Assessing the relative contribution of CYP3A-and P-gp-mediated pathways to the overall disposition and drug-drug interaction of dabigatran etexilate using a comprehensive mechanistic physiological-based pharmacokinetic model

Dabigatran etexilate (DABE) is a clinical probe substrate for studying drug-drug interaction (DDI) through an intestinal P-glycoprotein (P-gp). A recent in vitro study, however, has suggested a potentially significant involvement of CYP3A-mediated oxidative metabolism of DABE and its intermediate monoester BIBR0951 in DDI following microdose administration of DABE. In this study, the relative significance of CYP3A- and P-gp-mediated pathways to the overall disposition of DABE has been explored using mechanistic physiologically based pharmacokinetic (PBPK) modeling approach. The developed PBPK model linked DABE with its 2 intermediate (BIBR0951 and BIBR1087) and active (dabigatran, DAB) metabolites, and with all relevant drug-specific properties known to date included. The model was successfully qualified against several datasets of DABE single/multiple dose pharmacokinetics and DDIs with CYP3A/P-gp inhibitors. Simulations using the qualified model supported that the intestinal CYP3A-mediated oxidation of BIBR0951, and not the gut P-gp-mediated efflux of DABE, was a key contributing factor to an observed difference in the DDI magnitude following the micro-versus therapeutic doses of DABE with clarithromycin. Both the saturable CYP3A-mediated metabolism of BIBR0951 and the solubility-limited DABE absorption contributed to the relatively modest nonlinearity in DAB exposure observed with increasing doses of DABE. Furthermore, the results suggested a limited role of the gut P-gp, but an appreciable, albeit small, contribution of gut CYP3A in mediating the DDIs following the therapeutic dose of DABE with dual CYP3A/P-gp inhibitors. Thus, a possibility exists for a varying extent of CYP3A involvement when using DABE as a clinical probe in the DDI assessment, across DABE dose levels.


In vitro-to-in vivo extrapolation factors CYP3A4/5 pathways
Intersystem extrapolation factors (ISEF) of intrinsic clearance (CLint) was calculated, using the following equation (Proctor et al.,2004): (HLM) CL int (rhCYP) × CYP abundance (HLM) where CLint (HLM) is HLM intrinsic clearance, CLint (rhCYP) is recombinant human CYP3A4/5 intrinsic clearance of DABE, BIBR095, and CYP abundance (HLM) is an abundance of CYP3A4/5 in the liver.All in vitro kinetic parameters, specifically Km and Vmax of DABE and its metabolites, were obtained from Udomnilobol et al. (2023) .Using the Simcyp population based ADME simulator software version 20 (Certara, Sheffield, UK), the in vitro Vmax values were scaled up to whole liver Vmax values through multiplying with ISEF, liver CYP3A4/5 abundance, amount of microsomal protein per liver, and liver weight.Then the well-stirred model was used for extrapolating organ CL to in vivo CL values.

CES1/2 pathways
The in vitro values of Km and Vmax (HLM and HIM) were obtained from Udomnilobol et al. (2023).Through the software, the well-stirred model was applied to obtain in vivo CL values upon conversion of in vitro Vmax to whole liver Vmax values (by multiplying with the amount of microsomal protein per liver and liver weight for HLM and the amount of microsomal protein per intestine).

Plasma esterase
In vivo clearance from plasma esterase was estimated from the in vitro half-life values of DABE and BIBR0951, using the equation derived from Imai et al., 2022 in the Simcyp software:

Rapid equilibrium dialysis
Fractions unbound in plasma (fu,p) of BIBR0951 and BIBR1087 were determined by rapid equilibrium dialysis method.Briefly, the compound at 1 μM in pooled human plasma was dialyzed against phosphate buffer saline (0.1 M sodium phosphate + 150 mM NaCl, pH 7.40) at 37°C in a humidified CO2 incubator for 4 h.The organic solvent in the system was kept not more than 1%.For inhibition of BIBR0951 and BIBR1087 hydrolysis in plasma, the experiment was performed in the presence of esterase inhibitor, 1 mM phenylmethylsulfonyl fluoride (PMSF).After incubation, the dialyzed plasma and buffer were collected and added with an equal volume of blank buffer and plasma, respectively.Then, the samples were quenched with 3 volumes of ice-cold acetonitrile + 0.1 % formic acid containing labetalol as an internal standard.After centrifugation, the supernatants were collected and diluted with water + 0.1% formic acid prior to LC-MS/MS analysis.The fu,p was calculated by dividing the peak area ratio in the dialyzed buffer to that in the dialyzed plasma.The % recovery was calculated by 100 × (a summation of peak area ratio in the dialyzed buffer and plasma)/peak area ratio in the initial plasma before dialysis.In this experiment, ketoconazole (1 μM) was used as a positive control.

Blood/plasma partitioning
Blood to plasma partition coefficients (B:P) of BIBR0951 and BIBR1087 were determined as described previously (Yu et al., 2005).Briefly, the compounds were separately incubated at 1 μM in whole blood and corresponding plasma (reference plasma) at 37°C for 1 h.The organic solvent in the incubations was kept not more than 1%.For inhibition of BIBR0951 and BIBR1087 hydrolysis in plasma, the experiment was performed in the presence of esterase inhibitor, 1 mM phenylmethylsulfonyl fluoride (PMSF).At 0 and 1 h, both whole blood and reference plasma were collected and centrifuged at 2,000 g for 10 min at 25°C.Then, plasma samples were collected from the centrifuged whole blood and reference plasma, and immediately quenched with ice-cold acetonitrile + 0.1 % formic acid containing internal standard (labetalol).After centrifugation, the supernatants were collected and diluted with water + 0.1% formic acid prior to LC-MS/MS analysis.The B:P values were calculated by dividing the peak area ratio in the partitioned plasma (from whole blood) at 1 h to that in the reference plasma at 1 h.The % recovery was calculated from the formula: 100 × (peak area ratio in the reference plasma at 1 h/peak area ratio in the reference plasma at 0 h).In this experiment, chloroquine (5 μM) was used as a positive control.

Table 1 : Data sets for model development and qualification
95% B for 1 min, column reconditioning to 5% B for 0.2 min, and post-analysis equilibration at 5% B for 0.8 min.Turbo V TM ion source was operated in the electrospray positive ionization mode under 40 psi of curtain gas, 4500 volts of ion spray voltage, 450°C of source temperature,