Preclinical assessment of anthracycline cardiotoxicity in laboratory animals: predictiveness and pitfalls.

Doxorubicin is one of the most prescribed anticancer drugs, due to its important activity in hematological malignancies as in solid tumors. However, its important cardiac toxicity still limits its long-term use and prevents from reaching optimal benefits. Numerous ways have been proposed to avoid cardiac toxicity, such as protracted infusions or special formulations, development of less cardiotoxic analogues and of cardioprotectors. There is a need for preclinical models able to screen rapidly these various approaches and to provide rational bases for clinical trials. The first model is the long-term rabbit model. Weanling rabbits given weekly injections of doxorubicin for 4 months developed a cardiomyopathy which was obvious from a clinical (cardiac failure) and from a pathological point of view. This model has been widely used afterwards for the discovery of cardioprotective molecules. Models in other animals such as rats or mice were similarly implemented, also with long-term exposures to the drug, resulting in cardiac failure and severe pathological alterations which could be graded for comparison. Starting from the evidence that the damage caused by anthracyclines on cardiomyocytes was immediate after each injection and that the functional efficiency of the myocardium should be affected by the anthracyclines long before the morphological alterations become detectable, we developed a short-term model studying the cardiac performances of isolated perfused hearts of rats that had been treated within 12 days by repetitive administrations of the molecule(s) to be tested. This model appeared easy to implement and provided the data expected from clinical experience: epirubicin appeared less cardiotoxic than doxorubicin; liposomal formulations appeared less cardiotoxic than free drug formulations; dexrazoxane strongly protected against doxorubicin cardiotoxicity. We were then to show that paclitaxel could potentiate doxorubicin cardiotoxicity, but that docetaxel did not so; or that a high dose of dexrazoxane brought significantly higher protection than a conventional dose. Based upon these various contributions, we can encourage the use of the short-term model of isolated perfused rat heart to screen the preclinical cardiotoxicity of anthracycline molecules, formulations and combinations.