Integration of the LeiCNS-PK3.0 PBPK framework and multiscale data to select anti-COVID19 drugs with favorable brain pharmacokinetics
Mohammed A. A. Saleh (1), Makoto Hirasawa (1), Ming Sun (1), Berfin Gülave (1), Jeroen Elassaiss-Schaap (1,2), Elizabeth C. M. de Lange (1)
(1) Systems Pharmacology and Pharmacy, LACDR, the Netherlands; (2) PD-value, the Netherlands
Objectives
Increasing concerns about the CNS involvement in COVID-19 have been raised by the associated neurological manifestations [1–3]. SARS-CoV-2 has also been shown to infect the brain cells, with viral RNA detectable for up to 230 days [4]. While a causal relationship between the viral neurotropism and the CNS manifestations is still unestablished, yet the reduction of viral load in brain is crucial to avoid a latent viral state and recurrent encephalitis.
Remdesivir, Molnupiravir, and Nirmatrelvir have been approved for COVID-19 treatment. Information on their CNS distribution is however lacking, particularly at the CNS target sites in the brain extracellular (brainECF) and intracellular (brainICF) fluids.
Using LeiCNS-PK3.0 physiologically based PK (PBPK) framework, we have previously demonstrated adequate predictions of the unbound PK of the human CNS, including that of brainECF, within a two-fold errors limit [5]. LeiCNS-PK3.0 integrates information on drug physicochemical and the human physiological properties along with multiscale data on CNS drug disposition for accurate prediction of the human brain PK.
Here, we applied LeiCNS-PK3.0 to predict the human brain unbound PK profiles of Remdesivir, Molnupiravir, and Nirmatrelvir. These predictions were compared against EC90 of the delta and omicron variants of SARS-CoV-2 to identify the extent to which effective concentrations in the brain are reached.
Methods
LeiCNS-PK3.0 simulations: Brain PK predictions were simulated using the LeiCNS-PK3.0 framework, physiological parameters of healthy adult human [5], and the recommended dosing regimen. Simulations were performed in R (V4.1.2) [6] using the package RxODE (V1.1.4) [7].
Intracellular PK assessment: The prodrugs Molnupiravir and Remdesivir are metabolized intracellularly to the active nucleoside analogues, EIDD-2061 and GS-443902, respectively. The two analogues are hydrophilic and hence might accumulate intracellularly with repeated dosing, particularly GS-443902 given its long half-life (43.3 hours) [8]. In vitro [9] and in vivo [10] intracellular kinetic profiles of the parent drugs and their metabolites were used to predict the intracellular levels in humans.
Drug efficacy: Evaluation of drug efficacy was performed by comparing the predicted brainECF PK profile against the EC90 of the omicron and delta variants [11]. In addition, the average efficacy of the drug over the treatment duration was calculated using the brainECF concentrations and EC50 [12].
Sensitivity analysis: A sensitivity analysis was performed to assess the impact of the pathophysiological changes associated with COVID-19 on brain PK profiles. Model parameters were altered by 10% and 200% and the impact on PK profiles was evaluated using the PK measures: Cmax, Tmax, AUC, and half-life.
Results
The predicted brainECF PK profile of Nirmatrelvir was above the EC90, with a calculated average efficacy of 87 and 96% against the delta and omicron variants, respectively. On the other hand, the brainECF PK profile of Remdesivir was below the EC90 of both variants, expect at Tmax. Also, the intracellular levels of GS-443902, while increasing with repeated dosing, were below the calculated intracellular EC90. Similarly, the predicted brainECF PK profile of EIDDD-1931, the stable plasma metabolite of Molnupiravir, was consistently below the EC90 of both variants. Intracellularly, EIDD-2061 did not achieve sufficient levels for adequate antiviral activity.
Based on the sensitivity analysis, the CNS pathophysiology of COVID-19 resulted in very little to no impact on brainECF PK profiles. Hence, it is rational to assume that the PK predictions of healthy humans will reflect that of COVID-19 patients.
Conclusions
We identified Nirmatrelvir as a promising candidate for future studies investigating the CNS efficacy of anti-COVID-19 drugs against the delta and the omicron variants. The study also represents an application of how integrating our LeiCNS-PK3.0 PBPK framework and multiscale PK data can help explore the uncharted territory of the human brain PK.
[1] Philippens, I. H. C. H. M. et al. BIORXIV (2021).
[2] Chou, S. H. Y. et al. JAMA Netw. Open (2021) 4, 1–14.
[3] Douaud, G. et al. Nature (2022).
[4] Stein, S. et al. Res. Sq. (2021).
[5] Saleh, M. A. A. et al. J. Pharmacokinet. Pharmacodyn. (2021) 48, 725–741.
[6] Gupta, M., et al. ACS Chem. Neurosci. (2020) 11: 205–224.
[7] Wishart, D. S. et al. Nucleic Acids Res. (2017) 46, D1074–D1082.
[8] R Core Team (2019). R Foundation for Statistical Computing, Vienna, Austria.
[9] Fidler, M., et al. (2019).
[10] Humeniuk, R., et al. Clin. Pharmacokinet. (2021) 60: 569–583.
[11] Gilead Sciences. Prod. Doc. (2020) 1–127.
[12] Painter, G.R., et al. Antiviral Res. (2019) 171:1–10.
[13] Gonçalves, A. et al. CPT Pharmacometrics Syst. Pharmacol. (2020) 9, 509–514.
[14] Rosales, R., et al. BioRxiv (2022).
[15] Douaud, G., et al. Nature (2022).
[16] Erickson, M.A., et al. Int. J. Mol. Sci. (2021) 22: 1–28.
[17] Qin, Y., et al. J. Clin. Invest. (2021) 131:1-12.
[18] Saleh, M. A. A. & de Lange, E. C. M. Pharmaceutics (2021) 13, 1–17.
[19] European Medicines Agency (2021). Assessment report of Paxlovid in COVID-19. 1–131.