The last two authors contributed equally to this paper.
Differences in antitumor effects of various statins on human pancreatic cancer
Article first published online: 20 NOV 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 122, Issue 6, pages 1214–1221, 15 March 2008
How to Cite
Gbelcová, H., Leníček, M., Zelenka, J., Knejzlík, Z., Dvořáková, G., Zadinová, M., Poučková, P., Kudla, M., Balaž, P., Ruml, T. and Vítek, L. (2008), Differences in antitumor effects of various statins on human pancreatic cancer. Int. J. Cancer, 122: 1214–1221. doi: 10.1002/ijc.23242
- Issue published online: 21 JAN 2008
- Article first published online: 20 NOV 2007
- Manuscript Accepted: 19 SEP 2007
- Manuscript Received: 4 JUL 2007
- Czech Ministry of Health. Grant Number: NR8132-3
- Czech Ministry of Education. Grant Numbers: MSM 6046137305, ME849
- Czech Science Foundation. Grant Number: 204/03/H066
- pancreatic cancer;
- HMG-CoA reductase;
- K-ras oncogene;
Statins are widely used for the treatment of hypercholesterolemia. However, their inhibitory action on HMG-CoA reductase also results in the depletion of intermediate biosynthetic products, which importantly contribute to cell proliferation. The aim of the present study was to compare the effects of the individual commercially available statins on experimental pancreatic cancer. The in vitro effects of individual statins (pravastatin, atorvastatin, simvastatin, lovastatin, cerivastatin, rosuvastatin and fluvastatin) on the viability of human pancreatic cancer were evaluated in CAPAN-2, BxPc-3 and MiaPaCa-2 cell lines. The in vivo experiments were performed on nude mice xenotransplanted with CAPAN-2 cells. The mice received oral treatments either with a placebo, or with the statins mentioned earlier in a daily dose corresponding to a hypocholesterolemic dose in humans. The effect of these statins on the intracellular Ras protein, trafficking in MiaPaCa-2 transfected cells, was also investigated. Substantial differences in the tumor-suppressive effects of all statins were detected in both in vitro and in vivo experiments. While simvastatin exerted the highest tumor-suppressive effects in vitro, rosuvastatin (p = 0.002), cerivastatin (p = 0.002) and fluvastatin (p = 0.009) were the most potent compounds in an animal model. All statins (except pravastatin) inhibited intracellular Ras protein translocation. In summary, substantial tumor-suppressive effects of various statins on the progression of experimental pancreatic adenocarcinoma were demonstrated, with marked differences among individual statins. These results support greatly the potential of statins for the chemoadjuvant treatment of pancreatic cancer. © 2007 Wiley-Liss, Inc.
Inhibitors of hydroxy-methyl-glutaryl coenzyme A (HMG-CoA) reductase (statins) are widely used for treatment of hypercholesterolemia. However, the effects of statins on human tissues are pleiotropic, involving inhibition of atherogenous plaque formation, platelet aggregation, the improvement of both endothelial function and fibrinolytic activity, or even direct protective effects of the statins upon mortality in acute myocardial infarction.1, 2 In addition, the inhibition of HMG-CoA reductase, a key enzyme in the cholesterol biosynthesis, also results in the depletion of several important intermediates, including farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP), which modify and target small GTPases to their site of action.3 Farnesylated Ras proteins are associated with multiple mitogenic signal transduction pathways in response to growth factor stimulation,4 including external signal regulating kinase 1/2 and phosphatidylinositol 3′-kinase/Akt signaling pathways to name the most important effectors.5 On the other hand, geranylgeranylated proteins of the Ras homologous (Rho) family regulate signal transduction from membrane receptors in a variety of cellular events related to cell adhesion and invasion.6 Accordingly, inhibition of farnesylation/geranylgeranylation became a plausible approach to modify cell proliferation in tumor tissues. During the last decade, the antiproliferative effects of statins were demonstrated in numerous in vitro as well as in vivo studies on various tumor cell lines including hepatocellular carcinoma,7 lung,8 colorectal9 or pancreatic cancer.10–15 In this regard, pancreatic cancer is of particular interest, since more than 90% of human pancreatic cancers bear activating mutations in the K-ras proto-oncogene.16 These mutations result in the loss of GTPase activity (physiologically associated with the Ras protein as a negative feedback mechanism), which leads to protracted K-Ras activation. Suppression of this event via statin-mediated inhibition of K-Ras farnesylation thus seems to be a promising therapeutic approach. In fact, antitumor activities of statins were also demonstrated in some human studies,17, 18 and also interestingly in several human epidemiological studies, primarily focused on cardiovascular outcomes.19–21
The antiproliferative effects for all marketed statins have been described. However, some data suggest certain differences in the antitumor effects of individual statins. Wong et al.22 demonstrated substantially higher inhibitory effects on the growth of acute myeloid leukemia cells of cerivastatin, compared to lovastatin, atorvastatin and fluvastatin. In another in vitro study on leukemia cell lines, simvastatin was the most effective statin; while, pravastatin had the weakest effect.23 A similar tendency was also observed in osteosarcoma cell lines treated with simvastatin or pravastatin,24 as well as in an in vitro breast cancer study by Mueck et al.25 Based on incomplete and scattered data, it seems that individual statins act differently on various diseases and cell populations,2 and this may be the reason for the inconclusive or controversial epidemiological data published so far. This might all result from the differences in the statins' structure, pharmacokinetics and biotransformation rates.26 For instance, large variations in the modulation of hepatic cytochrome P450 activity have been documented among individual statins.27
Although, as pointed out earlier, statins have been extensively studied as possible chemotherapeutic agents, and no complex data have as yet been provided on the differences in antitumor activities of the individual statins. Therefore, the aim of the present study was to compare the tumor-suppressive effects of the various statins routinely used in clinical medicine upon the growth of human pancreatic cancer in both in vitro and in vivo experimental models.
Material and methods
For the in vitro study, the following pure forms of statins were used: pravastatin, atorvastatin, lovastatin, simvastatin, fluvastatin, cerivastatin and rosuvastatin (all obtained from Alexis; San Diego, CA except for simvastatin, kindly provided by Merck, Sharp and Dohme, NJ). Simvastatin was either used in its native (lactone) form (believed not to inhibit HMG-CoA reductase) or in the active (lactam) form, prepared as described previously.28 Mevalonate, FPP and GGPP were purchased from Sigma (St. Louis, MO).
The following pancreatic cancer cell lines were used for the invitro studies: CAPAN-2, MiaPaCa-2 and BxPc-3 (ATCC, Manassas, VA). All cell lines were both maintained and grown in a humidified atmosphere containing 5% CO2 at 37°C, in the following media, supplemented with 10% fetal bovine serum: MiaPaCa-2 in DMEM, BxPc-3 in RPMI 1640 and CAPAN-2 in McCoy's 5A medium with 1.5 mM L-glutamine containing sodium bicarbonate (1.5 g/l).
All statins in the in vitro study were used in the concentration range of 0–40 μM (0; 10; 20; 30 and 40 μM). Three hundred microliters of the cell suspension (∼2.7 × 105 cells/ml) were used for inoculation of individual wells in the 6-well plate. Two milliliters of medium were added to each well, and the plates were incubated at 37°C in a humidified atmosphere of 5% CO2 in air. Statins from stock solutions in methanol (20 μl per well) were added 24 hr later to the final concentrations, indicated earlier. Twenty microliters of methanol were added to control wells. After 24, 48 and 72 hr of statin treatment, cells in triplicate were washed with PBS, harvested by 0.25% trypsin and resuspended. Both cell growth and viability were assessed by direct counting of 0.4% trypan blue dye-excluding cells.
Ras protein translocation assay
Reverse-transcription polymerase-chain reaction.
Total RNA was isolated from HeLa (with wild type K-ras gene) and MiaPaCa-2 (with activation mutation in K-ras gene) cell lines by RNeasy Kit (Qiagen, MD), and wild type; and the mutated 570-bp K-ras cDNA genes, respectively, were amplified by reverse-transcription polymerase-chain reaction (RT-PCR) using Enhanced Avian HS RT-PCR kit (Sigma) and following primers:
5′-primer of K-ras: 5′- TTCAGATCTATGACTGAATATAAACTTGTGGTAGTTGGAG -3′
3′-primer of K-ras: 5′- AAGGATCCTTACATAATTACACACTTTGTCTTTGACTTC -3′
The PCR products were purified using QIAquick PCR Purification Kit (250) (Qiagen).
Owing to the C-terminal processing of Ras, both wild type and mutant ras DNA sequences were ligated into pEGFP-C1 vector (Clontech, CA), downstream of the coding sequence for the green fluorescent protein (GFP), to generate the pEGFP-K-Ras and pEGFP-K-RasG12C vectors, respectively. Vectors were amplified in the Escherichia coli DH5α cells (Invitrogen, CA) and verified by sequencing. Vectors were then used for expression of N-terminally tagged K-Ras with GFP (GFP-K-Ras) in MiaPaCa-2 cells.
Transfection and localization imaging
MiaPaCa-2 cells were seeded in a single 6-well cell culture plate, with sterile glass coverslips 5 hr before transfection. Transfection with aforementioned plasmids was carried out by FuGene 6 (Roche, Basel, Switzerland) according to manufacturer instructions. After 24 hr the medium was changed, and statins to final concentration of 20 μM were added. After next 24 hr, cells were washed by PBS and fixed for 20 min with 4% formaldehyde in PBS. Actin filaments were stained with TRITC phalloidine (Sigma). The pEGFP-C1 vector was used as a control of transfection efficiency, and for observing the localization of the GFP protein (alone, with, and without the drug treatments). Intracellular localization of the individual proteins (GFP, GFP-K-Ras or GFP-K-RasG12C) and actin filaments was visualized by fluorescent microscopy, using QuickPHOTO CAMERA 2.1 processing software (Olympus, Tokyo, Japan).
The in vivo study was performed on nude mice (strain CD-1, Charles River WIGA, Sulzfeld, Germany) xenotransplanted subcutaneously with human pancreatic adenocarcinoma cell line CAPAN-2 (107 cells; n = 6 for each treatment group). After initiation of tumor growth (7–10 days after xenotransplantation; tumor size at the beginning of treatment was 0.27 ± 0.04 cm3), the mice received oral treatment with a placebo (saline) or one of the following commercially available statins: pravastatin (Lipostat, Bristol-Myers Squibb, NY), atorvastatin (Sortis, Pfizer, NY), lovastatin (Mevacor, Merck, Sharp and Dohme), simvastatin (Zocor, Merck, Sharp and Dohme), cerivastatin (Cholstat, Laboratories Fournier, Paris, France) or fluvastatin (Lescol, Novartis, Basel, Switzerland); given in a daily dose approximately corresponding to the hypocholesterolemic dose used in humans (Fig. 3). Drugs were administered intragastrically once daily via gastric tube. The primary endpoint was the survival time. Simultaneously, an assessment of tumor size was performed by measurements of the 2 greatest perpendicular diameters of the subcutaneous tumors, measured every 3 days with a caliper.29
All aspects of the animal studies met the accepted criteria for the care and experimental use of laboratory animals, and all protocols were approved by the Animal Research Committee of the 1st Faculty of Medicine Charles University in Prague.
Data are presented as the median and 25–75% range, or the mean ± SD. The statistical significance of differences between variables was evaluated by the Mann–Whitney Rank Sum test. The effect of statin treatment on the survival of animals was analyzed by a standard nonparametrical analysis (Mann–Whitney Rank Sum test), as well as Kaplan Meier Log Rank survival analysis with Holm–Sidak posthoc testing, when survival p-value was significant. Group mean differences in tumor size were measured by repeated measures analysis of variance (RM ANOVA) with Holm–Sidak posthoc testing, when p value was significant. When needed, log transform values of tumor size were used for comparisons to comply with normality and equal variance requirements. Differences were considered statistically significant when p values were less than 0.05.
The in vitro inhibitory effects of statins on viability of pancreatic adenocarcinoma cell lines
Tumor-suppressive effects of individual statins were analyzed in vitro by experiments on pancreatic cancer-derived cell lines. In contrast to well-differentiated adenocarcinoma cells, CAPAN-2 and MiaPaCa-2 cells are poorly differentiated. Both of these cell lines harbor activating K-ras mutations in codon 12, prevalently associated with pancreatic cancer.12 Generally, the presence of statins in the growth medium dramatically reduced the numbers of cancer cells (Figs. 1a–f, Table I). However, various statins exhibited significantly different inhibitory efficacy and we have also observed notable differences between the sensitivity among the individual cell lines. In the CAPAN-2 and MiaPaCa-2 cell lines, even the lowest tested concentrations (10 and 15 μM, respectively) of cerivastatin, simvastatin and lovastatin induced a cytostatic effect, evaluated as zero increase of the cell number, compared to the control at the time of statin application (Figs. 1a–d, Table I). The tumor-suppressive effect of fluvastatin and atorvastatin was slightly lower, corresponding to a 20 μM concentration. The tumor-suppressive effect of rosuvastatin was significantly lower than the aforementioned statins, and was observed at a concentration of 40 μM. Only slight cytotoxic effects (the prevalence of cell dying) were observed at the lower tested concentrations (Fig. 1). All statins (except pravastatin) exhibited an acute cytotoxic effect in concentrations higher than 40 μM, manifested by release of the cells into the growth medium and permeabilization of the plasma membrane, which was detected by trypan blue staining (data not shown).
|24 hr||48 hr||72 hr||24 hr||48 hr||72 hr||24 hr||48 hr||72 hr|
BxPc-3 cells are moderately differentiated cancer cells producing the wild type K-ras proto-oncogene30 and overexpressing cyclooxygenase-2.31 This cell line was less sensitive to the tumor-suppressive effect of statins especially in the short-term experiments (24 and 72 hr, Figs. 1e and 1f, Table I; data for 48 hr not shown). This is well documented by an increase of IC50 values for individual statin, compared to these values for the other cell lines tested (Table I). Interestingly, all the tested compounds (except pravastatin) exhibited a comparable cytostatic effect for both the BxPc-3 cells and the other cell lines, after 72-hr exposure. A high proportion of the cell population exhibited signs of degeneration, such as granulation of cytoplasmic material, as well as decreased cell viability, according to trypan blue staining. A total cytostatic effect was observed at the 60 μM concentration of cerivastatin, simvastatin, lovastatin and fluvastatin (data not shown).
The viability of cancer cells was substantially prevented by all statins, except pravastatin, which only reduced the growth progression when administered at a high concentration (40 μM). Although pravastatin was generally the least effective statin after 48-hr and 72-hr exposure, compared to rosuvastatin, it exhibited a better inhibitory effect in the 24-hr experiment (Table I).
Addition of mevalonate, as well as FPP or GGPP, substantially abrogated the inhibitory growth effect of all statins (Table II), suggesting that the effect was caused by effective inhibition of farnesylation and not by the possible toxicity of statins. When mevalonate was added in a concentration far exceeding 20× that of statins, their tumor-suppressive effect was completely eliminated, while equimolar concentrations of mevalonate only had a partial effect (Table II). Surprisingly, both forms of simvastatin (i.e., lactam and lactone) were found to have similar efficacy (data not shown).
|CAPAN-2 (% of control cells)||MiaPaCa-2 (% of control cells)||BxPc-3 (% of control cells)|
|Mevalonate 17 μM||76||–||–|
|Mevalonate 600 μM||98||92||81|
|Pravastatin + mevalonate 17 μM||91||–||–|
|Pravastatin + mevalonate 600 μM||98||112||87|
|Prava + FPP||90||–||–|
|Prava + GGPP||115||–||–|
|Lovastatin + mevalonate 17 μM||32||–||–|
|Lovastatin + mevalonate 600 μM||86||93||90|
|Lovastatin + FPP||56||–||–|
|Lovastatin + GGPP||79||–||–|
|Atorvastatin + mevalonate 17 μM||23||–||–|
|Atorvastatin + mevalonate 600 μM||91||71||79|
|Atorvastatin + FPP||59||–||–|
|Atorvastatin + GGPP||76||–||–|
|Simvastatin + mevalonate 17 μM||10||–||–|
|Simvastatin + mevalonate 600 μM||78||84||85|
|Simvastatin + FPP||29||–||–|
|Simvastatin + GGPP||75||–||–|
|Fluvastatin + mevalonate 17 μM||55||–||–|
|Fluvastatin + mevalonate 600 μM||105||99||67|
|Fluvastatin + FPP||41||–||–|
|Fluvastatin + GGPP||62||–||–|
The effect of individual statins on K-Ras protein translocation
Since K-Ras proteins translocation from cytoplasm to the cell membrane is dependent on its farnesylation, we investigated the potential effect of statins on this cellular event in MiaPaCa-2 pancreatic cancer cells transfected with pEGFP-K-Ras (HeLa, wild type) plasmids.
As demonstrated in Figure 2, all the statins tested at the 20 μM concentration efficiently inhibited K-Ras protein trafficking from cytoplasm to the cell membrane. The only exception was pravastatin, which did not influence this process at all. Similar effect of statins on the localization of K-ras carrying the activation mutation (MiaPaCa-2 G12C) was observed. Again, pravastatin did not have any effect (data not shown).
Antitumor effects of individual statins on human pancreatic adenocarcinoma cell line CAPAN-2 xenotransplanted to nude mice
All of the statins used significantly prolonged survival in mice suffering from human pancreatic adenocarcinoma. While the median survival rate of placebo treated animals was 31.5 [29–32] days (median [25–75%]), the mice treated with statins survived significantly longer, depending on statin used (39–106.5 days, Fig. 3). Compared to placebo-treated animals, mice treated with pravastatin survived 39.5 [36–42] days (p = 0.01), while survival rates in the atorvastatin and simvastatin group were 39 [36–45] (p = 0.04) and 42 [39–44] days (p = 0.002), respectively. The highest survival rate was found in the mice treated with cerivastatin (106.5 [47–122] days, p = 0.002) and rosuvastatin groups (104.5 [65–122] days, p = 0.002). Substantial anticancer effects were also observed in a fluvastatin group (56 [42–246] days, p = 0.009), where 2 complete remissions were even recorded. Partial tumor regressions accompanied with central necrosis were also observed in several mice treated with cerivastatin and rosuvastatin. The significance of these results was also confirmed in a Kaplan–Meier survival analysis, demonstrating that the longest survivals in animals were those treated with rosuvastatin, cerivastatin, fluvastatin and simvastatin (Fig. 3). However, marked differences in the survival rate were also observed among these most effective statins. After detailed analysis of the survival data, significant differences among the most and the least effective statins were detected (Table III).
|Statin treatment||p value|
|Rosuvastatin > pravastatin||0.0005|
|Rosuvastatin > atorvastatin||0.002|
|Rosuvastatin > simvastatin||0.013|
|Rosuvastatin > lovastatin||0.042|
|Cerivastatin > pravastatin||0.0005|
|Cerivastatin > atorvastatin||0.006|
|Cerivastatin > simvastatin||0.005|
|Cerivastatin > lovastatin||0.065|
|Fluvastatin > pravastatin||0.025|
|Fluvastatin > atorvastatin||0.060|
|Fluvastatin > simvastatin||0.165|
|Fluvastatin > lovastatin||0.582|
The survival rate correlated well with the progression of tumor size. The increase in volume of CAPAN-2 tumors during treatment was significantly depressed in the statin-treated groups (Fig. 4a). Similar to the results of the survival analysis, significant differences were detected among the individual statins, with the best results found in fluvastatin and cerivastatin groups (Fig. 4b).
Compared to initial values of tumor size, treatment with cerivastatin and rosuvastatin did not lead to significant tumor progression, as measured by tumor size before death (p > 0.05, data not shown); while treatment with fluvastatin resulted in a subtle, but again insignificant, diminishing of tumors (due to 2 remissions observed in this group). In all other treatment groups, tumor progression was delayed but not prevented.
Routine histological examination of pancreatic cancer treated with saline and fluvastatin did not show any major difference in any of the examined samples, despite the fact that placebo-treated animals exhibited rapid progression of the tumors (data not shown).
Statins have been intensively studied drugs because of their deep impact on the human organism caused, in particular, by their cholesterol-lowering actions. However, their effects are pleiotropic and include potentially remarkable and clinically relevant antitumor effects. Indeed, the antiproliferative effects of statins were described for various experimental tumors, including pancreatic cancer.10–15, 32–35 The first study on the effect of statins upon pancreatic cancer was published as early as 1992 by Sumi et al.,10 who described the substantial antiproliferative effect of lovastatin on both poorly and well-differentiated human pancreatic cancer cell lines in vitro as well as in vivo. However, only extremely high doses of lovastatin (60 mg/kg body wt./day) were used in vivo.10 A similar effect on pancreatic cancer was also described for lovastatin by Mikulski et al.11 In studies by Kusama et al.,13, 14 performed upon 3 human pancreatic cancer cell lines, micromolar concentrations of fluvastatin and lovastatin inhibited EGF-induced translocations of RhoA and cancer cell invasion in a dose-dependent manner. These effects were reversed by the addition of all-trans-geranylgeraniol, but not by the addition of all-trans-farnesol, suggesting involvement of Rho activation, and not Ras signaling, in the process of cancer cell invasion. In another in vitro study on 5 pancreatic cancer cell lines by Sumi et al.,12 as well as in a study by Muller et al.,32 lovastatin inhibited the proliferation of pancreatic cancer, even in the absence of activating point mutations in the K-ras gene. In addition, the chemoadjuvant effect of statins on experimental pancreatic cancer has been described in several studies, including the beneficiary action of fluvastatin and gemcitabine34 or lovastatin and troglitazine, a PPAR-γ agonist.35
These data seem to be in contrast to recent observations on the inducing effects of statins on the heme oxygenase-1 (HO-1),36 a potent bioactive enzymatic system supposed by some authors to have procancerogenic role.37, 38 However, opposing, i.e., protective effects of HO-1 expression have been demonstrated by other authors for breast,39 liver40 and colon cancers.41 In addition, both pravastatin and simvastatin in a concentration range of 50–250 μM were not able to anyhow modulate HO-1 activity in CAPAN-2 pancreatic cancer cells (unpublished data, R. Motterlini and L. Vítek).
The current study is, according to the authors' knowledge, the first study to have compared differences in the effects of the individual commercially available statins upon the viability and growth of human pancreatic cancer in vitro and in vivo. All of the statins used, except pravastatin, substantially inhibited the growth of all 3 different pancreatic cell lines in vitro (Fig. 1, Table I). However, the intensity of this effect was dependent on the type of statin used, with marked differences among the individual compounds. These inhibitory effects were partially prevented by concomitant addition of mevalonic acid, FPP or GGPP, indicating the contribution of downstream intermediates in cholesterol biosynthesis for growth and viability of pancreatic cancer cells (Table II). The effect of GGPP can be related to the activation of another protein family than Ras, i.e. the Rho family (Rho, Rac and Cdc42), where geranylgeranylation is the predominant mechanism, whereas Ras proteins are mainly farnesylated. However, both K- and N-ras can be geranylgeranylated in the cells treated with inhibitors of farnesyltransferase.42
The most potent statin under in vitro conditions was cerivastatin, followed by simvastatin and lovastatin. Slightly less effective were fluvastatin and atorvastatin. The effect of these statins far exceeded the effects of rosuvastatin and pravastatin. However, these findings did not entirely parallel in vivo results, in which fluvastatin, cerivastatin and rosuvastatin had much better tumor-suppressive responses (Fig. 3). This is very interesting, especially for fluvastatin, if we consider that fluvastatin has the lowest hypocholesterolemic effects among the statins on the market.43 Marked differences in the impact of individual statins on cancer cell viability, with the worst effects found for pravastatin, may be due to many factors such as different chemical structures leading to changes in their pharmacokinetics and pharmacodynamics.26 Except for pravastatin and rosuvastatin, all statins are lipophilic substances which certainly affect their behavior and bioavailability for cells and tissues in the organism (where the situation is far more complex). In fact, lipophilic statins lovastatin and simvastatin were found to inhibit HMG-CoA reductase (also in the cells of peripheral origin), while pravastatin exhibited this action only in cultured hepatocytes.44 In contrast to pravastatin, rosuvastatin, the second hydrophilic statin, is able to inhibit HMG-CoA reductase in nonhepatic cells with 2 orders of magnitude higher efficacy.45 However, it should also be noted that the half-life elimination of rosuvastatin is ∼20 hr, compared to a half-life of all other statins between 2 and 3 hr. In addition, whereas lipophilic atorvastatin, fluvastatin, lovastatin and simvastatin were described to have direct proapoptotic effects,46, 47 pravastatin lacks this action.47 These data are also consistent with the lack of inhibitory effects of pravastatin on K-Ras protein translocation in pancreatic cancer cells, as compared to other statins. Interestingly, the effect of rosuvastatin on K-Ras translocation, which exhibits the highest IC50 value, was comparable to the most effective statins (Table I, Figs. 1 and 2).
Differences in the biological behavior of statins are further evidenced by their antioxidant effects, since it has been demonstrated that fluvastatin exerts 2 orders of magnitude higher antioxidant capacity (which obviously might be of clinical relevance) as compared to pravastatin.48 Although lactone prodrugs of lovastatin and simvastatin are ineffective for the inhibition for HMG-CoA reductase, these isomers are potent modulators of the 20S proteasome.49 However, it is important to emphasize that both (open and lactone) forms were identically active against tumor viability in our in vitro experiments. Other factors, such as the aforementioned different bioavailability, protein binding, conversion to metabolites, as well as elimination half-life26 may contribute to the observed diversity.
Our data are corroborated by an in vitro breast cancer study by Mueck et al.,25 who also reported different tumor-suppressive effects of various statins. Interestingly, in this study, fluvastatin and simvastatin were also the most effective statins, whereas pravastatin had no effect up to 50 μM concentrations.25 Negligible antitumor effects of pravastatin were also observed in other cancer studies,22–24 as well as no antiviral effect having been demonstrated for pravastatin compared to the strong antiviral activity on hepatitis C of the other tested statins.50
Genetic heterogeneity is another important factor which affects sensitivity of particular cancer cells to antiproliferative/proapoptotic effects of statins. This has been reported in a very recent paper by Wong et al.,51 who demonstrated that only half of the 17 multiple myeloma cell lines tested was sensitive to lovastatin-induced apoptosis, while resistant cell lines had different genetic profile. Moreover, specific adaptation mechanisms may lead to selection of variant resistant cancer cell clones as was shown for pravastatin-treated human breast and gastric cell lines.52 It is not clear whether other statins are more efficient in preventing development of statin resistance, but this may be further factor accounting for the weakest anticancer effects of pravastatin.
In vitro inhibitory concentrations of statins, similarly as in the majority of previous studies, may seem biologically irrelevant and not accounting for marked tumor-suppressive effects of much lower doses of statins in vivo. However, it must be stressed that in the human body statins may undergo biotransformation to become substantially more active inhibitors. All statins are metabolized in the liver where lovastatin, simvastatin, atorvastatin and cerivastatin share a common metabolic pathway through cytochrome P-450 3A4. Simvastatin is metabolized by hepatic CYP3A4 onto its dihydroxy acid form, whose antiproliferative effects were ∼1 order of magnitude higher, compared to simvastatin with IC50 (being 0.5–2.3 mg/l for 8 different pancreatic cancer cell lines tested).33 Similar biotransformation into active hydroxyl metabolites was also described for lovastatin, atorvastatin and fluvastatin,26 which is subjected to the action of cytochrome P-450 2C9. Pravastatin has multiple metabolic pathways. Since pancreatic cancer cell lines are not equipped with these biotransforming enzymes, it is not surprising that much higher doses must be used in vitro to mimic similar in vivo effects.
In support of the tumor-suppressive effects of statins, there are also the results of either randomized clinical trials or observational epidemiological studies, which were the subject of recent reviews and meta-analyses (for review, see Ref.2). Although several epidemiological studies have proved the beneficial effects of statins on cancer risk,2 these recent meta-analyses53–56 did not confirm cancer protection in statin users. This fact might be due to a too short follow-up period, as well as a methodology bias of the epidemiologic studies on cardiovascular outcomes, which were not designed or statistically powered to appropriately evaluate any beneficial or detrimental effects on a relatively rare event (such as a subsequent malignancy). This also has been recognized as a possible explanation for the slightly increased cancer incidence in elderly individuals taking pravastatin in the PROSPER study.57 Moreover, based on our in vitro and in vivo results, it seems that pravastatin has only negligible tumor-suppressive effects on pancreatic cancer. Interestingly, pravastatin also does not inhibit the proliferation of human breast cancer cells25 or gastric carcinoma cells in culture,58 but it does have beneficial effects on colon tumors in an animal model.51 This may account for the negative results of the meta-analytic studies mentioned earlier. In fact, in a meta-analysis of 7 trials by Hebert et al.,53 43% (3 out of 7) of studies involved were those with pravastatin. The same was true for a study by Bjerre and LeLorier54 (3 of 5 studies involved), CTT Collaborator's study55 (5 of 14 studies involved), as well as in the most recent meta-analysis by Dale et al.56 (10 of 20 studies). Interestingly, studies not demonstrating a drop of total cancer incidence or mortality also include data showing selective protective effects for certain types of cancer. This is true, for instance, for a large pravastatin clinical study CARE, in which a 43% decreased incidence of colorectal cancer was detected,59 consistent with the experimental data for pravastatin mentioned earlier.60 Similar results on the risk of colorectal cancer in patients mainly treated by pravastatin and simvastatin were also found in the study by Poynter et al.61
In conclusion, on the basis of our findings as well as results of other studies, we propose that the inhibitors of HMG-CoA reductase might be of potential for the chemoadjuvant treatment of pancreatic cancer. Indeed, promising results of a large, very recent study on the risk of pancreatic cancer in a population of almost half million US veterans support this.62
Fruitful and stimulating discussions on the mechanisms of statins, as well as material support, are greatly appreciated from Professor Henning Schröder from the University of Halle, Germany. Authors also thank Dr. Roberto Motterlini from Northwick Park Institute for Biomedical Research, UK, for his help with preliminary experiments on the effects of statins on heme oxygenase activity. The authors also thank Drs. Alžběta Kračmarová, Michaela Merkerová and Eva Michalová for their help with molecular biology studies.
- 2Cytoprotective and antiproliferative effects of HMG-CoA reductase inhibitors. Curr Enz Inhib 2006; 2: 261–80., .
- 4The role of p21ras in receptor tyrosine kinase signalling. Biochim Biphys Acta 1994; 1198: 131–47., .
- 45Pharmacokinetic and pharmacodynamic alterations of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors: drug–drug interactions and interindividual differences in transporter and metabolic enzyme functions. Pharmacol Ther 2006; 112: 71–105., .