What is a relevant statin concentration in cell experiments claiming pleiotropic effects?

Abstract

Apart from their lipid-lowering effects statins have been ascribed a number of other functions including anti-inflammatory and immunomodulatory properties 1. Notably, most available evidence for the proposed pleiotropic effects of statins is based on in vitro studies. We claim that the results from these in vitro studies are potentially misleading, because of the high statin concentrations used. We performed a systematic literature search identifying in vitro experiments where mechanistic insights behind pleiotropic effects of statins are claimed. Statin concentrations in these cell experiments were compared with concentrations detected in human plasma. We focused on studies published in journals with an impact factor >15, as these reports were likely to have a broad impact on this field of research. Indeed, the choice of statin concentration is frequently based on previous experiments from papers in high-impact journals. In these articles, pleiotropic effects of statins were only detected at statin concentrations of 1–50 µmol L−1 (e.g. 2-6). In contrast, the mean concentration of statins in human serum (at therapeutic doses) is only 1–15 nmol L−1 (Table 1). In addition, the protein binding of statins in human blood is high, 95–99%, and it is only the free fraction (0.01–0.5 nmol L−1) that is pharmacologically active. Compared with the mean statin concentration in human blood, the peak concentration (Cmax) is 2–6 times higher (Table 1), but this concentration is only present in serum for a very short time (minutes) while the in vitro incubations have lasted for up to 4 days (Table 1). The large discrepancy between statin concentrations employed in cell experiments and those detected in human plasma is remarkable and, in our opinion, unjustified. The potential problems associated with using too high concentrations in vitro can be illustrated by data from our own laboratory where we have found that simvastatin inhibits growth of pneumococci with a MIC value of 15 µg ml−1 (36 µmol L−1; unpublished data). This MIC value is approximately in the same range as for other antibiotics (tetracycline MIC = 1 µg ml−1, vancomycin MIC = 2 µg ml−1). Hence, one could make the assumption that statins may be used as antibiotic drugs. However, when the true statin concentration in vivo (>1000-fold lower) is considered, statins no longer appear attractive as anti-infective agents. A possible rationale for using 1000-fold higher concentrations of statins in vitro than in vivo could be if the drug accumulated in the target organ. Indeed, one study demonstrated statin concentrations in liver tissue exceeding those in serum, but the difference was only twofold 7. In contrast, the concentration in brain and muscle tissue is only one-third of that in serum 7, 8. The intracellular concentration of statins in endothelial, blood and immunologically active cells is unknown. What statin doses have been used in animal experiments? In humans the dosage of statins varies between approximately 0.1–1 mg kg−1 bodyweight, while most studies in rodents have used doses of 1–100 or even 500 mg kg−1 body-weight (e.g. 3, 7, 8). These doses result in concentrations much higher than those achievable in patients 7, better resembling the exposure used in vitro. The use of very high doses in rodents is probably possible because of a pharmacodynamic resistance to the pharmacological effect in these animals, e.g. statins do not lower cholesterol concentrations in serum of rodents even though they significantly inhibit the mevalonate pathway in the liver 9. We fully acknowledge that more clinically relevant doses of statins are used in some animal experiments. However, given that the metabolism and actions of statins differ between humans and rodents, animal experiments claiming that statins are anti-inflammatory should be interpreted with caution. In conclusion, we believe that a pleiotropic effect of statins is an interesting area that deserves studies on underlying mechanisms both in vivo and in vitro. However, we do call for an active discussion about relevant statin concentrations in these experiments. If concentrations much higher than those achieved in therapeutic use are chosen, the rationale for this approach should be presented by the authors. There are no competing interests to declare.