Background: Irinotecan hydrochloride and its active metabolite (SN38) function as topoisomerase I inhibitors, with the active metabolite being 100-1000 times more potent. Although irinotecan is widely used as a first-line therapy in metastatic colorectal cancer (mCRC), significant interindividual pharmacokinetic (PK) variability and dose-limiting toxicities restrict its application. We have developed a novel lipid-drug conjugate of SN38 (SN38-undecanoate) as an oral therapy for mCRC. SN38-undecanoate is a fatty acid ester prodrug of SN38, wherein the C20 hydroxyl-group is conjugated with undecanoic acid. PK characterisation of the prodrug-metabolite in an animal model supports optimisation of treatment regimens for the next steps of drug development.
Objectives: To characterise the PK behaviour of SN38-undecanoate and SN38 in plasma following oral administration of the prodrug in three different formulations.
Methods: Twelve rats (three per group) received a single dose of either 1 mg/kg intravenous (IV) SN38 or 10 mg/kg oral SN38-undecanoate in three different formulations (liquid self-microemulsifying drug delivery system (SMEDDS), solid lipid carrier system (SLC), and lipid emulsion). Plasma concentrations of parent and metabolite were quantified using a simultaneous liquid chromatography-tandem mass spectrometry (LC-MS/MS) method. Noncompartmental analysis (NCA) and population PK (popPK) modelling of all concentration data were performed in Phoenix WinNonlin and NLME (version 8.4). Models of various complexities were fitted using the Quasi-Random Parametric Expectation Maximization (QRPEM) engine. The final model was selected by evaluating the goodness-of-fit diagnostic plots and the objective function values.
Results: The SMEDDS and SLC formulations provided higher exposures to SN38-undecanoate (AUC0-48h, geometric mean [SD] values 550 [1.07] and 574 [2.58] nmol.h/L) and SN38 (414 [1.08] and 403 [1.88] nmol.h/L) compared to the emulsion formulation (327 [1.18] and 167 [1.20] nmol.h/L). The metabolic ratio, defined as the percentage of the SN38 AUC divided by the prodrug AUC, for the oral formulations was considerably higher compared to intravenous irinotecan (>50% vs. <5%). Comparing the oral and IV curves, a flip-flop phenomenon was observed for SN38, indicating a slow systemic release of the active metabolite following absorption and conversion of the prodrug. Fitting IV and SMEDDS data together, a model with three and two disposition compartments for SN38-undecanoate and SN38, respectively, and linear elimination clearance of SN38-undecanoate (29.6 L/kg.h [between subject variability, 2%]) and SN38 (2.02 L/kg.h [50%]) had good predictive performance. Similar first-order rate constants for conversion of SN38-undecanoate to SN38 were estimated in the central (0.00668 1/h [3%]) and the first peripheral (0.00644 1/h [4%]) compartments. With the SLC formulation, a second rise in parent and metabolite levels was observed (Tmax≥ 10h). This was explained by splitting the prodrug dose into immediately absorbed (FI= 0.259 [68%]) and lagged absorbed (FLag= 0.741 (1-FI), Tag= 9.91 h [4%]) components, attributable to different release and absorption mechanisms.
Conclusions: The plasma PK of SN38-undecanoate and SN38 were adequately described by a compartmental model with linear elimination clearance and prodrug conversion. Preclinical NCA and model estimates can be used to set a first-in-human starting dose for this novel oral lipophilic prodrug with the aim of achieving an overall SN38 systemic exposure equivalent to that of irinotecan. The sustained systemic release of the active metabolite from oral SN38-undecanoate, compared to intravenous irinotecan, would likely reduce the impact of interindividual PK variability. Additionally, the direct targeting of mesenteric lymphatics, facilitated by its highly lipophilic nature, could enhance the efficacy-toxicity balance in therapy for mCRC patients.