Christelle Rodrigues [1], Najah Harouki [1], Julie Legrand [1], Niels J Skartved [2], Janus S Jakobsen [2], Camilla Frölich [2], Rikke Hald [2], Maria Almena-Carrasco [1], Fanny Cazade [1], Adrien Tessier [1]
[1] Servier, Gif-sur-Yvette, France, [2] Servier, Ballerup, Denmark
Objectives:
S095024 (Sym024) is a monoclonal Fc-effector function attenuated antibody inhibiting CD73. CD73 is a key enzyme involved in the conversion of extracellular AMP to adenosine, which via adenosine receptors can elicit immunosuppresion and tumour growth [1]. Increased CD73 expression has been observed in various types of cancer, including but not limited to colorectal cancer and lung cancer. The upregulation of CD73 is associated with tumour progression and evasion of immune responses in these malignancies [1]. The clinical safety, tolerability, pharmacokinetics (PK), and pharmacodynamics (PD) of S095024 alone or in combination with an anti-PD1 antibody was investigated in a first-in-human (FIH) study in patients with selected types of advanced solid tumours. The study was designed with a part I, assessing S095024 as single agent (100 to 1500 mg IV Q2W), and part II, assessing S095024 in combination (300 to 1500 mg IV Q2W). A part IIa was added assessing S095024 at 3000 mg as single agent in Cycle 1 and in combination from Cycle 2 onwards. The aim of this abstract is to explain the rationale behind the addition of the 3000 mg dose and highlight the role of modelling
Methods:
The determination of the initial doses for the FIH study integrated in vitro and in vivo study data and predicted PK in human. The human PK was predicted from a cynomolgus monkey population PK model via one-species allometric scaling [2]. An Emax relationship was used to simulate CD73 inhibition in the patient tumours, using predicted human serum concentrations and a EC50 of 1 μg/mL derived from in vitro (mean EC50 ≈ 1 µg/mL on primary human B cells and EC50 ≈ 1.8 µg/mL in cancer cells) and in vivo (EC50= 1.24 µg/mL in a NODscid mouse model (no immune cells) with A375 human xenograft tumours) S095024 inhibition data. Additionally, a tumour growth inhibition (TGI) model developed with NODscid mice data [3] was also used to derive the exposure needed for pharmacological activity in human.
New in vivo experiments were performed with NOG mice (with human immune cells). Predictions of CD73 inhibition in patient tumours were refined using the EC50 of 22.4 µg/mL returned from this mouse strain. The TGI model was also refined using the PK/PD observed in NOG mice.
Results:
Early predictions of CD73 activity, using the EC50 obtained in NODscid mice, indicated near to full inhibition with a 900 mg Q2W dosing regimen of S095024. Additionally, the initial PK/PD relationship based on NODscid mice data indicated that an average systemic plasma concentration of 300 μg/mL was needed for inhibiting tumour growth, corresponding to an expected dose of 900-1500 mg Q2W in humans. Based on these PK/PD simulations, doses ranging from 100 mg to 1500 mg administered every two weeks (Q2W) were selected for the FIH study. The dose of 900 mg was predicted to meet the defined criteria of full CD73 inhibition and thus anticipated to be the recommended phase 2 dose (RP2D).
During the FIH study, CD73 enzymatic activity was measured in patient tumours, at baseline and end of Cycle 1, to quantify the degree of inhibition induced by S095024 treatment. Preliminary results in patients (up to 900 mg) did not meet the predefined criteria of full CD73 inhibition. The two PK/PD models were then refined using new preclinical data from NOG mice. Using the higher EC50 obtained in this strain, lower levels of inhibition were obtained leading to the need of a 3000 mg Q2W dosing regimen to obtain near to full inhibition of CD73 activity. Regarding the TGI model using the PKPD relationship observed in NOG mice, an average systemic plasma concentration of 1500 μg/mL in human was needed for tumour growth inhibition.
These results imply that a higher dose than initially predicted may be necessary to attain the full pharmacological effect, thereby prompting the inclusion of an additional cohort at 3000 mg Q2W.
Conclusions:
The change from NODscid to NOG mice for EC50 determination in the PK/PD model significantly influenced the predicted human CD73 activity inhibition profiles, probably due to different tumour micro-environments in the two models. This finding highlights the importance of addressing different preclinical models to optimise predictions of clinically active doses. In this case, modelling of new preclinical data led to the addition of a new cohort at 3000 mg Q2W, instilling more confidence in later selecting the RP2D.
References:
[1] Bach N, Winzer R, Tolosa E, Fiedler W, Brauneck F. The Clinical Significance of CD73 in Cancer. Int J Mol Sci. 2023 Jul 21;24(14):11759
[2] Dong JQ, Salinger DH, Endres CJ, et al. Quantitative prediction of human pharmacokinetics for monoclonal antibodies: retrospective analysis of monkey as a single species for first-in-human prediction. Clin Pharmacokinet. 201; 50:131-42.
[3] Simeoni, M. P. (2004). Predictive Pharmacokinetic-Pharmacodynamic Modeling of
Tumor Growth Kinetics in Xenograft Models after Administration of Anticancer Agents.
Cancer Research, 64:1094-1101.
Reference: PAGE 32 (2024) Abstr 10910 [www.page-meeting.org/?abstract=10910]
Poster: Drug/Disease Modelling - Oncology