PI-102 - PREDICTION OF A CLAUDIN 18.2 TARGETED ANTIBODY DRUG CONJUGATE (ADC) PHARMACOKINETICS IN CANCER PATIENTS USING PBPK MODELING AND SIMULATION

C. ZUNINO¹, S. Wang², S. URDY³, Y. Zhang⁴, X. DECLEVES³, A. Puszkiel³, N. DJEBLI⁵; ¹Phinc Development, Massy France, Phinc Development, 36 rue Victor Basch, 91300 Massy, France, ²Jiangsu Hengrui Pharmaceuticals, Shanghai, China, Jiangsu Hengrui Pharmaceutical, Shanghai, China, ³Université Paris Cité, Inserm UMRS1144, Cochin Hospital, Université Paris Cité, Inserm UMRS1144, Cochin Hospital, 75014 Paris, France, ⁴Jiangsu Hengrui Pharmaceuticals, Shanghai, China, Jiangsu Hengrui Pharmaceuticals - Shanghai,, China, ⁵Luzsana / Hengrui, Luzsana Biotechnology (subsidiary of Jiangsu Hengrui Pharmaceutical), Basel, Switzerland.

Background: ADCs represent a promising anticancer drug modality. Although physiologically based pharmacokinetics (PBPK) modeling has become an essential approach in Pharmacometrics to mechanistically characterize exposure in different tissues, very few PBPK models have been published so far for ADCs, none within the PK-Sim® / MoBi® software.

Methods: To capture the PK of this anti-Claudin 18.2 ADC, a human PBPK model was built in PK-Sim® and MoBi® and the predictions were compared to observations from 3 clinical studies after IV administration in 109 patients with gastric or pancreatic cancers (including 908 and 824 plasma observations for ADC and toxophore, respectively). The PK parameters were considered accurate if the predicted error ratios were within the two-fold error range (0.5-2).

Results: In PK-Sim®, we defined one PBPK model for each compound (ADC, toxophore and naked antibody), with input parameters either measured in vitro or predicted from the compound structure. While this model captured the ADC PK profile following IV administration of multiple escalating doses, additional clearance mechanisms were essential to improve the fit of the ADC elimination phase. After integration of target-mediated drug disposition (TMDD), nonspecific catabolism and deconjugation of the toxophore in MoBi®, 3 and 4 parameters were optimized for the ADC and the toxophore, respectively (degradation rate constant and reference concentration of the target, deconjugation rate constant, lipophilicity, nonspecific hepatic clearance, passive renal clearance of the toxophore). The PK data was adequately captured for both ADC and toxophore (Fig. for the 6 mg/kg dose as an example), with predicted error ratio included within the two-fold range: Cmax_ADC (1.07-1.50), Cmax_Toxophore (0.69-1.44), AUC0-504h_ADC (0.59-0.98) and AUC0-504h_toxophore (0.82-1.38).

Conclusion: The “parameter optimization” of parameters such as those driving TMDD and deconjugation elimination pathways allowed to accurately capture the observed data for both ADC and toxophore in cancer patients for the anti-Claudin 18.2 ADC. This analysis paves the way for PBPK modeling of other ADCs currently in development.