PERSONALISING ANTICANCER DRUG CIRCADIAN TIMING
Recent preclinical and clinical studies highlighted large inter-patient variabilities regarding circadian rhythms of the organism and of drug response. Those individual variations sugget that a high patient benefit may be obtained by personalizing drug circadian timing, especially in the field of cancer as antitumor agents are often responsible of severe toxicities. We design combined experimental, clinical and mathematical approaches to individualize anticancer drug circadian timing towards and improved treatment tolerability and antitumor efficacy.
MULTI-SCALE APPROACH TO PERSONALISE IRINOTECAN CIRCADIAN TIMING
Irinotecan is an anticancer drug currently used in the clinics against digestive cancers. The drug pharmacology and its control by the circadian timing system was studied in cell culture and in mice to investigate molecular determinants of irinotecan chronotoxicity rhtyhmes towards its circadian timing optimization.

An in vitro study of irinotecan chronopharmacology led to the design of a cellular chronoPK-PD model (A) incorporating multitype experimental data, including the extra- and intracellular concentrations of active metabolite SN 38 and irinotecan-induced apoptosis after irinotecan exposure at three CTs (B–D) Dulong et al., 2015). This cellular investigation provided the basis for a mouse study and the development of a whole-body model of irinotecan chronoPK-PD explicitly incorporating the cellular model in relevant organs (E) (Ballesta et al., 2012). The model was first developed for B6D2F1 male mice in which several chronopharmacology datasets were available, including plasma and colon chronoPK profiles of SN38 after irinotecan at best and worst time of tolerability (F–G). The next step will consist in fitting intestinal chronotoxicity data available for the same mouse category (H) (Li et al., 2013). Dots or bars represent experimental results, and solid lines represent best-fit models.
THERAPEUTIC OPTIMISATION
Mathematical models which are calibrated to experimental data can now be utilized in optimization procedures in order to determine the best possible circadian manner to administer the drug.

Optimisation of irinotecan chronotherapy. A. Model-predicted chronotoxicity of irinotecan in Caco-2 cell culture according to exposure duration and circadian time (CT) of exposure start. B. Optimal irinotecan exposure schemes for various toxicity threshold: we considered a well-synchronised cell population representing healthy cells and a cell population without circadian rhythms representing cancer cells. Optimization procedures consisted in finding the drug exposure scheme that would maximize the damage on cancer cells while damage in healthy cells stay under a toxicity threshold (y-axis). All Model-computed optimal schemes target the same circadian time window between CT2 and CT9. The maximum allowed dose as predicted by the model (green numbers) logically increases with the toxicity threschold.