A physiologically-based pharmacokinetic/pharmacodynamic framework to support clinical development of V937, a novel oncolytic virus
Zinnia P Parra-Guillen(1,2), Iñaki F Trocóniz(1,2), Sara Zalba(1,2), María J Garrido(1,2), Kapil Mayawala(3), Dinesh de Alwis(3), Tomoko Freshwater(3)
(1)Pharmacometrics & Systems Pharmacology, Department of Pharmaceutical Technology and Chemistry, School of Pharmacy and Nutrition, University of Navarra, Pamplona, Spain; (2)IdiSNA, Navarra Institute for Health Research, Spain; (3) Quantitative Pharmacology and Pharmacometrics Immune/Oncology, Global Clinical Development, Merck & Co, Inc, Kenilworth, New Jersey, USA
Objectives:
Immuno-oncology (IO), based on the activation of the endogenous anti-tumor immune response, is a growing strategy in cancer treatment. Oncolytic viruses (OVs), which rely on the ability to selectively infect and replicate in cancer cells leading to direct and/or immune dependent tumor lysis, represents a potential therapeutic strategy [1]. Currently, there is a lack of comprehensive quantitative models characterizing viral kinetics and distribution to the tumor for the aforementioned treatment scenarios. Our work aims to develop a mechanistic modelling framework for the novel oncolytic virus V937, which integrates viral kinetics/replication, distribution to tumor and tumor response after intratumoral (IT) or intravascular (IV) administration in cancer patients.
Methods:
To build the model, circulating V937 mRNA (n=798) as well as individual tumor size grouped by injected (n=165) or non-injected (n=73) tumor lesions from phase 1/2 clinical trials were used [2,3]. Patients with missing tumor information or receiving doses in non-measured lesions were excluded. In total, data from (i) 18 cancer patients receiving a dose of 1x108, 3x108 or 10x108 TCID50 (half-maximal tissue culture infectious dose) of V937 as a 30 min IV infusion on days 1, 3, 5 of cycle 1 and on day 1 of the following 21-day cycles and (ii) 31 cancer patients receiving a IT dose of 3x108 TCID50 on days 1, 3, 5, 8 and 22 of cycle 1, followed by an administration at 21-day intervals were analyzed.
The model consists of 5 compartments: plasma, peripheral, injected tumor, non-injected tumor and lymph. Both tumor compartments are further divided into vascular, interstitial and cellular sub-compartments. Viral biodistribution is represented using organ volumes, and physiological processes, e.g., blood and lymph flows taken from a previously published model [4]. Viral dynamics at the tumor level were incorporated; following uptake in the tumor cell, driven by interstitial viral levels in the tumor compartment, the virus could replicate, elicit a direct tumor cell killing and/or get released upon cell death [5]. In addition, an indirect killing (immune) effect activated by the death of the infected immune cells (iTcells) was explored. Cell proliferation of uninfected tumor cells was also included to describe tumor growth. SimBiology v5.10 and NONMEM 7.4 were used for model development and simulation.
Results:
An adequate description of circulating viral levels was obtained estimating a high serum viral clearance (ca. 30 L/h) and a large distribution to the peripheral compartment. In the absence of viral replication, tumor interstitial viral levels could be up to 4 orders of magnitude lower after IV administration compared to IT. Those differences in exposure after IV and IT administration can be overcome if infectivity and/or replication takes place to a sufficient degree. Two different killing mechanisms were identified: a direct killing triggered by the death of tumor cells infected by the virus (kdeath=0.0012 day-1) and an indirect killing induced by the immune response that acted on both injected and non-injected lesions (KCD8 = 9.93 x 10-14 day-1cell-1). The infectivity rate constant was estimated to be consistent with previous in vitro experiments (b= 2.35x10-11 L/copies/day) [6] and thus fixed for model stability, while current data did not support the inclusion of viral replication at tumor level, therefore it was fixed to 0. In this regard, viral levels in tumor (not available in this current study) would be needed to well characterize this replication process.
The final model described the observed systemic viral PCR levels and tumor size profiles well by accounting for variability in tumor growth and infectivity processes.
Conclusions:
This work represents the first modelling framework to successfully describe clinical circulating levels of V937 and the tumor size profile after both, IV and IT administration. The platform developed here, that mechanistically accounts for the physiological viral distribution to the tumor environment, viral dynamics at the tumor level and the triggered immune response, can be used to: (i) investigate the contribution of the different processes on the observed outcome and (ii) formulate new hypotheses and suggest experimental protocols to increase our understanding of the system. Upon further validation, this model can additionally be employed to inform clinical drug development.
References:
[1] Kaufman et al. Nat Rev Drug Discov. 14:642-662 (2015).
[2] Shah et al. J Pharmacokinet Pharmacodyn. 39: 67-86 (2012).
[3] Titze et al. Eur J Pharm Sci. 97:38-46 (2017).
[4] Pandha et al. Cancer Res 2017;77(13 Suppl):Abstract nr CT115.
[5] Antbacka et al. J Clin Oncol. 39: 3829-3838 (2021).
[6] Parra-Guillen et al. Front Pharmacol. 23:705443 (2021).