"The encouraging results from the Phase 2 clinical trials with GW501516 have been shadowed by the preclinical toxicology findings. There are several conflicting studies showing the direct effects of oral dosing rats and mice of GW501516 leading to either the induction or in- hibition of carcinoma (Table 3). Looking closely at the experimental de- sign it seems that there is a direct correlation between dose and outcome, with doses of GW501516 10 mg/kg/day for 6 weeks inducing detrimental side effects such as rectal bleeding in mice (Gupta et al., 2004) and at 10 to 100 mg/kg/day for 104 weeks in rats and mice devel- oping tumours throughout the body (Geiger et al., 2009; Newsholme et al., 2009) and 30 to 275 mg/kg leading to placental malformation in rats (Nishimura et al., 2013). In these animals, this high dosing would lead to agonists reaching the μM range, and therefore the non- genomic and off target effects of PPARβ/δ agonists become apparent. On the other hand, similar or lower doses of GW501516 (2 to 10 mg/ kg for 5 weeks) lead to the inhibition of chemically induced carcinomas in mice when dosed for up to 22 weeks (Marin et al., 2006; Hollingshead et al., 2008), and inhibit growth of human cancer cell lines (Bility et al., 2008). Mice orally dosed with 4 mg/kg/day become more metabolically active (Lee et al., 2006); at 2 mg/kg/day mice, re- cover from chemically induced non-alcoholic steatosis (Nagasawa et al., 2006) and ApoE−/− mice exhibit significantly reduced lesions (Barish et al., 2008). A study of septic shock in mice used a single injec- tion of 0.03 mg/kg GW0742, whose beneficial effects were completely reversed by a selective PPARβ/δ antagonist GW0660 (Kapoor et al., 2010). This study did not measure any changes in cholesterol, although it does clearly indicate that far smaller doses of these compounds may provide us with more PPARβ/δ selective data."
