A medRxiv preprint simulated rosuvastatin exposure during ticagrelor coadministration in four age–ABCG2 genotype groups. Exposure-matching results differed by group, but the proposed dose scenarios were not clinically tested.
Rosuvastatin is transported by breast cancer resistance protein (BCRP), which is encoded by ABCG2. Ticagrelor can inhibit this transporter, increasing rosuvastatin exposure. Researchers at Mahidol University and collaborating institutions built physiologically based pharmacokinetic (PBPK) models to examine how that interaction might differ with age and the homozygous ABCG2 c.421C>A variant.
The authors developed and evaluated the component drug models against available clinical pharmacokinetic data, then combined them for a drug–drug interaction model. For the interaction, the simulated and observed AUC ratios were 2.25 and 2.47; the corresponding maximum-concentration ratios were 1.92 and 2.58. The modeled ratios met the authors’ twofold-error criterion.
The component models were assembled in stages. Intravenous data helped constrain distribution and elimination, followed by oral data to characterize absorption. The rosuvastatin model was revised because changing BCRP activity alone in an earlier parameterization did not reproduce the size of the observed genotype effect. The authors adjusted related model parameters before applying the model to the interaction question. The final predictions therefore depended on both physiological assumptions and data-fitted drug parameters.
Ticagrelor development used clinical pharmacokinetic studies with separate training and evaluation arms. For the combined model, however, the available rosuvastatin–ticagrelor study also informed the transporter-inhibition parameter. The agreement with that interaction study should be read in that context, rather than treated as a fully independent prospective test of a parameter chosen without those data.
The study simulated four groups of 100 virtual people each: younger adults and adults aged 65–85 years, each with either the wild-type genotype or the homozygous variant. Rosuvastatin exposure during ticagrelor coadministration was compared with the modeled steady-state AUC during rosuvastatin use alone in the corresponding virtual group. The investigators treated an AUC ratio of 0.80–1.25 as the exposure-matching target.
With half the rosuvastatin dose, the predicted AUC ratio was 1.44 in younger wild-type adults and 0.78 in younger homozygous-variant adults. Both fell outside the study’s 0.80–1.25 target: 1.44 exceeded the upper limit, while 0.78 fell below the lower limit. In the older simulated groups, a quarter dose gave a ratio of 1.06 for wild-type individuals, while a half dose gave 1.12 for homozygous-variant individuals.
The exercise aimed to reproduce an exposure, not to simulate clinical benefit or directly estimate the incidence of muscle toxicity.
The model predicted a larger relative interaction when baseline BCRP function was preserved. When the homozygous variant had already reduced transporter activity, inhibiting that pathway further produced a smaller relative change. This was the authors’ mechanistic explanation for the simulated pattern, not an experimental separation of transporter effects in the virtual participants. Age-related physiology and the stated reduction in CYP2C9 activity contributed additional model assumptions.
The older-adult simulations included an age-related 40% reduction in CYP2C9 activity. Clinical data for older adults, variant carriers and coadministration came from only one study in the relevant scenarios. Those data helped optimize the models but could not independently validate the age–genotype predictions; development and evaluation also relied mainly on data from Caucasian populations.
The proposed regimens were simulated rather than prospectively tested for safety or efficacy in patients.
A retrospective analysis of pharmaceutical regulatory data found better performance for Ames-test (Q)SAR models used in ICH M7 work than for models of mammalian-cell or in vivo genotoxicity. Reliability checks improved performance in some comparisons.
NIH researchers compared ADME@NCATS predictions with laboratory measurements for selected Tox21 compounds. Three basic ADME models reached ROC AUCs of 0.81–0.92; a separate passive blood–brain barrier model reached 0.77. The work supports compound triage, not validated prediction of in vivo toxicity.
