Structural model development with compartmental pharmacokinetic models for a range of small molecule subcutaneous data
Asina Gijasi (1,2), Jeroen Elassaiss-Schaap (1)
(1) Gedmore B.V., the Netherlands, (2) Leiden University, the Netherlands
Objectives: The development of pharmacokinetic (PK) and pharmacodynamic (PD) models demands considerable time from highly skilled scientists, and the need for such expertise exceeds the current availability in the pharmaceutical industry [1]. Automated model development thus represents an appealing opportunity. A good alternative to the published unsupervised global model space search using a hybrid genetic algorithm is a staged and supervised approach [2]. In this regard, the building block for an automated model development system for the approach of a structural pharmacokinetic model search has already been implemented [3]. Subcutaneous administration is a common route for many drugs including insulin, heparin and monoclonal antibodies [4], but the automated model development system, Gedmore, has yet to be expanded and validated for subcutaneous PK data. Therefore, we at Gedmore, have expanded the automated model PK development system to subcutaneous dosing PK data.
Objectives:
(1) Develop compartmental PK models for subcutaneous dosing PK data from literature.
(2) Examine whether the simple models are good enough to describe concentration data.
Methods: To evaluate subcutaneous dosing PK, a literature search was performed to collect data. Subcutaneously administered median concentration data from single and multiple dose studies in either female or male animals was collected. Data was restricted to small-molecule drugs and linear PK as the current Gedmore system strictly estimates linear PK models and to prevent loss of exposure due to target-mediated drug disposition (TMDD) or the presence of anti-drug antibodies (ADAs). For each data set, an optimal PK model was developed using non-linear mixed effects (NLME) modeling using NONMEM (v 7.5.1, ICON Development Solutions, Maryland). Model goodness-of-fit (GOF) was determined using differences in objective function value (OFV), akaike information criteria (AIC) and GOF plots. Lastly, the resulting models were then also visually compared to population PK models for subcutaneously administered drugs in animals to examine whether the simple structural models are good enough to describe the concentration data.
Results: Mean or median preclinical subcutaneously administered concentration data was collected and for the 5 literature PK studies retrieved, an optimal structural PK model was developed that described the data adequately. The optimal models appeared to be 1- or 2-compartmental PK models with a combination of zero and/or first order absorption with reasonable residuals between model and data. Comparison of the developed optimal PK models with a published preclinical population PK model [5] for a subcutaneously administered drug demonstrated concordance, highlighting the capability of the simple structural models to accurately describe subcutaneous concentration data.
Conclusions: Subcutaneous dosing PK data was described using straightforward mammillary compartmental models featuring zero order with or without first order absorption. To strengthen this finding, more data should be collected and diversified to explore the capability for a wider range of experimental conditions, including multiple dosing data. The model development trajectories will thereafter be implement in a subcutaneous administration procedure in Gedmore.
References:
[1] Z. Huang, P. Denti, H. Mistry, and F. Kloprogge, “Machine Learning and Artificial Intelligence in PK‐PD Modeling: Fad, Friend, or Foe?,” Clinical Pharmacology & Therapeutics, Jan. 2024, doi: https://doi.org/10.1002/cpt.3165.
[2] M. Sale and E. A. Sherer, “A genetic algorithm based global search strategy for population pharmacokinetic/pharmacodynamic model selection,” British Journal of Clinical Pharmacology, vol. 79, no. 1, pp. 28–39, Dec. 2014, doi: https://doi.org/10.1111/bcp.12179.
[3] PAGE 16 (2007) Abstr 1188 [www.page-meeting.org/?abstract=1188]
[4] I. Usach, R. Martinez, T. Festini, and J.-E. Peris, “Subcutaneous Injection of Drugs: Literature Review of Factors Influencing Pain Sensation at the Injection Site,” Advances in Therapy, vol. 36, no. 11, pp. 2986–2996, Oct. 2019, doi: https://doi.org/10.1007/s12325-019-01101-6.
[5] Juan Manuel Serrano-Rodríguez et al., “Population pharmacokinetics and pharmacokinetic/pharmacodynamic evaluation of marbofloxacin against Coagulase-negative staphylococci, Staphylococcus aureus and Mycoplasma agalactiae pathogens in goats,” Research in Veterinary Science, vol. 159, pp. 1–10, Jun. 2023, doi: https://doi.org/10.1016/j.rvsc.2023.03.026.