Introduction The Autotaxin (ATX)/ Lysophosphatidic acid (LPA) axis has triggered a lot of interest in IPF due to a positive phase 2 proof of concept with GLPG1690, a potent ATX inhibitor in the FLORA study. In animal models GLPG1690 displayed a strong activity in the bleomycin (BLM) lung fibrosis mouse model, however, G889813, another potent ATX inhibitor from a distinct chemical series was found inactive in this model despite high potency at reducing LPA levels in blood. Aim and objectives To compare the activity of GLPG1690 and G889813 (two potent ATX inhibitors with different binding modes) in vitro and in vivo in a mouse BLM model. Methods Biochemical potency was assessed using human recombinant ATX. Ex vivo potency was determined in plasma by measuring LPA levels (LC-MS) after incubation with ATX inhibitors. Lung exposure was assessed by measuring compound levels by (LC-MS) after oral administration. Lung weight, Ashcroft score and hydroxyproline were used as a measure of efficacy in the BLM model. The binding modes of GLPG1690 and G889813 was derived from co-crystal structures with human or rat ATX. Results In vitro, GLPG1690 and G889813 were shown to be potent inhibitors of human ATX (IC50 values 131 nM and 6 nM respectively). In a rat plasma assay GLPG1690 and G889813 displayed high potency (IC50 values 541 nM and 13nM respectively). In vivo GLPG1690 was shown to be significantly active at 3mg/kg bid as assessed by Ashcroft score whereas G889813 was found inactive at the same dose. Lung exposure was measured at a dose of 10mg/kg, an AUC of 1427 ng.h/ml and 6123 ng.h/ml was obtained for GLPG1690 and G889813 respectively. GLPG1690 was classified as a type IV inhibitor occupying both the hydrophobic pocket and the allosteric tunnel and G889813 was classified as type I inhibitor occupying the hydrophobic pocket and the catalytic site.Conclusion GLPG1690, displayed high efficacy in a preclinical model and surprisingly G889813, a 30 fold more potent compound and well exposed in the lung was shown inactive in BLM model. It is postulated that ATX, independently of its LPA synthesis activity, can trap exogenous LPA within the allosteric tunnel, protect it from degradation and serve as a shuttle to deliver it to its cognate receptor. If physiologically relevant in lung, this model could explain the superior in vivo efficacy observed with GLPG1690 and suggest that type IV ATX inhibitors may be needed to treat lung fibrosis.