Tel.: 416-946-2276, Fax: +416-946-2840
Differential interactions between statins and P-glycoprotein: Implications for exploiting statins as anticancer agents
Article first published online: 3 MAR 2010
Copyright © 2010 UICC
International Journal of Cancer
Volume 127, Issue 12, pages 2936–2948, 15 December 2010
How to Cite
Goard, C. A., Mather, R. G., Vinepal, B., Clendening, J. W., Martirosyan, A., Boutros, P. C., Sharom, F. J. and Penn, L. Z. (2010), Differential interactions between statins and P-glycoprotein: Implications for exploiting statins as anticancer agents. Int. J. Cancer, 127: 2936–2948. doi: 10.1002/ijc.25295
- Issue published online: 3 MAR 2010
- Article first published online: 3 MAR 2010
- Manuscript Revised: 11 FEB 2010
- Manuscript Received: 9 SEP 2009
- Province of Ontario (Ontario Institute for Cancer Research, Ministry of Health and Long Term Care, Ontario Graduate Scholarship)
- Canadian Cancer Society, Canadian Breast Cancer Foundation Ontario Region
- Leukemia and Lymphoma Society of Canada
- Canadian Institutes for Health Research (Excellence in Radiation Research for the 21st Century Strategic Training Initiative in Health Research)
- Canada Research Chairs Program
- multidrug resistance;
- Top of page
- Material and Methods
- Supporting Information
Statins, prescribed for decades to control cholesterol, have more recently been shown to have promising anticancer activity. Statins induce tumor-selective apoptosis by inhibiting the mevalonate (MVA) pathway. In addition, we have recently demonstrated that lovastatin modulates drug accumulation in a MVA-independent manner in multidrug-resistant (MDR) tumor cells overexpressing the P-glycoprotein (P-gp) multidrug transporter. P-gp-mediated drug efflux can contribute to chemotherapy failure. However, direct statin-mediated inhibition of P-gp in human MDR tumor cells at clinically achievable concentrations remains unexplored. An understanding of these interactions is crucial, both to appreciate differences in the anticancer potential of different statins and to safely and effectively integrate statins into traditional chemotherapy regimens that include P-gp substrates. Here we evaluate interactions between 4 statins (lovastatin, atorvastatin, fluvastatin and rosuvastatin) and P-gp, at both the molecular level using purified P-gp and at the cellular level using human MDR tumor cells. Lovastatin bound directly to purified P-gp with high affinity and increased doxorubicin accumulation in MDR tumor cells, potentiating DNA damage, growth arrest and apoptosis. By contrast, while atorvastatin inhibited substrate transport by purified P-gp in proteoliposomes, it had no effect on doxorubicin transport in MDR tumor cells. Finally, fluvastatin and rosuvastatin only interacted with P-gp in vitro at high concentrations and did not inhibit doxorubicin transport in MDR cells. These differential interactions should be considered when combining statins with traditional chemotherapeutic drugs.
Statins, commonly prescribed as cholesterol control agents, are currently being evaluated as potential anticancer therapeutics.1, 2 In the acid form, these drugs inhibit 3-hydroxy-3-methylglutaryl CoA reductase (HMGCR), the rate-limiting enzyme of the mevalonate (MVA) pathway.1, 3 Consequently, statins prevent the generation of important MVA-derived products, including cholesterol and isoprenoids (Fig. 1a). By depleting the MVA pool, statins can trigger apoptosis in a variety of tumor cell types, while nontransformed cells remain viable.1, 2, 4–7 This has led to clinical evaluation of statins in the treatment of several cancers. Positive responses have been observed in some subsets of patients, particularly in tumor type-specific clinical trials combining statins with traditional chemotherapeutic agents.8–11 However, the rationale for the choice of statin used has often remained undefined. An understanding of differences in the anticancer properties of each statin is necessary to effectively exploit this family of drugs in cancer therapy.
One characteristic that may differ between the statins is the capacity to interact with the xenobiotic transporter P-glycoprotein (P-gp, encoded by the ABCB1 gene). Although P-gp plays a physiological role in drug absorption and excretion, its deregulation in tumor cells can also lead to multidrug resistance (MDR).12–14 Classical MDR occurs when tumor cells develop cross-resistance to drugs of diverse structure and function, and may play a role in the failure of chemotherapy in several malignancies.13, 14 It has been well defined in cell culture systems that, in MDR tumor cells, upregulation of P-gp expression and activity can prevent intracellular accumulation of several chemotherapeutic drugs and can increase inherent resistance to apoptosis.13, 15 Thus far, attempts to evade MDR in tumor cells in vivo by pharmacologically inhibiting P-gp have been generally disappointing in clinical trials due to high toxicity and low efficacy (first generation modulators) or adverse pharmacokinetic interactions with coadministered drugs (second generation modulators).13, 16 However, highly selective and less toxic third generation modulators with minimal pharmacokinetic interactions are currently being evaluated in clinical trials.17
Some statins have been shown to inhibit the activity of human P-gp in ectopic expression systems18–20 or when P-gp is expressed endogenously in human colon epithelial cells.21, 22 These reports of P-gp inhibition have been inconsistent, and dependent on both the experimental systems and whether the acid or lactone chemical form of statin was used. Importantly, direct inhibition mediated by a physical interaction of statins with P-gp has not yet been demonstrated or quantified. Furthermore, the statin concentrations required for P-gp inhibition in cell systems have usually exceeded those achievable in systemic circulation. Up to 12.3 μM lovastatin has been detected in the plasma following administration of a high but well-tolerated dose,23 and based on the relative pharmacokinetics of each statin administered at a cholesterol-controlling dose (Table 1) it is likely that the other statins examined here could also achieve this level. It is therefore unclear how statins interact with P-gp in human MDR tumors and what therapeutic implications this would have for combining statins with chemotherapeutic agents that are P-gp substrates. To begin to address these questions, we have recently demonstrated that lovastatin increases accumulation of the P-gp substrate doxorubicin in P-gp-overexpressing ovarian tumor cells, thereby potentiating its anticancer therapeutic effect.26 Ongoing clinical trials (www.cancer.gov/clinicaltrials) indicate that statins will likely be incorporated into combination chemotherapy regimens that often include P-gp substrates, making an understanding of how statins interact with P-gp in human MDR tumor cells essential.
Here we have characterized the interactions between P-gp and the open-ring, acid form of 4 statins—lovastatin, atorvastatin, fluvastatin and rosuvastatin (Fig. 1b)—both in vitro and in various human MDR tumor cell types. We demonstrate that lovastatin interacted directly with P-gp in vitro and modulated MDR in tumor cells. Atorvastatin also interacted with P-gp in vitro, but could not modulate doxorubicin resistance in MDR tumor cells. Finally, rosuvastatin and fluvastatin interacted with P-gp only at high concentrations and did not inhibit doxorubicin transport in MDR cells. These results must be considered when integrating statins into combination treatments for both preclinical and clinical evaluation of anticancer efficacy, allowing statin selection to be customized to whether modulation of P-gp is desired.
Material and Methods
- Top of page
- Material and Methods
- Supporting Information
Lovastatin (Apotex, Toronto, Canada), fluvastatin sodium (Alexis Biochemicals, Lausen, Switzerland) atorvastatin calcium and rosuvastatin calcium (21 CEC Pharmaceuticals, East Sussex, UK) were obtained in powder form. Lovastatin was activated by alkaline hydrolysis, as previously described.27 For cell culture experiments, statins were dissolved in ethanol except rosuvastatin, which was dissolved in DMSO. For in vitro assays statins were dissolved in DMSO except lovastatin, which was prepared as above. Vincristine and cyclosporin A (CsA) were obtained from Calbiochem (San Diego, CA). Tetramethylrosamine (TMR) and Hoechst 33342 (H33342) were purchased from Invitrogen (Burlington, Canada), and dimyristoylphosphatidylcholine (DMPC) was obtained from Avanti Polar Lipids (Alabaster, AL). MB Biomedicals (Solon, OH) supplied the detergent 3-[3-(cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS). All other compounds were obtained from Sigma (Oakville, Canada).
Cell lines and cell culture
Cells were grown in a humidified incubator at 37°C in 5% CO2 and were regularly confirmed to be mycoplasma-free (MycoAlert mycoplasma detection kit, Lonza, Shawinigan, Canada). Cell culture medium (Ontario Cancer Institute Tissue Culture Media Facility, Toronto, Canada) was supplemented with 10% fetal bovine serum (Hyclone, Logan, UT). The ovarian carcinoma-derived A2780 and A2780ADR cell lines28 were provided by Dr. Jeremy Squire (Kingston General Hospital, Kingston, Canada) and cultured in RPMI 1640. MDR was maintained in A2780ADR cells by weekly treatment with 0.1 μM doxorubicin. The ovarian carcinoma-derived SKOV3 and SKVCR2.0 cell lines29 were obtained from Dr. Victor Ling (British Columbia Cancer Agency, Vancouver, Canada) and cultured in αMEM. SKVCR2.0 cells were grown under continuous selection with 2 μg/mL vincristine. The multiple myeloma cell lines 8226/S and 8226/Dox4030 were acquired from Dr. William Dalton (H. Lee Moffit Cancer Center and Research Institute, Tampa, FL) and cultured in RPMI 1640. 8226/Dox40 cells were treated weekly with 0.4 μM doxorubicin. The breast carcinoma-derived MCF7, MCF7DOX and MCF7TAX cell lines31 were provided by Dr. Amadeo Parissenti (Sudbury Regional Hospital, Sudbury, Canada) and cultured in DMEM H21. MCF7DOX cells were continuously cultured with 0.3 μM doxorubicin and MCF7TAX cells with 6.64 nM paclitaxel. All drug-resistant cells were cultured in the absence of selecting drug for a minimum of 1 passage prior to performing experiments.
ATPase activity of purified P-gp
P-gp was purified from the plasma membrane of CHRB30, a MDR Chinese hamster ovary cell line, using extraction with the detergent CHAPS, as described previously.32, 33 Purified P-gp was estimated to be >90% pure by SDS-PAGE, and reacted with the P-gp-specific antibody C219 on Western blots.33 The ATPase activity of purified P-gp in CHAPS buffer was determined in the presence of increasing statin concentrations as described elsewhere,34, 35 using a final ATP concentration of 2 mM and an incubation time of 20 min. Purified P-gp displayed basal ATPase activity in the range of 1.8–2.0 μmol/min/mg protein. The percent activity in the presence of statins was determined relative to a control with added DMSO, which changed the activity by <5%. The percent activity was determined relative to DMSO vehicle control. Triplicate samples were assayed in at least 2 independent experiments.
P-gp drug binding affinity
The binding affinity of purified P-gp for statins was determined as previously described, by measuring saturable quenching of the intrinsic tryptophan (Trp) fluorescence of the protein upon drug binding.36 Data were fitted to an equation describing drug binding to a single site, and the dissociation constant (Kd) and maximum percent fluorescence quenching (ΔFmax) were estimated. ΔFmax is variable, depending on the chemical nature of each drug, and the mechanisms by which Trp fluorescence is quenched. Kd depends on the concentration dependence of quenching, and is independent of ΔFmax. Triplicate samples were assayed in at least 2 independent experiments.
Substrate transport by P-gp in proteoliposomes
Purified P-gp was reconstituted into proteoliposomes of DMPC by gel filtration chromatography.32, 37 Transport of the fluorescent substrates TMR and H33342 was carried out at 27°C in a similar manner to that described earlier,32, 38 using final TMR and H33342 concentrations of 1 and 3.5 μM, respectively, 10 μg of P-gp per assay, and ATP and an ATP regenerating system. The initial rate of transport into proteoliposomes was determined relative to a control with added DMSO, which changed the activity by <5%. This initial rate of transport was estimated from the slope of the first 10 sec of the trace of fluorescence vs. time in 2 experiments, and the IC50 value for inhibition of transport was estimated from the plot of the mean initial rate vs. statin concentration.
P-gp cell surface expression
P-gp expression was determined essentially as previously described,39 resuspending 5 × 105 cells in 0.5 mL of staining buffer with or without 10 μL of a FITC-tagged anti-P-gp antibody (cat# 557002; BD Biosciences, Mississauga, Canada). Expression was detected by flow cytometry using a Becton Dickinson FACScan and analyzed using FlowJo software (v7.2.2, Tree Star). Three independent experiments were performed.
MTT assay for cell proliferation
The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was performed as described.6 Cells were seeded at subconfluence in 96-well plates, seeding 3,750 cells per well for all cell lines except 8226/S, 8226/Dox40 (18,750 cells per well) and SKOV3 (1,500 cells per well). The following day, cells were treated with a range of doses of doxorubicin for 48 hr. Three to six independent experiments were performed, each in triplicate. MTT50 values were computed from dose–response graphs using Prism (v5.0, GraphPad Software), referring to the concentration at which mitochondrial dehydrogenase activity in the cell population is decreased by 50%, as an indicator of viability.
Synergy assay for combination treatment
MTT assays were performed in A2780 and A2780ADR cells as described above. On a single 96-well plate, cells were treated with lovastatin or doxorubicin alone in a range of doses, and with a combination of these 2 drugs in a fixed 1:1 ratio as previously described.40 Lovastatin doses ranged from 12.5 nM to 2 μM in A2780 cells and from 2.4 to 38 μM in A2780ADR cells, ensuring a range of inhibition of cell proliferation. Doxorubicin doses ranged from 0.75 to 12 nM in A2780 cells and from 1.1 to 18 μM in A2780ADR cells. Combination Index (CI) analysis was performed using Calcusyn software (v2, Biosoft), according to the algorithms of Chou and Talalay.41 According to these calculations, the interaction between lovastatin and doxorubicin is additive if CI = 1, antagonistic if CI > 1, and synergistic if CI < 1.
Isoprenylation assay and immunoblotting
A2780ADR cells were treated with 20 μM statin or vehicle for 8–24 hr and harvested for Rap1 immunoblot, performed as described42 using either 15% or 8–16% gradient SDS-PAGE gels (Bio-Rad, Mississauga, Canada) and Rap1 (1:500; sc-65, Santa Cruz Biotechnology, Santa Cruz, CA) or α-tubulin (1:2,000; Calbiochem) antibodies. At least 3 independent experiments were performed.
Doxorubicin accumulation assay
Cells (5 × 105) were seeded in 6-well plates and incubated overnight. Cells were then treated for 3 hr with 1 μM doxorubicin alone or in combination with up to 10 μM lovastatin, atorvastatin and fluvastatin, up to 100 μM rosuvastatin, 200 μM MVA, 10 μM CsA, or vehicle. The geometric mean intracellular doxorubicin signal of 10,000 events was detected by flow cytometry for each sample. Relative intracellular doxorubicin accumulation was determined for each experiment by dividing geometric means by that of cells treated with only doxorubicin. Four to seven independent experiments were performed.
Alkaline comet assays were performed using reagents from the Trevigen Comet Assay Kit (Trevigen, Gaithersburg, MD), according to the manufacturer's instructions. A2780ADR and 8226/Dox40 cells were treated for 24 hr with vehicle, 1 μM doxorubicin, 10 μM lovastatin or 10 μM atorvastatin alone or in combination. The Olive tail moment (OTM) was determined for 100 cells per sample according to the calculation Olive tail moment = (Tail Length) × (Percent of DNA in Tail) / 100, using Komet software (v5.5, Kinetic Imaging). Three to four independent experiments were performed.
One million A2780ADR and 5 × 105 8226/Dox40 cells were seeded overnight in 10-cm dishes or 6-well plates, respectively, and then treated with 10 μM lovastatin, 10 μM atorvastatin, 10 μM doxorubicin and/or 1 μM doxorubicin for 48 hr. Dual staining was performed on ethanol-fixed cells with propidium iodide (PI) and terminal deoxynucleotide transferase dUTP nick end labeling (TUNEL) reagents according to the manufacturer's instructions, using the APO-BRDU Apoptosis Kit (Phoenix Flow Systems, San Diego, CA). Ten thousand events were scored and analyzed by flow cytometry as above. Four independent experiments were performed.
Modulation of doxorubicin accumulation was analyzed by performing 2-tailed 1-sample t-tests comparing the relative intracellular doxorubicin fluorescence of statin-treated cells to 1, using SPSS software (v16.0). Pairs of treatment groups were compared using 2-tailed, 2-sample unpaired t-tests with Welch's adjustment for heteroscedasticity, using SPSS software (v16.0). Multivariate effects, in particular synergy between statins and doxorubicin, were assessed with a general linear model (GLM), modeling the observed data points as: Y = StatinEffect + DoxoEffect + Statin:Doxo + BatchEffect. That is, the observed values (Y) were treated as a linear sum of the effect of statins alone (StatinEffect), the effect of doxorubicin alone (DoxoEffect), an interaction between the 2 treatments (synergy or antagonism), and a batch effect. A Statin:Doxo interaction term of 0 reflects additivity, < 0 reflects subadditivity (either saturation or antagonism) and > 0 reflects synergy. Models were fit in the R statistical environment (v2.8.1) and coefficients along with their standard errors and statistical significances were extracted for each term. p-values < 0.05 were considered statistically significant.
- Top of page
- Material and Methods
- Supporting Information
The 4 statins selected for study (Fig. 1b) are representative of the 6 currently available in North America, described in Table 1.24, 25 For this investigation, we did not include simvastatin because its physicochemical properties and putative interactions with P-gp closely mirror those of lovastatin,18, 43 or pravastatin because its lower lipophilicity leads to poor cellular uptake.44 Statins were studied in their acid forms, which inhibit HMGCR.
Statins interact directly with purified P-gp
Direct interaction between the 4 statins and P-gp was first investigated in vitro using purified P-gp isolated from the MDR Chinese hamster ovary cell line CHRB30. As an initial screen for direct functional interactions between the statins and P-gp, we examined the effect of statins on the constitutive ATPase activity of the transporter (Fig. 2). Substrates and modulators of P-gp-mediated drug transport typically modulate this activity; some drugs stimulate activity, others inhibit activity, and many display a biphasic pattern.45 At concentrations predicted to be clinically achievable (Table 1; Ref. 23), only lovastatin and atorvastatin had appreciable effects. Although lovastatin modulated ATPase activity in a biphasic manner, atorvastatin inhibited ATPase activity at concentrations > 1 μM. Rosuvastatin required high concentrations (at least 100 μM) to modulate ATPase activity, and fluvastatin had little effect at any concentration tested. Importantly, 3 of the 4 statins (lovastatin, atorvastatin and rosuvastatin) appeared to alter the ATPase activity of purified P-gp in some capacity, supporting further evaluation of the interactions between P-gp and this class of drugs. These different patterns of ATPase modulation are not understood at a mechanistic level. It should be noted that both substrates and modulators can affect P-gp ATPase activity, and the ATPase profile of a compound is not linked to its ability to reverse MDR in intact cells.
Direct binding of statins to P-gp was next determined by measuring quenching of tryptophan fluorescence within the protein. The Kd value of a P-gp substrate estimated from this approach was previously shown to correlate well with its competition for drug transport.46 All 4 statins demonstrated saturable quenching of tryptophan fluorescence, indicating that all were capable of binding directly to P-gp (Fig. 3a). Lovastatin had the highest binding affinity with a mean Kd of 2.7 with a range of ± 0.8 μM. The Kd values of the other tested statins were above reported clinically achievable plasma levels, ranging from intermediate values of 31 ± 1.9 μM and 55 ± 5.5 μM for fluvastatin and atorvastatin, respectively, to 203 ± 69 μM for rosuvastatin.
We next evaluated the capacity of statins to inhibit P-gp-mediated transport of 2 fluorescent substrates, TMR and H33342. Wang et al. have previously used flow cytometry to show that statins can inhibit uptake of a similar fluorescent rhodamine dye in intact cells.19 Here, purified P-gp was reconstituted into proteoliposomes, allowing substrates to be pumped into the lumen by inward-facing P-gp molecules when ATP is supplied, resulting in a rapid decrease in fluorescence intensity. The initial rate of substrate transport can be estimated from this decrease.32, 38 In the presence of statins, transport was inhibited with different potencies (IC50 values) (Fig. 3b). Although lovastatin was found to bind to P-gp with the highest affinity (Fig. 3a), it inhibited substrate transport in this system with only an intermediate potency (IC50 of 70 μM for TMR and 200 μM for H33342). By contrast, atorvastatin inhibited transport at concentrations < 1 μM for TMR and < 10 μM for H33342, despite its lower binding affinity. Fluvastatin inhibited transport at moderate concentrations (IC50 of 75 μM for TMR and 200 μM for H33342), consistent with its intermediate binding affinity. Rosuvastatin required high concentrations to have any inhibitory effect on TMR or H33342 transport (IC50 values of 600 and 250 μM, respectively). Transport inhibition studies in proteoliposomes cannot differentiate between P-gp substrates and modulators. Although both of these groups of compounds can compete for transport of fluorescence substrates in vitro, only studies of drug resistance in intact cells can identify modulators that reverse MDR.
Not all statins modulate doxorubicin transport in P-gp-overexpressing MDR tumor cells
The interactions between statins and P-gp were next examined in the context of tumor cells that have acquired MDR through continuous drug selection. A panel of paired parental (drug-sensitive) and MDR tumor cell lines was acquired, including the parental and MDR pairs A2780 and A2780ADR (ovarian carcinoma), SKOV3 and SKVCR2.0 (ovarian carcinoma), 8226/S and 8226/Dox40 (multiple myeloma) and a panel of MCF7, MCF7DOX and MCF7TAX lines (breast carcinoma). Each of the 5 MDR cell lines was confirmed to have overexpressed P-gp on the cell surface (Fig. 4a). Although most MDR cell lines displayed a robust upregulation of P-gp protein expression relative to parental cells, the paclitaxel-selected MCF7TAX cells demonstrated only moderate P-gp overexpression, consistent with their previous characterization.31 Performing an MTT assay for doxorubicin sensitivity demonstrated that all MDR cell lines were 2- to 129-fold more resistant to this P-gp substrate drug compared to parental cells (Supporting Information Table 1).
Whether the 4 statins could be taken up by MDR tumor cells to inhibit the MVA pathway was determined by assessing the geranylgeranylation status of the Rap1 protein, which is modified and proteolytically processed only when the pathway is intact.47 A2780ADR cells were treated with each statin at 20 μM for up to 24 hr. Unprocessed Rap1 was detected by anti-Rap1 immunoblot in MDR cells treated with the more lipophilic drugs, lovastatin, atorvastatin and fluvastatin, as a higher molecular weight band that was present after 16 hr of statin treatment (Fig. 4b, Supporting Information Fig. 1). In comparison, a relatively lower level of unprocessed Rap1 was detected in rosuvastatin-treated cells, even after 24 hr. This demonstrates that while all 4 statins can accumulate in MDR tumor cells and inhibit HMGCR activity, rosuvastatin appears to be less potent in this system.
To extend our in vitro results indicating that statins interact with P-gp (Figs. 2 and 3), we evaluated the effects of the 4 compounds on intracellular accumulation of doxorubicin in the panel of tumor cell lines. As expected, statins alone did not emit confounding fluorescence (data not shown), and the parental cell lines lacking P-gp expression exhibited no change in doxorubicin accumulation upon cotreatment with statins (Supporting Information Fig. 2). Concentrations of doxorubicin and the lipophilic statins were chosen based on what is predicted to be physiologically achievable.23, 48 By contrast, up to 100 μM rosuvastatin was evaluated, since it had reduced antiproliferative effects on tumor cells (data not shown), consistent with its delayed uptake (Fig. 4b). In MDR cells, the interactions between doxorubicin and the 4 statins differed (Fig. 5a, Supporting Information Fig. 3A). Only lovastatin increased doxorubicin accumulation in MDR cells in a dose-dependent manner. In all MDR cell lines, a statistically significant increase of ∼ 2-fold was observed at the highest lovastatin concentration tested (10 μM). This effect was not reversed by exogenous MVA, which rescues HMGCR inhibition (Fig. 5b, Supporting Information Fig. 3B).
The activity of lovastatin as an enhancer of doxorubicin accumulation in MDR cells was compared to that of CsA, a potent modulator of P-gp and other MDR transporters13, 49 (Fig. 5c, Supporting Information Fig. 3C). Administered at 10 μM, CsA was expected to completely inhibit doxorubicin transport by MDR transporters, since intracellular accumulation of coadministered doxorubicin was similar to that observed in parental cells (data not shown), and as such, this treatment served as a positive control for MDR pump inhibition, to which lovastatin could be compared. CsA increased doxorubicin accumulation within the MDR cells to a greater degree than lovastatin. In the case of the MCF7TAX cell line, the difference between the activities of CsA and lovastatin did not reach statistical significance, which may be due in part to the lower level of P-gp expressed in these cells. Finally, we assessed whether lovastatin increased intracellular doxorubicin accumulation primarily by inhibiting doxorubicin transport. To do this, MDR cells were again treated with a concentration of CsA expected to completely inhibit doxorubicin transport by MDR drug pumps. Adding lovastatin to this treatment did not lead to any further enhancement in intracellular doxorubicin levels, as expected for an agent modulating doxorubicin transport.
Lovastatin potentiates the anticancer effects of doxorubicin in MDR cells
Given that lovastatin partially reversed efflux-mediated doxorubicin resistance in MDR tumor cell lines, the anticancer therapeutic potential of combination treatments including lovastatin and doxorubicin were evaluated in 2 MDR cell lines. To gain an initial appreciation of the nature of the interaction between lovastatin and doxorubicin and whether it may be therapeutically beneficial, we performed synergy assays in A2780 and A2780ADR cells. Cells were treated with either drug alone or in combination, cell proliferation was monitored by MTT assay, and CI analysis was performed. In agreement with previous observations over a shorter time point,26 we found that in the parental A2780 cells lovastatin and doxorubicin have an additive effect (CI = 1), while in A2780ADR cells, lovastatin and doxorubicin have synergistic antiproliferative effects (CI < 1) (Supporting Information Fig. 4).
Next, the alkaline comet assay was used to determine that, in both A2780ADR and 8226/Dox40 MDR cell lines, combining lovastatin and doxorubicin at sublethal doses led to a trend toward potentiation of the DNA damaging effects of doxorubicin, measured as an increase in the mean OTM (Fig. 6a). Atorvastatin had no effect on doxorubicin-induced DNA damage, consistent with its inability to modulate doxorubicin accumulation.
We next examined changes in cell cycle profile and apoptosis levels by fixed PI and TUNEL staining, respectively, upon combined treatment for 48 hr with a low, clinically relevant concentration of doxorubicin (1 μM) and either lovastatin or atorvastatin (Fig. 6b). In addition, cells were treated with 10 μM doxorubicin as a positive control for induction of cell cycle arrest and/or apoptosis. In A2780ADR cells, 1 μM doxorubicin had no significant effect on cell cycle and induced only low levels of apoptosis. However, combining lovastatin and doxorubicin potentiated doxorubicin-induced cell cycle arrest at the G2/M phase. This growth arrest was not reversed by MVA. Furthermore, the combination treatment led to a statistically significant potentiation of apoptosis, yielding greater levels of apoptosis than when treating cells with 10-fold more doxorubicin alone. Consistent with the accumulation and DNA damage assays, atorvastatin did not alter growth arrest or apoptosis in doxorubicin-treated cells.
In 8226/Dox40 cells, 1 μM doxorubicin induced both G2/M arrest and apoptosis to a minor degree. Combining lovastatin with doxorubicin significantly potentiated doxorubicin-induced G2/M arrest and a led to a ∼ 2-fold increase in apoptosis. Although the potentiation of apoptosis was not as substantial as that seen in A2780ADR cells, the combination of lovastatin and 1 μM doxorubicin had comparable effects to treatment with 10 μM doxorubicin alone. Atorvastatin was not found to enhance doxorubicin-induced growth arrest or apoptosis. Overall, lovastatin consistently potentiated the anticancer effects of doxorubicin in MDR cells.
- Top of page
- Material and Methods
- Supporting Information
An understanding of how statins interact with P-gp is essential when considering coadministering statins with chemotherapeutic drugs that are P-gp substrates for 2 reasons. First, if statins modulate MDR in tumor cells, this may increase the efficacy of chemotherapy and response. Second, if statins modulate the physiological function of P-gp, they may affect the absorption and clearance, and hence toxicity, of coadministered P-gp substrate drugs. Here, we have examined these interactions in preclinical systems to provide a strong rationale for further preclinical and clinical studies.
This is the first report to comprehensively characterize how the acid forms of 4 different statins interact directly with purified P-gp and their ability to modulate drug transport in P-gp-containing proteoliposomes. In addition, we have examined the ability of these statins to modulate the MDR phenotype in a variety of drug-selected human cell types, at clinically achievable concentrations. The fluorescence-based in vitro assays used have the advantage of yielding quantitative data characterizing the direct interactions of statins with P-gp at a molecular level, separated from the complex context of a cell that may both express other MDR drug transporters and exhibit pleiotropic responses to statin administration. On the other hand, examining how statins modulate MDR in human tumor cells may provide insights into how statins may affect the efficacy of coadministered chemotherapeutic drugs in future clinical studies. Here, we have focused on characterizing the interactions of the acid forms of the statins since atorvastatin, fluvastatin and rosuvastatin are administered to patients in this form and, while there is likely dynamic interconversion between lactone and acid forms in vivo, most of the lactone form of lovastatin commonly administered to patients as a prodrug is rapidly converted to the acid form under cell culture conditions.6 However, it remains possible that the lactone forms of these statins or metabolites thereof that may be present following oral administration in vivo may also act as modulators of P-gp and MDR and this warrants further investigation.
Although certain statins have previously been shown to modulate transport of P-gp substrates in intact cell systems,18–22, 43 it has remained unclear whether statins directly bind to P-gp to modulate its activity. The lactone form of lovastatin has been shown to inhibit P-gp in nonhuman cells ectopically expressing human P-gp18–20 and murine leukemia cells expressing murine P-gp.43 In these cell systems, the acid form was not found to inhibit P-gp. Interestingly, we have recently demonstrated that lovastatin acid can modulate doxorubicin resistance in MDR ovarian tumor cells.26 Here, we have shown that, in vitro, lovastatin stimulated the ATPase activity of purified P-gp at low concentrations and bound to it directly with high affinity. The low Kd value for lovastatin indicates that it may be the only drug in this class capable of directly binding P-gp expressed in tumors in vivo. Lovastatin also behaved as a moderate inhibitor of transport of 2 P-gp substrates. It has been previously noted, as demonstrated here, that the IC50 for substrate transport assessed in this system is often several-fold larger than that the Kd for binding of the candidate modulator to purified P-gp. Different potencies in the binding affinity and transport assays may be reconciled with binding being only the first event in the complex process of inhibiting substrate transport, and the effect of the proteoliposomal bilayer in drug partitioning.50 In addition, P-gp displays complex interactions between pairs of drugs (see below), which may be bound to the protein simultaneously.51 Overall, this is the first report of the acid form of lovastatin binding directly to P-gp and inhibiting transport of P-gp substrates in vitro.
Importantly, a clinically achievable concentration (10 μM) of lovastatin doubled doxorubicin accumulation within several types of MDR cells. Although this concentration has been well-tolerated in a patient with an advanced malignancy in a dose escalation trial,23 this does represent a dose substantially higher than that commonly administered for the purpose of cholesterol lowering. However, we have found that the anticancer actions of statins in tumor cells are both time- and dose-dependent, such that lower doses of statins administered over longer periods of time may have the equivalent effect of a higher dose administered more briefly.4, 26 Since the phenomenon of lovastatin-induced increased doxorubicin accumulation was not reversed by MVA, it may be a novel pleiotropic effect of lovastatin that, while not as potent as the action of CsA, did appear to involve inhibition of doxorubicin transport. In combination with the in vitro data, this suggests that P-gp is likely the transporter being directly targeted. It also remains possible that lovastatin may have complementary indirect effects on P-gp52 and may target additional MDR drug transporters.53 The enhanced doxorubicin accumulation in lovastatin-treated MDR cells was accompanied by potentiation of both DNA damage and growth arrest/apoptosis, which is also reflected in the detection of a synergistic interaction between the 2 drugs in MDR, but not parental, cells. This adds a level of complexity to the possible molecular mechanisms behind reported synergy between lovastatin and chemotherapeutic drugs that are substrates of P-gp.54
The differences observed in the potency of lovastatin in proteoliposome dye flux experiments and in reversing resistance to doxorubicin in intact cells may be explained by the fact that P-gp is known to have several interacting subsites for binding drugs within a large flexible pocket.55, 56 If 2 compounds compete for the same subsite, they inhibit each other's transport, while if they interact with different subsites, they may mutually stimulate each other's transport, or have little effect on each other. As shown in the recent X-ray crystal structure of P-gp, even very closely related structures (stereoisomers) can bind to different locations within the drug-binding pocket.51 Thus, compared to atorvastatin, the lower ability of lovastatin to inhibit transport of a fluorescent substrate in proteoliposomes suggests that its binding site overlaps poorly with the subsite occupied by the fluorescent probe, while the binding site for atorvastatin overlaps well. In intact cells, on the other hand, the statins compete with doxorubicin, and the degree of overlap between the statin and doxorubicin binding subsites will be different.
Considering the capacity of lovastatin to modulate P-gp-mediated MDR, it merits further study as part of a 2-pronged therapeutic approach for MDR tumors (Fig. 6c). If lovastatin is coadministered with P-gp substrates, there exists the possibility that inhibiting the physiological function of P-gp may result in decreased elimination of the substrate, requiring dose reduction. However, lovastatin and doxorubicin have been coadministered in mice with no noted adverse toxicity.57, 58 Further evaluation of lovastatin in combination with chemotherapeutic agents that are P-gp substrates for the treatment of MDR tumors is therefore warranted to examine whether the therapeutic potentiation observed in cell culture models is replicated in vivo.
Similarly to lovastatin, the lactone form of atorvastatin has been associated with P-gp inhibition in comparable cell systems. Results for the acid form have been mixed, depending on the system used.18–22, 43 Importantly, the pharmacokinetic properties of atorvastatin are appropriate for an anticancer agent, given its relatively high, sustained achievable plasma concentration, and that it does not require metabolic activation.24, 25 Here we have found that atorvastatin inhibited the ATPase activity of P-gp in vitro, in contrast to the biphasic stimulatory effect of lovastatin. These results suggest that the interactions of lovastatin and atorvastatin with P-gp may have subtle differences. Atorvastatin bound directly to P-gp with moderate affinity and performed as a potent P-gp transport inhibitor in vitro. However, in MDR tumor cells, atorvastatin had no effect on doxorubicin transport. The interaction between atorvastatin and P-gp may therefore depend on the P-gp substrate examined or the experimental systems used. Importantly, in the more complex MDR cell systems, the MDR machinery results in the upregulation of several other proteins that may circumvent the P-gp inhibition observed in vitro. In addition, when examining the activity of atorvastatin in cells, one must also consider the possible contributions of other transporter proteins affecting the uptake and intracellular distribution of atorvastatin, although we have shown here that uptake was sufficient to inhibit the MVA pathway (Fig. 4b). Given that the interaction of atorvastatin with P-gp and its substrates appears to be complex and context-dependent, the behavior of this statin in combination with P-gp substrate drugs requires further characterization.
Rosuvastatin and fluvastatin were not predicted to inhibit P-gp function in a clinically achievable concentration range. In a single report, rosuvastatin showed no evidence of inhibition of ectopically expressed human P-gp.18 Similarly, fluvastatin was not found to inhibit ectopic human or endogenous murine P-gp.18, 43 We have also found no evidence of P-gp inhibition, either in vitro or in MDR cells, by rosuvastatin or fluvastatin at physiologically relevant concentrations. It is unclear whether the low antiproliferative potency of rosuvastatin we have observed (data not shown) will be replicated in vivo. Fluvastatin, however, has pharmacokinetics appropriate for sustained systemic delivery and evidence of anticancer activity in conjunction with chemotherapy.54 Fluvastatin should therefore be further evaluated as a therapeutic approach when targeting P-gp is either unnecessary or undesirable.
Ultimately, this report marks a significant advance toward the ability to customize statin selection in clinical trials to a patient's MDR status. As outlined in Figure 6d, if lovastatin can be combined with chemotherapeutic agents that are P-gp substrates while avoiding adverse pharmacokinetic interactions in vivo, it could be evaluated as a promising therapeutic strategy in P-gp-expressing tumors. However, if general toxicity becomes a concern, a statin that does not interact with P-gp, like fluvastatin, could be clinically evaluated for P-gp-independent synergy with other chemotherapeutic drugs. Although all statins share the ability to induce tumor-selective apoptosis through depletion of MVA, statins clearly differ in their ability to modulate P-gp activity and MDR. Preclinical work should be continued and moved steadily forward into clinical trials, selecting statins with a rational approach geared toward whether P-gp inhibition is desired.
- Top of page
- Material and Methods
- Supporting Information
The authors thank members of the Penn and Sharom laboratories for helpful discussions and critical review of the manuscript, Dr. Susan Cole for helpful discussions, Ms. Amanda R. Wasylishen for technical assistance, all sources that have kindly provided cell lines, and Apotex (Toronto, Canada) for generously providing lovastatin. This study was conducted with the support of the Ontario Institute for Cancer Research through funding provided by the Province of Ontario (L.Z.P.), the Canadian Cancer Society (F.J.S.), an Ontario Graduate Scholarship (C.A.G.), the Canadian Breast Cancer Foundation Ontario Region (C.A.G., L.Z.P.), a fellowship from the Leukemia and Lymphoma Society of Canada (A.M.) and an Excellence in Radiation Research for the 21st Century Strategic Training Initiative in Health Research award from the Canadian Institutes for Health Research (J.W.C.). This research was also undertaken, in part, thanks to funding from the Canada Research Chairs Program (L.Z.P. and F.J.S.) and the Ontario Ministry of Health and Long Term Care (OMOHLTC). The views expressed do not necessarily reflect those of the OMOHLTC.
- Top of page
- Material and Methods
- Supporting Information
- 31Cross-resistance studies of isogenic drug-resistant breast tumor cell lines support recent clinical evidence suggesting that sensitivity to paclitaxel may be strongly compromised by prior doxorubicin exposure. Breast Cancer Res Treat 2004; 85: 31–51., , , , , , .
- Top of page
- Material and Methods
- Supporting Information
Additional Supporting Information may be found in the online version of this article.
|IJC_25295_sm_suppfigure-1.tif||22060K||Supplementary Figure 1. Statins inhibit the isoprenylation of Rap1, a surrogate assay for statin uptake and HMGCR inhibition. Total protein (25 μg) from lysed A2780ADR cells treated with 20 μM statins for up to 24 h was subjected to immunoblotting for Rap1. The full-length blot shown is representative of at least three independent experiments. lova, lovastatin; atorva, atorvastatin; fluva, fluvastatin; rosuva, rosuvastatin.|
|IJC_25295_sm_suppfigure-2.tif||19812K||Supplementary Figure 2. Statins do not alter doxorubicin accumulation in parental cells. Cells were treated as indicated for 3 h and intracellular doxorubicin fluorescence was detected by flow cytometry. Relative intracellular doxorubicin fluorescence was determined by normalizing all values to that of cells treated with doxorubicin alone (normalized to 1). In all panels, bars represent the mean of at least four independent experiments, with error bars indicating standard deviation. doxo, doxorubicin.|
|IJC_25295_sm_suppfigure-3.tif||19810K||Supplementary Figure 3. Lovastatin, but not all statins, modulates doxorubicin transport in additional MDR tumor cell lines. Cells were treated as indicated for 3 h and intracellular doxorubicin fluorescence was detected by flow cytometry. Relative intracellular doxorubicin fluorescence was determined by normalizing all values to that of cells treated with doxorubicin alone. A. Only lovastatin significantly increases intracellular doxorubicin accumulation in additional MDR tumor cell lines. p-values shown denote statistically significant differences from cells treated with doxorubicin alone (normalized to 1) by one-sample t-test. B. The effect of lovastatin is independent of inhibition of the MVA pathway. The p-values shown represent results from two-sample t-tests between bracketed groups. C. Lovastatin is less potent than CsA at inhibiting doxorubicin transport mediated by MDR drug transporters. The p-values shown represent results from two-sample t-tests between bracketed groups. In all panels, bars represent the mean of at least four independent experiments, with error bars indicating standard deviation. doxo, doxorubicin; lova, lovastatin; MVA, mevalonate; CsA, cyclosporin A.|
|IJC_25295_sm_suppfigure-4.tif||12218K||Supplementary Figure 4. Lovastatin synergizes with doxorubicin in an MDR cell line. Combination index analysis of MTT assay data gathered from parental A2780 cells treated with lovastatin and doxorubicin alone, and both drugs in combination, reveals no significant difference from an additive effect (CI=1) in the combination treatment. In the MDR A2780ADR cells, similar treatments reveal a synergistic interaction (CI<1) between lovastatin and doxorubicin. Bars represent the mean CI, from at least four independent experiments, at a given MTT reduction inhibitory concentration value, with error bars representing standard deviation. The p-values shown represent results testing differences from additivity (CI=1) by one-sample t-test.|
|IJC_25295_sm_suppinfotables.doc||55K||Supplementary Table 1. Sensitivity of human parental and multidrug-resistant tumor cell lines to doxorubicin by MTT assay. Supplementary Table 2. Coefficients from General Linear Modeling of Statin:Doxo interactions.|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.