were held in quarantine for 1 week. They had free access to modified AIN-76A semipurified control diet.35 The animals were ran-

GASTROENTEROLOGY Vol. 117, No. 4

domly distributed by weight into the various dietary groups and were transferred to the animal holding room. They were housed in plastic cages, 3 rats per cage, under controlled conditions, a 12-hour light/12-hour dark cycle, 50% relative humidity, and 21°C room temperature. Experimental diets were prepared by mixing lovastatin and sulindac, and their combination with modified AIN-76A control diet. The study was approved by American Health Foundation animal care committee, and the protocol was reviewed and approved by the Institutional Animal Care and Use Committee. Experimental procedure. At 5 weeks of age, groups of animals were fed the modified AIN-76A (control) and experimental diets containing 50 ppm lovastatin, 80 ppm sulindac, or 50 ppm lovastatin plus 80 ppm sulindac. At 7 weeks of age, all animals except the vehicle-treated rats received subcutaneous injections of azoxymethane once weekly for 2 weeks at a dose of 15 mg/kg body wt per week. The animals intended for vehicle control treatment were administered an equal volume of normal saline. The experimental diets were continued until termination of the study at 17 weeks of age. At that time, all animals were killed by CO2 inhalation, and colons were removed, flushed with Krebs–Ringer solution, opened from cecum to anus, and fixed flat between 2 pieces of filter paper in 10% buffered formalin. Segments of the colon (0.5 cm) were removed for histological analysis. Microscopic slides were placed on top of the filter paper to ensure that the tissue remained flat during fixation. After a minimum of 24 hours in buffered formalin, the colons were cut into 2-cm segments and stained with 0.2% methylene blue in Krebs–Ringer solution for 5 minutes. They were then placed, mucosal side up, on a microscope slide and observed under a light microscope. ACF were recorded by standardized procedures routinely used in our laboratory.36,39 Aberrant crypts were distinguished from the surrounding normal crypts by their increased size, significantly increased distance from the laminae to basal surface of cells, and the easily discernible pericryptal zone. The parameters used to assess the aberrant crypts were number per colon and multiplicity. Crypt multiplicity was determined as the number of crypts in each focus and was categorized as containing 1, 2, 3, or 4 or more aberrant crypts per focus. All colons were scored by one observer and checked at random by a second observer.

Statistical Analysis The Student t test was used to calculate the differences between the different treatment groups in in vitro studies. Analysis of variance (ANOVA) test and unpaired t test were used for the statistical analysis for differences in the treatment groups in the in vivo study.

Results In Vitro Studies Lovastatin induces apoptosis in colon cancer cells. Lovastatin in concentrations ranging from 10 to 30 µmol/L induced modest apoptosis in 3 different colon cancer cell lines (Figure 1). Apoptosis was characterized

October 1999

by electron microscopy, showing chromatin condensation and nuclear fragmentation (data not shown). Apoptosis, quantified as the percentage of subdiploid cells, increased from 2.0% to 20.7% in HCT-116 cells, 4.2% to 57.5% in SW480 cells, and 7.5% to 45.3% in LoVo cells after lovastatin treatment (Figure 1). Because the 3 cell lines showed markedly different levels of apoptosis, the concentrations of lovastatin used to demonstrate augmentation of sulindac-induced apoptosis were adjusted to achieve about the same levels of apoptosis with lovastatin alone (10 and 30 µmol/L for HCT-116, 5 and 10 µmol/L for SW480, and 10 and 20 µmol/L for LoVo cells). Lovastatin augments the effect of apoptosis induced by sulindac in colon cancer cells. Cells were

pretreated with lovastatin for 48 hours before exposure to 100, 250, and 500 µmol/L sulindac. There was a dose-dependent augmentation of sulindac-induced apoptosis with lovastatin before treatment in HCT-116, SW480, and LoVo cells. Similar results were found when apoptosis was quantified by measuring the percentage of subdiploid cells (Figure 2A) or by measuring the percentage of TUNEL-positive cells (Figure 2B). The augmentation of sulindac-induced apoptosis was observed at all doses of sulindac. Using the MTT assay to determine cell viability, lovastatin also had an additive effect with sulindac in decreasing the number of viable cells (Figure 3). Incubation with sulindac alone at concentrations up to 500 µmol/L produced only a modest decrease in the percentage of viable cells in SW480 and HCT-116 cells. When the cells were pretreated with lovastatin, there was

LOVASTATIN SENSITIZES COLON CANCER CELLS TO APOPTOSIS 841

a dramatic decrease (P ⬍ 0.0001) in the percentage of viable cells at all concentrations of sulindac (100–500 µmol/L). Effect of intermediates in the cholesterol synthetic pathway on lovastatin-induced apoptosis and cell viability. To explore possible mechanisms for lova-

statin augmentation of apoptosis induced by sulindac, we performed add-back experiments using intermediates in the cholesterol synthetic pathway. In SW480 and HCT116 cells, mevalonate at concentrations of 100–500 µmol/L, farnesylpyrophosphate at 100–250 µmol/L, or geranylgeranylpyrophosphate at 10 µmol/L was added during preincubation with lovastatin and again during incubation with lovastatin and sulindac. Geranylgeranylpyrophosphate and mevalonate completely reversed the lovastatin augmentation of sulindac-induced apoptosis, but farnesylpyrophosphate had no effect (Figure 4A). Similar results were obtained when cell viability was measured using the MTT assay (Figure 4B). Mevalonate, farnesylpyrophosphate, and geranylgeranylpyrophosphate used in similar doses failed to alter sulindacinduced apoptosis in the absence of lovastatin (data not shown). Lovastatin decreases the expression of COX isoforms in colon cancer cell lines. To study the effects

of lovastatin on the expression of COX isoforms (COX-1 and COX-2) and their role in the augmentation of sulindac-induced apoptosis, their expression was determined in HCT-116 and LoVo cells (Figure 5). Pretreatment lovastatin decreased the expression of COX-1 and COX-2 in LoVo cells. In HCT-116 cells, COX-2 expression remained undetectable and COX-1 expression increased with lovastatin. Because augmentation of apoptosis occurred in HCT-116 and SW-480 cells despite absent COX-2 expression, it is unlikely that lovastatin acts by modulating COX-2 expression. Similarly, in HCT-116 cells COX-1 expression was greater after lovastatin treatment, yet sulindac-induced apoptosis increased. Together these data suggest that lovastatin-induced augmentation of apoptosis is probably not caused by changes in COX-1 or COX-2 expression. In Vivo Studies Effect of lovastatin and sulindac alone or in combination on colonic ACF formation. The body

Figure 1. Lovastatin induces apoptosis in colon cancer cells. HCT116, SW480, and LoVo colon cancer cells were grown to 50% confluence, incubated with lovastatin (䊏, 0 µmol/L; R, 10 µmol/L; O, 30 µmol/L) for 48 hours, and harvested for quantification of apoptosis by flow cytometry. Apoptosis was determined as the percentage of subdiploid cells. Data are mean ⫾ SD. P ⬍ 0.05 was considered statistically significant (vs. no lovastatin). Similar data were obtained in 3 separate experiments. *Not statistically significant.

weights of azoxymethane-treated animals given lovastatin or sulindac alone or in combination were comparable to those of animals fed the control diet throughout the study (P ⬎ 0.05; data not shown). Examination of sections of proximal and distal colon by light microscopy showed no evidence of histological damage. Table 1

842 AGARWAL ET AL.

GASTROENTEROLOGY Vol. 117, No. 4

Figure 2. Lovastatin before treatment augments sulindac-induced apoptosis in HCT-116, SW480, and LoVo colon cancer cells. HCT-116 (left panels), SW480 (middle panels), and LoVo (right panels) colon cancer cells were grown to 50% confluence and preincubated with lovastatin for 48 hours. Cells then were incubated for 48 hours with lovastatin with or without 0, 100, 250, and 500 µmol/L sulindac and harvested for quantification of apoptosis by flow cytometry. 䊏, Apoptosis induced by sulindac alone. Lovastatin before treatment resulted in a dose-dependent augmentation of apoptosis. Because of varying sensitivities to lovastatin-induced apoptosis, different doses of lovastatin were used for HCT-116 (0, 10, and 30 µmol/L), SW480 (0, 5, and 10 µmol/L), and LoVo (0, 10, and 20 µmol/L) cells. Data are mean ⫾ SD. P ⬍ 0.05 was considered statistically significant (vs. no lovastatin). Similar data were obtained in 3 separate experiments. *Not statistically significant. (A ) Apoptosis quantified as percentage of subdiploid cells. (B ) Apoptosis quantified by BrdU/terminal deoxynucleotidyl transferase staining. Similar data were obtained in 3–6 separate experiments.

shows the results of studies of colonic ACF in F344 rats treated with azoxymethane. In animals fed the control diet without either sulindac or lovastatin, the number of ACF counted was 161 ⫾ 11 (mean ⫾ SEM). Salinetreated rats had no ACF (data not shown). Administration of 50 ppm lovastatin or 80 ppm sulindac resulted in a slight (12% and 15%, respectively) nonsignificant reduction in the mean number of ACF. In striking contrast, administration of lovastatin and sulindac combination reduced the mean number of colonic ACF per rat to 116 ⫾ 8, representing a decrease from control levels of 28% (P ⫽ 0.004).

To evaluate this reduction in ACF in more detail, we determined the number of individual crypts found in the ACF under the different experimental conditions (Table 1). We found no significant differences between the experimental conditions in the number of ACF containing only a single crypt. In ACF containing 2 or 3 crypts, there was a slight but significant reduction with administration of sulindac or combination of lovastatin plus sulindac. A highly significant reduction (30%; P ⫽ 0.02) in the number of ACF containing 4 or more crypts was found only with the combination of lovastatin plus sulindac.

October 1999

LOVASTATIN SENSITIZES COLON CANCER CELLS TO APOPTOSIS 843

Figure 3. Lovastatin before treatment reduces the number of viable cells after sulindac treatment. (A ) SW480 and (B ) HCT-116 cells were grown in 96-well plates and preincubated with lovastatin for 48 hours. Lovastatin with and without sulindac was added, and cells were incubated for an additional 48 hours. 䊏, 0 µmol/L lovastatin; O, 10 µmol/L lovastatin; 䊐, 30 µmol/L lovastatin. MTT assay then was performed as described in the protocol. The data are presented as the fraction of the optical density of the control wells and plotted as a percentage. Differences in cell viability after sulindac treatment were statistically significant (P ⬍ 0.001) between control incubation (no lovastatin) and preincubation with 10 and 30 µmol/L lovastatin at all concentrations of sulindac used. Data are mean ⫾ SD.

Discussion This study shows that lovastatin induces apoptosis in 3 colon cancer cell lines, HCT-116, SW480, and LoVo. Lovastatin also augmented sulindac-induced apoptosis in all 3 cell lines. The augmentation of apoptosis was prevented by cotreatment with mevalonate and geranylgeranylpyrophosphate but not by farnesylpyrophosphate, suggesting that inhibition of geranylgeranylation of some cellular protein is the principal mechanism of this effect. These isoprenoids did not alter sulindac-induced apoptosis in the absence of addition of lovastatin. Lovastatin treatment was associated with decreased expression of COX-1 and COX-2 in LoVo cells. In HCT-116 cells, no COX-2 was detected and COX-1 expression increased. In an established model of colon cancer prevention in F344 rats treated with azoxymethane, combination of lovastatin and sulindac was significantly more effective in reducing ACF formation than sulindac or lovastatin used alone. These data suggest a role for this combination in the chemoprevention of colon cancer. Lovastatin augmented sulindac-induced apoptosis in HCT-116, SW480, and LoVo cells. The increase in apoptosis was measured by 3 independent methods. This increase in apoptosis was more than the sum induced by lovastatin or sulindac when used alone at similar doses. Inhibition of isoprenylation, particularly geranylgeranylation of small G proteins, possibly rho, rac, cdc42, etc., appears to be critical because augmentation of apoptosis was reversed by small concentrations of geranylgeranyl-

pyrophosphate but not farnesylpyrophosphate. The downstream events may involve specific interactions such as a decrease in the expression of COX enzyme or alteration in ␤-catenin signaling, or lovastatin may nonspecifically increase the susceptibility to apoptosis by modulating the expression of apoptosis-related genes. Lovastatin potentiated sulindac-induced apoptosis in LoVo cells, which express COX-1 and COX-2, and decreased their expression, which could account for their increased susceptibility. However, lovastatin also increased sulindac-induced apoptosis in HCT-116 and SW480 cells that do not express COX-2.40,41 In HCT116 cells, COX-2 levels remained undetectable and COX-1 expression increased, yet lovastatin treatment markedly increased sulindac-induced apoptosis, suggesting that alterations in COX-1 and COX-2 are not primarily responsible. Furthermore, in HCT-116 cells, lovastatin caused similar augmentation of apoptosis induced by 50–150 µmol/L sulindac sulfide (which inhibits COX-1 and COX-2) and also 100–500 µmol/L sulindac sulfone (which does not inhibit COX-1 or COX-2) (data not shown). Together these data suggest that lovastatin augments sulindac-induced apoptosis by a COX-independent mechanism. Another possible mechanism may involve lovastatininduced alterations in ␤-catenin expression or function to sensitize cells to sulindac-induced apoptosis. ␤-Catenin binds to either hTcf4/Lef, E-cadherins, or the adenomatous polyposis coli (APC) gene product at any one time.

844 AGARWAL ET AL.

GASTROENTEROLOGY Vol. 117, No. 4

Figure 5. COX expression in lovastatin-treated cells. LoVo and HCT116 cells were incubated with lovastatin for 48 hours and then collected and studied as described in Materials and Methods. Lane 1, controls for COX-1 and COX-2; lane 2, untreated LoVo cells; lane 3, LoVo cells treated with 20 µmol/L lovastatin; lane 4, untreated HCT-116 cells; and lane 5, HCT-116 cells treated with 30 µmol/L lovastatin. In LoVo cells, lovastatin treatment results in decreased expression of COX-2 and expression of COX-1 remains undetectable. In HCT-116 cells, COX-1 expression increases and that of COX-2 remains undetectable. Similar results were obtained in 3 separate experiments.

Figure 4. Effect of addition of intermediates in the cholesterol synthetic pathway with lovastatin on sulindac-induced apoptosis. (A ) SW480 cells were grown to 50% confluence and preincubated for 48 hours with 15 µmol/L lovastatin with or without isoprenoids (100 µmol/L mevalonate [Mev], 100 µmol/L farnesylpyrophosphate [FPP], or 10 µmol/L geranylgeranylpyrophosphate [GGPP]). Cells were then incubated for 48 hours with 250 µmol/L sulindac with or without lovastatin and isoprenoids and then collected for flow cytometry. Apoptosis was quantified as number of subdiploid cells and is plotted as the fraction of total cell number. Data are mean ⫾ SD. (B ) SW480 cells were grown in 96-well plates and preincubated for 48 hours with 30 µmol/L lovastatin alone or with 100 µmol/L mevalonate, 100 µmol/L farnesylpyrophosphate, or 10 µmol/L geranylgeranylpyrophosphate. Sulindac at 100, 250, or 500 µmol/L was then added with or without lovastatin with the isoprenoids, and cells were incubated for 48 hours. MTT assays then were performed as described in the protocol. Data are the fraction of the optical density of control wells and plotted as a percentage (mean ⫾ SD). Similar data were obtained in 3 separate experiments and also in HCT-116 cells (not shown).

Binding of ␤-catenin with hTcf4/Lef results in transcription of several genes important in malignant transformation. Binding of APC with ␤-catenine results in degradation of ␤-catenin. In addition, the APC gene product has an important role in apoptosis, and insertion of wild-type APC gene in colon cancer cells with mutated APC gene induces apoptosis.42 Sulindac and other NSAIDs are believed to simulate the APC gene product function to induce apoptosis.42 E-Cadherin/␤-catenin binding requires rho proteins (for review, see Ilyes and Tomlinson43), and lovastatin can alter rho function by inhibiting its geranylgeranylation and thereby may alter the amount of ␤-catenin bound to E-cadherin. These lovastatininduced changes in ␤-catenin that may result in the increase in sulindac-induced apoptosis are being studied in our laboratory. The isoprenoids do not appear to have a direct role in sulindac-induced apoptosis because addition of mevalonate, farnesylpyrophosphate, or geranylgeranylpyrophosphate failed to alter apoptosis induced by sulindac alone. Alternately, lovastatin may nonspecifically increase susceptibility to apoptosis in colon epithelial cells. We showed previously that lovastatin augments apoptosis induced by the chemotherapeutic drugs 5-fluorouracil and cisplatin in colon cancer cell lines.34 In those studies, lovastatin treatment resulted in increased expression of a proapoptotic protein bax and decreased expression of an antiapoptotic protein bcl-2. In other studies, upregulation of caspase-7 has been reported after lovastatin treatment.44 Thus, lovastatin may modulate these apoptosis-related genes to make cells become more susceptible to apoptosis. Sulindac and other NSAIDs have been shown to reduce the incidence of colon neoplasia in patients with familial adenomatous polyposis9,13 and also in animal models of

October 1999

LOVASTATIN SENSITIZES COLON CANCER CELLS TO APOPTOSIS 845

Table 1. Effect of Lovastatin and Sulindac Individually and in Combination on ACF Crypt Multiplicity No. of crypts per colonic ACF Experimental group Control Lovastatin 50 ppm Sulindac 80 ppm Lovastatin 50 ppm ⫹ sulindac 80 ppm

1/ACF

2/ACF

Total ACF/colon

3/ACF

4/ACF

23 ⫾ 2.2 47 ⫾ 2.7

51 ⫾ 3.2

40 ⫾ 4.5

161 ⫾ 12

22 ⫾ 2.0 42 ⫾ 2.5

35 ⫾ 3.1b 43 ⫾ 2.9

142 ⫾ 14

30 ⫾ 3.0 39 ⫾ 2.7a 34 ⫾ 2.8b 36 ⫾ 2.1

137 ⫾ 5.4

21 ⫾ 2.1 32 ⫾ 2.4b 34 ⫾ 2.3b 28 ⫾ 3.4a 116 ⫾ 8.1b

NOTE. The number of crypts present in individual colonic ACFs were determined in 12 rats in each experimental group. Data are mean ⫾ SE. aP ⬍ 0.05, bP ⬍ 0.01 vs. controls.

colon carcinogenesis.10 The chemopreventive effect of NSAIDs is currently believed to result from increased apoptosis.20 Lovastatin, by augmenting the apoptosis induced by NSAIDs, could potentially increase their chemopreventive effect. To study this possibility, we used an azoxymethane model of colon carcinogenesis in F344 rat in which the formation of ACF was used as the end point. In this experimental model, a reduction in the total number of ACF and the number of ACF containing 4 or more crypts are accepted as valid markers for screening compounds with potential chemopreventive effect on colon neoplasia.36,37,45,46 Pretlow et al.47 showed that ACF also can be found in the colons of patients with colon cancer,47 and a recent human study showed that the number of ACF decreases after sulindac administration.48 In previous chemopreventive studies of colon neoplasia by Reddy et al.,49 320 ppm sulindac (80% of the maximum tolerated dose) reduced the number of colonic ACF by 36% and their multiplicity by 40%.49 In the present study, 80 ppm sulindac (20% of the maximum tolerated dose) produced a modest but nonsignificant reduction in colonic ACF number and crypt multiplicity. However, the combination of lovastatin and sulindac showed a pronounced reduction in the number of ACF (P ⬍ 0.004) and crypt multiplicity (P ⬍ 0.02). This reduction in colonic ACF is almost identical to that found previously when 320 ppm sulindac was administered alone. These data strongly imply that the combination of an HRI such as lovastatin with an NSAID such as sulindac may be more effective than sulindac alone in reducing the development of colonic neoplasia. Whether this combination of the 2 agents is appropriate to treat patients with FAP must await future studies in animals with genetically induced defects in the FAP gene. In our study, 50 ppm lovastatin induced a slight but

insignificant reduction in the total number of colonic ACF in azoxymethane-treated rats. However, Narisawa et al.32,33 found that pravastatin (5 ppm and 25 ppm) or simvastatin (10 ppm) reduced the number of colonic tumors in mice treated with dimethylhydrazine or N-methyl-N-nitrosourea.32,33 More long-term studies using tumor reduction in carcinogen-treated animals as end points are needed to clarify this discrepancy. Three clinical trials of HRIs with a follow-up of 5 years or longer have reported colon cancer incidence rates. In these trials, the chemopreventive effect of HRIs increases with the usage of NSAIDs (Table 2). These data suggest that HRIs may not be effective by themselves but potentiate the chemopreventive effects of NSAIDs. The chemopreventive effect of HRIs alone cannot completely be discounted because a 5-year follow-up may not be sufficient to observe the chemopreventive effect. In this study, we observed that the combination of HRIs with sulindac inhibits development of ACF, which is an early step in colon cancer carcinogenesis. Recently, several studies examined the potential chemotherapeutic effect of farnesyl protein transferase inhibitors (FTIs) on colon cancer, although no studies have explored their role in colon cancer prevention.51 Lovastatin has a somewhat similar mechanism of action as the FTIs (both inhibit isoprenylation of cellular proteins52; it is conceivable that FTIs also potentiate the chemopreventive effect of sulindac in colon cancer). FTIs are currently being studied in phase I and II trials so that there are no data on their long-term safety. On the other hand, HRIs such as lovastatin have been in clinical use for more than 15 years with no significant toxicity even when used for long periods. HRIs such as lovastatin seem to be ideal candidates to be combined with NSAIDs in studies of prophylaxis of colon cancer. NSAIDs have major dose-dependent gastrointestinal side effects that limit their use in chemoprevention. If lovastatin can amplify the chemopreventive effects of NSAIDs, then lower doses of potentially toxic NSAIDs such as sulindac could be used. Table 2. Chemopreventive Effect of HRIs on Colon Cancer in Published Clinical Trials

Trial

HRI used

No. of patients

Mean followup ( yr )

CARE30 4S31 TexCAPS50

Pravastatin Simvastatin Lovastatin

4159 4444 6605

5.0 5.4 5.2

% of patients on NSAIDs

% Reduction in colon cancer

83 37 17

43 19 0

NOTE. The protective effect of HRIs on colon cancer increases with the use of NSAIDs. HRIs by themselves do not seem to have a chemopreventive effect.

846 AGARWAL ET AL.

The concentrations of lovastatin and sulindac used in cell culture experiments are high and are not achievable in vivo. Previous in vitro studies also used similar doses of sulindac and HRIs. Of note, the levels of apoptosis induced also are very high; it is unlikely that such a degree of apoptosis is desirable in vivo, particularly for chemoprevention. The dose of lovastatin used in in vivo studies translates to approximately 1.5–2.5 mg/kg weight per animal, which is only modestly greater than 0.25–1.0 mg/kg weight used clinically in humans for lowering serum cholesterol. Extrapolation of these amounts to humans is limited because the pharmacokinetics of lovastatin differ in rats and humans. A reduction in colon cancer incidence was noted with pravastatin and simvastatin at doses used for coronary artery disease prophylaxis in clinical trials.30,31 These observations suggest that lower doses than those used in present study are effective. Further studies are needed to show this conclusively. To summarize, lovastatin augments apoptosis induced by sulindac in colon cancer cell lines. This effect can be prevented by addition of mevalonate and geranylgeranylpyrophosphate but not by farnesylpyrophosphate. Lovastatin also augments the chemopreventive effect of sulindac for colon neoplasia in an experimental animal model. Our data imply that lovastatin combined with NSAIDs may have an important synergistic role in the chemoprevention of colon cancer.

GASTROENTEROLOGY Vol. 117, No. 4

10.

11. 12.

13.

14.

15. 16. 17.

18.

19.

References 1. Shiff SJ, Qiao L, Tsai L-L, Rigas B. Sulindac sulfide, an aspirin-like compound, inhibits proliferation, causes cell cycle quiescence, and induced apoptosis in HT-29 colon adenocarcinoma cells. J Clin Invest 1995;96:491–503. 2. Shiff SJ, Koutsos MI, Qiao L, Rigas B. Nonsteroidal antiinflammatory drugs inhibit the proliferation of colon adenocarcinoma cells: effects on cell cycle and apoptosis. Exp Cell Res 1996;222:178– 188. 3. Metzger U, Meier J, Uhischmid G, Weihe H. Influence of various prostaglandin synthesis inhibitors on DMH-induced rat colon cancer. Dis Colon Rectum 1984;27:366–369. 4. Pollard M, Luckert PH. Effect of piroxicam on primary intestinal tumors induced in rats by N-methylnitrosourea. Cancer Lett 1984;25:117–121. 5. Reddy BS, Maruyama H, Kelloff G. Dose-related inhibition of colon carcinogenesis by dietar y piroxicam, a non-steroidal antiinflammatory drug, during different stages of rat colon tumor development. Cancer Res 1987;47:5340–5346. 6. Rao CV, Rivenson A, Simi B, Zang E, Kelloff G, Steele V, Reddy BS. Chemoprevention of colon carcinogenesis by sulindac, a nonsteroidal anti-inflammatory agent. Cancer Res 1995;55:1464– 1472. 7. Reddy BS, Rao CV, Rivenson A, Kelloff GJ. Inhibitory effect of aspirin on azoxymethane-induced carcinogenesis in F344 rats. Carcinogenesis 1993;14:1493–1497. 8. Tonelli F, Valanzano R, Dalora P. Sulindac therapy of colorectal polyps in familial adenomatous polyposis. Dig Dis 1994;12:259– 264. 9. Giardiello FM, Hamilton SR, Krush AJ, Piantadosi S, Hylind LM, Celano P, Booker SV, Robinson CR, Offerhaus GJ. Treatment of

20.

21. 22.

23.

24.

25.

26.

27.

28.

colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 1993;398:1313–1316. Beazer-Barclay Y, Levy DB, Moser AR, Dove WF, Hamilton SR, Vogelstein B, Kinzler KW. Sulindac suppresses tumorigenesis in the min mouse. Carcinogenesis 1996;17:1757–1760. Thun MJ, Namboodiri NM, Heath CW. Aspirin use and reduced risk of fatal colon cancer. N Engl J Med 1991;325:1593–1596. Nugent KP, Spigelman AD, Philips RKS. Tissue prostaglandin levels in familial adenomatous polyposis patients treated with sulindac. Dis Colon Rectum 1996;39:659–662. Winde G, Schmid KW, Schlegel W, Fischer R, Osswald H, Bunte H. Complete reversion and prevention of rectal adenomas in colectomized patients with familial adenomatous polyposis by rectal low-dose sulindac maintenance treatment. Dis Colon Rectum 1995;38:813–830. Giardiello FM, Spannhake EW, DuBois RN, Hylind LM, Robinson CR, Hubbard WC, Hamilton SR, Yang VW. Prostaglandin levels in human colorectal mucosa: effects of sulindac in patients with familial adenomatous polyposis. Dig Dis Sci 1998;43:311–316. Marnett LJ. Aspirin and potential role of prostaglandins in colon cancer. Cancer Res 1992;52:5575–5589. Smith WL. Prostanoid biosynthesis and mechanisms of action. Am J Physiol 1992;263:F181–F191. Hanif R, Pittas A, Feng Y, Koutsos MI, Qiao L, Staiano-Coico L, Shiff SLl, Rigas B. Effect of non-steroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin-independent pathway. Biochem Pharmacol 1996;52:237–245. Piazza GA, Rahm ALK, Krutzsch M, Sperl G, Paranka NS, Gross PH, Brendel K, Burt RW, Alberts DS, Pamukcu R, Ahnen DJ. Antineoplastic drugs sulindac sulfide and sulfone inhibit cell growth by inducing apoptosis. Cancer Res 1995;55:3110–3116. Piazza GA, Alberts DS, Hixson LJ, Paranka NS, Li H, Finn T, Bogert C, Guillen JM, Brendel K, Gross PH. Sulindac sulfone inhibits azoxymethane-induced colon carcinogenesis in rats without reducing prostaglandin levels. Cancer Res 1997;57:2909–2915. Piazza GA, Rahm AK, Finn TS, Fryer BH, Li H, Stoumen AL, Pamukou R, Ahnen DJ. Apoptosis primarily accounts for the growth-inhibitory properties of sulindac metabolites and involves a mechanism that is independent of cyclooxygenase inhibition, cell cycle arrest and p53 induction. Cancer Res 1997;57:2452– 2459. Goldstein JL, Brown MS. Regulation of mevalonate pathway. Nature 1990;343:425–430. Jackson JH, Cochrane CG, Bourne JR, Solski PA, Buss JE, Der CJ. Farnesol modification of Kirsten-ras exon 4B protein is essential for transformation. Proc Natl Acad Sci USA 1990;87:3042– 3046. Newman CM, Magee AI. Post-translational processing of ras superfamily of small GTP-binding proteins. Biochem Biophys Acta 1993;1155:79–96. Porfiri E, Evans T, Chardin P, Hancock JF. Prenylation of Ras proteins is required for efficient hsos-1 promoted guanine nucleotide exchange. J Biol Chem 1994;269:22676–22677. Jones KD, Couldwell WT, Hinton DR, Su Y, He S, Anker L, Law RE. Lovastatin induces growth inhibition and apoptosis in human malignant glioma cells. Biochem Biophys Res Commun 1994;205: 681–687. Padayatty SJ, Marcelli M, Shao TC, Cunningham GR. Lovastatin induced apoptosis in prostate stromal cells. J Clin Endocrinol Metab 1997;82:1434–1439. Reedquist KA, Pope TK, Roess DA. Lovastatin inhibits proliferation and differentiation and causes apoptosis in lipopolysaccharide-stimulated murine-B cells. Biochem Biophys Res Commun 1995;211:665–670. Perez-Sala D, Mollinedo F. Inhibition of isoprenoid biosynthesis

October 1999

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

induces apoptosis in human promyelocytic HL-60 cells. Biochem Biophys Res Commun 1994;199:1209–1215. Satoh S, Isobe H, Ayukawa K, Sakai H, Nawata H. The effects of pravastatin, an HMG-CoA reductase inhibitor on cell viability and DNA production of rat hepatocytes. Life Sci 1996;59:1103– 1108. Sacks FM, Pfeffer MA, Moye LA, Rouleau JL, Rutherford JD, Cole TG, Brown L, Warnica JW, Arnold JMO, Wun C, Davis BR, Braunwald E. The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. N Engl J Med 1996;335:1001–1009. Pederson TR, Berge K, Cook TJ, Faergeman O, Haghfelt T, Kjekshus J, Tatu M, Mushner T, Olsson A, Pyoeraelae K, Thorgeisson G, Tobert JA, Widel H, Wilhelmson H. Safety and tolerability of cholesterol lowering agents with simvastatin during 5 years in Scandinavian simvastatin survival study. Arch Intern Med 1996; 156:3085–2092. Narisawa T, Morotomi M, Fukaura Y, Hasebe M, Ito M, Aizawa R. Chemoprevention by pravastatin, a 3-hydroxy-3-methylglutaryl– coenzyme A reductase inhibitor, of N-methyl-N-nitrosourea– induced colon carcinogenesis in F344 rats. Jpn J Cancer Res 1996;87:798–804. Narisawa T, Fukaura Y, Tanida N, Hasebe M, Ito M, Aizawa R. Chemopreventive efficacy of low dose of pravastatin, an HMGCoA reductase inhibitor, on 1,2-dimethylhydrazine–induced colon carcinogenesis in ICR mice. Tohuku J Exp Med 1996;180:131– 138. Agarwal B, Halmos B, Moss SF, Ramey WG, Holt PR. Lovastatin sensitizes colon cancer HT-29 cells to apoptosis induced by chemotherapeutic agents. Proc Am Assoc Cancer Res 1998; 39:68. Reddy BS, Rivenson A, El-Bayoumy K, Upadhyaya P, Pittman B, Rao CV. Chemoprevention of colon cancer by organoselenium compounds and impact of high- or low-fat diets. J Natl Cancer Inst 1997;89:506–512. Wargowich MH, Chen D-D, Jiminez A, Steele VE, Velasco M, Stephens C, Price R, Gray K, Kelloff GJ. Aberrant crypt as a biomarker for colon cancer: evaluation of potential chemopreventive agents in a rat. Cancer Epidemiol Biomarkers Prev 1996;5: 355–360. Li X, Darzynkiewicz Z. Labelling DNA strand breaks with BrdUTP detection of apoptosis and cell proliferation. Cell Prolif 1995;28: 572–579. Moss SF, Agarwal B, Arber N, Guan RJ, Krajewska M, Krajewski S, Reed JC, Holt PR. Increased intestinal Bak expression results in apoptosis. Biochem Biophys Res Commun 1996;223:199–203. Rao CV, Desai D, Simi B, Kulkarni N, Amin S, Reddy BS. Inhibitory effects of caffeic acid esters on azoxymethane-induced biochemical changes and aberrant crypt foci formation in rat colon. Cancer Res 1993;53:4182–4188. Sheng H, Shao J, Kirkland SC, Isakson P, Coffey RJ, Morrow J, Beauchamp RD, DuBois RN. Inhibition of human colon cancer cell growth by selective inhibition of cyclooxygenase-2. J Clin Invest 1997;99:2254–2259.

LOVASTATIN SENSITIZES COLON CANCER CELLS TO APOPTOSIS 847

41. Parker J, Kaplon MK, Alvarez CJ, Krishnaswamy G. Prostaglandin H synthetase expression is variable in human colorectal adenocarcinoma cell lines. Exp Cell Res 1997;236:321–329. 42. Morin PJ, Vogelstein B, Kinzler KW. Apoptosis and APC in colorectal tumorigenesis. Proc Natl Acad Sci USA 1996;93:7950– 7954. 43. Ilyas M, Tomlinson IPM. The interactions of APC, E-cadherin and ␤-catenin in tumor development and progression. J Pathol 1997; 182:128–137. 44. Marcelli M, Cunningham GR, Haidacher SJ, Padayatty SJ. Caspase-7 is activated during lovastatin-induced apoptosis of the prostate cancer cell line LNCaP. Cancer Res 1998;58:76–83. 45. Deschner EE. Experimentally induced cancer of the colon. Cancer (Phila) 1974;34:824–828. 46. McLellan E, Medline A, Bird RP. Sequential analysis of growth and morphological characteristics of aberrant crypt foci: putative preneoplastic lesions. Cancer Res 1991;51:5270–5274. 47. Pretlow TP, O’Riordan MA, Pretlow TG, Stellato TA. Aberrant crypts in human colonic mucosa: putative preneoplastic lesions. J Cell Biochem 1992;16G(suppl):55–62. 48. Takayama T, Katsuki S, Takahashi Y, Ohi M, Noijiri S, Sakamaki S, Kato J, Koguwa K, Miyaka H, Niitsu Y. Aberrant crypt foci of the colon as precursors of adenoma and cancer. N Engl J Med 1998;339:1277–1284. 49. Reddy BS, Rao CV, Seibert K. Evaluation of cyclooxygenase-2 inhibitor for potential chemopreventive properties in colon carcinogenesis. Cancer Res 1996;56:4566–4569. 50. Downs JR, Clearfield DO, Weis S, Whitney E, Shapiro DR, Beere PA, Langendorfer A, Stein EA, Kruyer W, Gotto AM. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels. J Am Med Assoc 1998; 279:1615–1622. 51. Omer CA, Anthony NJ, Buser-Doepner CA, Burkhardt AL, deSolms SJ, Dinsmore CJ, Gibbs JB, Hartman GD, Koblan KS, Lobell RB, Oliff A, Williams TM, Kohl NE. Farnesyl: protein transferase inhibitors as agents to inhibit tumor growth. Biofactors 1997;6: 359–366. 52. Lebowitz PF, Davide JP, Prendergast GC. Evidence that farnesyltransferase inhibitors suppress ras transformation by interfering with rho activity. Mol Cell Biology 1995;15:6613–6622.

Received December 29, 1998. Accepted July 6, 1999. Address requests for reprints to: Peter R. Holt, M.D., Division of Gastroenterology, Department of Medicine, St. Luke’s–Roosevelt Hospital Center, 1111 Amsterdam Avenue, S&R12, New York, New York 10025. e-mail: [email protected]; fax: (212) 523-3683. The authors thank Dr. Jan Orenstein for reviewing the electron micrographs, Ann Washinton for immunoblotting, Dr. Steven Moss for reviewing the manuscript, Valerie Gray for typing the manuscript, and the Clark Foundation for providing the flow cytometry equipment.