Quantitative Structure Pharmacokinetic Relationships (QSPKR) and Mechanistic Models for Transporter-Mediated Renal Drug Disposition

The kidneys play an essential role in maintaining physiological homeostasis and eliminating toxic endogenous and exogenous waste products, including drugs and drug metabolites. The nephron represents the functional unit of kidneys, consisting of the glomerulus, proximal tubules, loop-of-Henle, distal tubules, and collecting ducts. Renal clearance (CLR) is an important route of elimination for drugs with low to negligible metabolism and biliary clearance. A survey of top 200 prescription drugs in the US in 2010 revealed that 32% of drugs are predominantly eliminated by renal clearance. Renal clearance is the net result of glomerular filtration, active secretion, active reabsorption, and passive reabsorption. Numerous membrane transporters, belonging to ATP-binding cassette (ABC) and solute carrier (SLC) super-families are expressed at the basolateral and brush-border membranes of the proximal tubule cells (PTC), mediate active secretion and active reabsorption of compounds across the PTC. Several studies have extensively evaluated aspects of nephron physiology, physicochemical determinants of renal clearance, and the role of renal transporters in clinically-relevant drug-drug interactions. However, in silico models for the prediction of renal clearance in humans and physiologically-relevant pharmacokinetic-pharmacodynamic (PK/PD) models for drug-transporter interactions that have been reported in the literature are limited. The overall objective of this research was to utilize the knowledge of physiochemical determinants of renal clearance and nephron physiology to develop quantitative structure pharmacokinetic relationships (QSPKR) and mechanistic models for the quantitative prediction of renal clearance and renal transporter-drug interactions. Using a structurally-diverse set of 382 drugs or drug-like compounds, we successfully developed in silico QSPKR models for the prediction of CLR of compounds undergoing net renal reabsorption and renal transporter-mediated renal clearance (Q2 > 0.75). We observed that compounds undergoing net reabsorption predominantly belonged to BDDCS classes 1 and 2 (extensive metabolism). We observed that compounds belonging to BDDCS classes 1 and 2 had a low extent of urinary excretion (UE) and those belonging to classes 3 and 4 (poor metabolism) had a high extent of UE. Therefore, we utilized the human data on amount of drug excreted unchanged into urine (%Ae) of 834 compounds from the literature to determine a quantitative threshold value of 16.8% using the receiver operating characteristic (ROC) curve analysis method to distinguish between compounds undergoing high and low extent of UE. We observed that 78% of compounds with cLogP ≤0.1 have Ae >16.8% and 82% of compounds with cLogP ≥1.5 have A e ≤16.8%. Our analysis indicated that cLogP is an indicator of the extent of UE, but not a quantitative predictor of %Ae (R 2adj. =0.32). Analysis of physiochemical space of compounds in our studies showed a significant overlap between compounds (1) undergoing net secretion and net reabsorption and (2) with high and low extent of UE. This indicated that compounds with similar physicochemical profiles display unique elimination mechanisms in vivo. In order to develop a semi-mechanistic and physiologically--relevant kidney model for evaluating drug-transporter interactions, we first evaluated various literature reports of semi-mechanistic pharmacokinetic models describing active renal secretion and reabsorption using data for two probe drugs, phenolsulfonphthalein (active secretion) and gamma-hydroxybutyric acid (active reabsorption), in rats. Our results indicated that mechanism-based models utilizing the Michaelis-Menten parameters, maximum capacity of transport (VMAX) and affinity (K M) are necessary for evaluating drug-transporter interactions to avoid mis-prediction of renal clearance of compounds. We then developed a semi-mechanistic kidney model incorporating physiologically-relevant fluid reabsorption and transporter-mediated active reabsorption. The model was successfully developed using the data for the drug of abuse gamma-hydroxybutyric acid (GHB) in rats, which exhibits monocarboxylate transporter (MCT1/SMCT1)-mediated active renal reabsorption. The model accounted for 67%, 15%, and 16% of the water reabsorption from the proximal tubules, loop-of-Henle, and distal nephron (distal tubules and collecting ducts), respectively, where the segmental fluid flow rates, volumes, and sequential reabsorption were incorporated as functions of the glomerular filtration rate. The active renal reabsorption of GHB was modeled as MCT1/SMCT1-mediated vectorial transport across proximal tubule cells. Additionally, the model was further applied to evaluate pharmacokinetics of L-lactate, also a substrate for MCT1/SMCT1, in rats. Overdose of GHB leads to coma and even death due to severe respiratory depression. We further developed this model to quantitatively evaluate the effects of non-competitive inhibition by AR-C155858 of MCT1-mediated GHB renal reabsorption and uptake on GHB PK/PD in rats as a potential treatment strategy in GHB overdose. Using simulations, we demonstrated that treatment with AR-C155858 up to 180 minutes post-GHB administration should completely reverse GHB-induced respiratory depression. In conclusion, this research presents model-based approaches for (1) the quantitative prediction of renal clearance and qualitative prediction of extent of urinary excretion in humans and (2) transporter-mediated non-linear drug disposition and drug-drug interactions in rats.