Transporter and Metabolic-based Drug-Drug Interactions: The Blood-Brain Barrier and The Liver

Abstract

The HIV Protease inhibitor (PI)-based DDIs are complex, unpredictable and often paradoxical. Investigations dedicated to evaluating PIs' potential as inducers of the CYPs and drug transporter, and PIs' as substrates of hepatic uptake transporters haven been scarce, especially those using human hepatocytes, an advanced in vitro systems that provides the most comprehensive and in vivo-like hepatocyte cellular characteristics. These investigations are much needed to better understand and predict the in vivo DDIs with PIs as precipitant and/or object drugs. Therefore, we used human hepatocytes and performed 1) comprehensive quantification of the net induction potential of the eight PIs for nine major hepatic CYPs (CYP3A, 2B6, 1A, 2C8, 2C9, 2C19, 2D6, 2E1, 2A6) at the activity level, as well as their induction of the mRNA transcripts of the same CYPs, and the hepatic drug transporters (P-gp, OATPs, MRPs), and 2) transport studies to examine whether the hepatic uptake of PIs is transporter mediated. We showed that the majority of PIs (particularly amprenavir) produced significant induction in CYP3A4 mRNA expression. They were also net and modest inducers of CYP2B6 (mRNA and activity), and OATP1B1 and P-gp (mRNA). Based on these findings, we were able to establish qualitative agreement between our in vitro results and those observed in vivo. Furthermore, from these studies, we also learned that the hepatic uptake of ritonavir, lopinavir, and nelfinavir was dominated by passive diffusion. However, amprenavir was transported into human hepatocytes, but not by OATP-mediated transport. P-gp is functionally important at the human blood-brain barrier (BBB). However, we question whether BBB P-gp can be maximally inhibited or induced by a drug at its clinically-approved concentration, and whether such P-gp mediated DDIs are clinically significant, and be translated to improve the treatment of CNS diseases. To address these questions, we recruited healthy human volunteers to conduct positron emission tomography (PET) imaging to quantify the magnitude of change in BBB P-gp activity in comparison to baseline by quinidine (model P-gp inhibitor), and after chronic treatment of rifampin (model P-gp inducer) using 11C-verapamil as our P-gp PET substrate. In addition, we evaluated whether the level of P-gp inhibition can be predicted using preclinical data. Our findings show quinidine can significantly inhibit P-gp at the human BBB, and therefore has the potential produce clinically significant DDIs with P-gp substrate drugs with narrow therapeutic index and/or significantly effluxed from brain by P-gp. The quinidine-verapamil P-gp-based DDI at the human BBB was also successfully predicted by the macaque model, but not by the rat model. Conversely, we found that chronic rifampin treatment did not induce P-gp at the human BBB, and highlighted the need for future investigation to determine whether nuclear receptors (e.g., PXR, CAR, VDR, GR) known to regulate P-gp in other tissues are present at the human BBB. In summary, we addressed and explained many unanswered questions and key issues that we believe have contributed to the complexity of DDIs associated with HIV protease inhibitors and those mediated by P-gp at the human BBB. However, many new and interesting questions are raised from these studies, and remained to be addressed.