Predicting human pharmacokinetics from preclinical data: absorption

Abstract

Predicting the rate and extent of oral absorption of drugs in humans has been a challenging task for new drug researchers. This tutorial reviews in vitro and PBPK methods reported in the past decades that are widely applied to predicting oral absorption in humans. The physicochemical property and permeability (typically obtained using Caco-2 system) data is the first necessity to predict the extent of absorption from the gut lumen to the intestinal epithelium (F a). Intrinsic clearance measured using the human microsome or hepatocytes is also needed to predict the gut (F g) and hepatic (F h) bioavailability. However, there are many issues with the correction of the inter-laboratory variability, hepatic cell membrane permeability, CYP3A4 dependency, etc. The bioavailability is finally calculated as F = F a × F g × F h. Although the rate of absorption differs by micro-environments and locations in the intestine, it may be simply represented by k a. The k a , the first-order absorption rate constant, is predicted from in vitro and in vivo data. However, human PK-predicting software based on these PBPK theories should be carefully used because there are many assumptions and variances. They include differences in laboratory methods, inter-laboratory variances, and theories behind the methods. Thus, the user's knowledge and experiences in PBPK and in vitro methods are necessary for proper human PK prediction.

Baicalin Enhanced Oral Bioavailability of Sorafenib in Rats by Inducing Intestine Absorption

“…Therefore, we examined the absorptive profiles of SOR with/without the treatment of BG in the in situ single-pass rat intestinal perfusion model. The P eff and K a of the drugs are the key biopharmaceutical variables to access the rate and extent of intestinal absorption ( Yim et al, 2020 ). In the present study, coperfusion of SOR with BG demonstrated significantly higher P eff and K a values in the intestine in the studied concentration range.…”

Physiologically‐Based Kinetic Modeling Predicts Similar In Vivo Relative Potency of Senecionine N‐Oxide for Rat and Human at Realistic Low Exposure Levels

“…The absorption rate constants for uptake from the GI tract into liver were calculated based on the apparent permeability coefficient ( P app ) values in Caco‐2 cells [ 48 ] and an absorption rate constant for uptake of RID from the GI tract to the liver of 0.72 h −1[ 38 ] using Equation (3). [ 16,38,55 ] The P app values of SENO and SEN were 0.72 (10 −6 cm s −1 ) and 16.26 (10 −6 cm s −1 ), [ 48 ] yielding absorption rate constants for uptake from the GI tract into the liver for SENO ( k a SENO ) and SEN ( k b SEN ) of 0.17 and 3.90 h −1 , respectively. The k a values thus obtained were compared to and combined with other reported absorption rate values from Wang et al., [ 46 ] and an average value was taken for the PBK model as shown in Table 1 .…”

“…The k a values thus obtained [46] 0.276 0.330 At 5.7 mg SEN kg −1 bw [46] 0.170 3.900 P app [48] and k a values based on Equation (3). [ 16,38,55] 0.245 1.552 Average value from literature as final k a or k b values were compared to and combined with other reported absorption rate values from Wang et al, [46] and an average value was taken for the PBK model as shown in Table 1. The transfer rate constants of SENO and SEN from the small intestine (lower ileum) to the intestinal microbiota (caecum) (k in1 for SENO and k in2 for SEN) were assumed to be the same as the rate constants for transfer of orally-administered drugs, amounting to 0.46 h −1 .…”

Beyond the Michaelis–Menten: Accurate Prediction of Drug Interactions Through Cytochrome P450 3A4 Induction

“…This is based on the assumption that the orally administered drug is absorbed completely from the intestine (i.e., the gastrointestinal absorption () is one), which is also recommended in the current FDA guidance for DDI if no relevant data are available 8 . However, several previous studies have reported that the values obtained from effective permeability experiments would be less than one 9–11 . Thus, obtained from the assumption of could be highly underestimated, leading to inaccurate predictions of AUCR.…”

“…8 However, several previous studies have reported that the F a values obtained from effective permeability experiments would be less than one. [9][10][11] Thus, F g obtained from the assumption of F a = 1 could be highly underestimated, leading to inaccurate predictions of AUCR.…”