393P - Pharmacodynamic (PD) biomarkers for the p70S6K/Akt inhibitor, M2698: translation from animal to human and relevance to dose selection

Date 11 September 2017
Event ESMO 2017 Congress
Session Poster display session
Topics Clinical research
Basic Scientific Principles
Presenter Wenyuan Xiong
Citation Annals of Oncology (2017) 28 (suppl_5): v122-v141. 10.1093/annonc/mdx367
Authors W. Xiong1, H. Tian2, A. Clark2, J. Shaw2, R. Kaleta3, I. Celik4, P. Girard2
  • 1Merck Institute For Pharmacometrics, Merck Serono S.A., 1015 - Lausanne/CH
  • 2Global Early Development, EMD Serono Research and Development Institute, Billerica/US
  • 3Clinical Oncology, EMD Serono Inc., Billerica/US
  • 4Global Research And Development, Merck KGaA, Darmstadt/DE



M2698 is an oral, potent and selective inhibitor of p70S6K/Akt1/3 in the PAM pathway with the potential to block signaling from the Akt feedback loop and overcome tumor resistance. Preclinical studies suggest a steep exposure – adverse event relationship. Therefore, insights into drug-dependent target modulation (S6 phosphorylation (pS6) inhibition) and its association with efficacy could inform dose selection.


Pharmacokinetic (PK) data from the phase I, first-in-human (FIH) dose escalation trial conducted in patients with advanced cancer who received daily (d) oral M2698 (15-320mg/d) were evaluated by nonlinear, mixed effect modeling. Using a PK/PD model developed with data from a breast tumor cell line derived xenograft (CLDX) in mice, tumor pS6 time profiles in humans were simulated using a mouse PD model driven by human PK. Model predictions were calibrated by comparing simulations to clinical observations, with an assumed variation in sensitivity (IC50) to pS6 between CLDX and human tumors. Predicted pS6 time profiles and clinically observed pS6 inhibition in human tumors and peripheral blood mononuclear cells (PBMC) informed the dose escalation decision in the FIH study.


M2698 PK profiles were best described by a two-compartment linear model with transit compartments for delayed absorption. Consistent exposure-dependent effects of pS6 inhibition in PBMCs were observed at 160-320 mg/d, with 70-80% pS6 inhibition observed in some tumors. Based on predicted pS6 inhibition, CLDX tumors were 2-3x more sensitive than human tumors. Applying a 2-3x higher IC50, simulations suggested that 250-350mg/d would achieve the PD threshold of continuous tumor pS6 inhibition ≥ 80% in 90% of a human population, leading to escalation to 380mg/d.


Understanding inter-species variation can improve the precision of preclinical-to-clinical translation. For M2698, preliminary clinical PD data showed that human tumors were 2-3x less sensitive to pS6 inhibition than CLDX tumors in mice. Collective PMBC and tumor PD outcomes suggest that 160 -320 mg/d M2698 may result in considerable pS6 inhibition; a range for selection of a phase 2 dose.

Clinical trial identification


Legal entity responsible for the study

Merck KGaA


Merck KGaA


W. Xiong, I. Celik, P. Girard: Merck employee. H. Tian, A. Clark, J. Shaw, R. Kaleta: EMD Serono employee