Category: Preclinical Development
Purpose: The unbound liver-to-blood partition coefficient (Kpuu) is utilized to provide estimates of unbound intracellular drug concentrations. It is generally accepted that estimation of tissue- or target-specific unbound drug concentration is imperative to accurately assess in vivo pharmacological efficacy, toxicological effects, and drug-drug interaction potential of therapeutic drugs. One may assume that active uptake would result in increased intracellular unbound drug accumulation, whereas efflux would decrease tissue-specific unbound drug concentration. Therefore, differential concentrations of unbound drug in the blood versus tissue are often anticipated and considered crucial in predictions related to drug disposition. The Biopharmaceutics Drug Disposition Classification System (BDDCS) is a tool used in early discovery to predict major routes of elimination, oral drug disposition, and anticipate clinically significant involvement of transporters. Here we explore the relationship of BDDCS class with human Kpuu values, anticipating that BDDCS class 1 drugs will have Kpuu values close to 1, as transporters are not expected to play a clinically relevant role.
Methods: A literature search identified 12 articles with experimental determinations of human Kpuu values. The following methods and number of Kpuu determinations were identified: Extended Clearance Method (ECM) (8 publications; 49 values); temperature method (3 publications; 68 values); homogenization method (3 publications; 40 values); logD 7.4 method (1 publication; 18 values) PET imaging (1 publication; 1 value). Within each methodology, drugs were classified by BDDCS. Investigations into consistency of Kpuu value for a single drug between laboratories, and across the various experimental methods were also investigated.
Results: The average Kpuu values by methodology and BDDCS class (1, 2, 3, 4, not classified) are as follows: ECM (n=13) (2.0, 0.61, 14, 0.41, none); ECM assuming no active basolateral efflux (n=26) (1.3, 1.0, 2.0, 0.7, 0.45); ECM assuming no active efflux nor apical secretion (n=16) (none, 5.5, 10, 9.5, none); ECM assuming no active efflux, no secretion, no metabolic clearance (n=6) (4.2, none, 4.0, 10, none); Temperature method (n=68) (4.6, 5.5, 7.5, 7.0, 7.8); Homogenization method (n=40) (5.6, 3.0, 7.0, 1.9, 8.0); LogD 7.4 method (n=18) (11, 7.5, 4.6, none, 17); PET imaging (n=1) (2; Class 3 drug). Analysis of Kpuu value by drug (across different laboratories using the same method, as well as when determined by differing experimental methods) revealed significant variability. For example, the average cerivastatin Kpuu value was 6.4, with a standard deviation of 5.3 and a range of 1.3 to 18 (n=9); and individual values were as follows: (ECM assuming no active efflux - 1.6; ECM assuming no active efflux, no secretion, no metabolic clearance – 5.0, 4.3, 3.3; temperature method – 7.7, 11, 5.4; homogenization method – 1.3; logD 7.4 method – 18).
Conclusion: The analysis indicates that there is no association between BDDCS class and average Kpuu value. Further, it is notable that the average Kpuu value of BDDCS class 1 drugs was not close to 1, as would be anticipated for drugs that are not substrates of transporters. This may be explained by the fact that BDDCS class 1 drugs may show in vitro transporter effects, but these will not be clinically significant in vivo, suggesting that perhaps an in vitro Kpuu value is irrelevant for use in predicting in vivo drug disposition for BDDCS class 1 drugs. Additionally, significant variability in the Kpuu value was observed both by laboratory as well as by experimental method, further highlighting the difficulty in performing these measurements. Currently, our lab is questioning the utility of Kpuu in predicting the magnitude of AUC change for clinical drug-drug interactions, based on theoretical considerations. In summary, the difficult-to-measure Kpuu value displays extensive variability due to the intricate experimentation necessary, may not be reliable for BDDCS class 1 drugs, and is theoretically unnecessary for predicting drug-drug interactions after oral dosing.