Session 2: Biological, Experimental and Methodological Challenges Associated with the Study of Drug Transporters
Chairs: Aleksandra Galetin, University of Manchester, Manchester, England, United Kingdom and Laurent Salphati, Genentech Inc., South San Francisco, California, USA
Critical in vitro Factors to Consider when Conducting IVIVE of Transporter-based Drug Disposition 
Jashvant Unadkat, University of Washington, Seattle, Washington, USA
Challenges in Translation of in vitro Transporter Data to Predict Unbound Tissue Concentrations
Xiaoyan Chu, Merck & Co., Inc., Rahway, New Jersey, USA
Challenges in Pharma Industry using in vitro and Preclinical Systems to Predict in vivo Clearance of Transporter Substrates 
Yurong Lai, Gilead Sciences, Foster City, California, USA
Investigations of Compounds Elimination Through Direct Transporter-Mediated Intestinal Secretion
Laurent Salphati, Genentech Inc., South San Francisco, California, USA
The Effect of Plasma Protein on Hepatic Drug Clearance in vitro and its Implications for IVIVE
Laura Francis, The University of Manchester, Manchester, United Kingdom
Is the protein-mediated transport effect an artifact or a real phenomenon?
Mengyue Yin, University of Washington, Seattle, Washington, USA
A Novel in vitro Tool for Testing Drug Interactions of Human MDR3 Transporter
Zsuzsanna Gaborik, SOLVO Biotechnology, a Charles River Company, Budapest, Hungary
Panel Discussion with All Speakers
Key:
Complete
Next
Failed
Available
Locked
Critical in vitro Factors to Consider when Conducting IVIVE of Transporter-based Drug Disposition | Dr. Jashvant Unadkat
Recorded 03/03/2021
Recorded 03/03/2021
In drug development, in vitro to in vivo extrapolation (IVIVE) of transporter-mediated drug clearance (TMCL) remains a challenge, especially when multiple transporters are involved (e.g. basal and apical). To conduct IVIVE, choice of the in vitro system to adopt is critical. Typically, one of the following is chosen: primary cells (e.g. human hepatocytes), transporter-expressing cells (TEC) or inside-out vesicles (for efflux transporters). Then, the transporter-mediated drug clearance (TMCL) obtained in these in vitro systems is extrapolated to that in vivo using a scaling factor. Scaling factors typically used are physiological scaling (e.g. number of hepatocytes per gram of liver), relative activity factor (RAF) or relative expression factor (REF). RAF scaling relies on the availability of transporter-selective probes where RAF is the ratio of the TMCL of the probe drug in vivo and in vitro. REF is the ratio of the abundance of transporters in human tissues (e.g. liver) and in vitro (e.g. TEC). For IVIVE of TMCL to be successful when using these scaling factors, the following must be true. First, the rate-determining step in the in vivo TMCL must be known or identified; second, the mechanisms (including co-factors) of transport in in vitro systems must recapitulate that in vivo; third, the scaling factors used should be correct. In my presentation, these assumptions will be dissected in detail and case studies will be presented where these assumptions were correct or incorrect and respectively resulted in success or failure in IVIVE of TMCL. Supported in part by UWRAPT, a public-private consortium funded (past or present) by Gilead, Takeda, Amgen, Genentech, Biogen, Merck & Co, BMS, Pfizer, AstraZeneca.
Challenges in Translation of in vitro Transporter Data to Predict Unbound Tissue Concentrations | Dr. Xiaoyan Chu
Recorded 03/03/2021
Recorded 03/03/2021
Unbound tissue/intracellular drug concentrations are important determinants of the efficacy and toxicity of drugs. Quantitative prediction of unbound drug concentrations in targeted tissues, such as liver and brain, remains a significant challenge, especially for drugs with their tissue exposure modulated by multiple uptake/efflux transporters. In vitro assays, integrated with preclinical/clinical studies, imaging techniques, and physiologically based pharmacokinetic (PBPK) modeling will help to advance our capability for quantitative prediction of unbound tissue/intracellular drug exposure. In recent years, various in vitro hepatocyte models/methodologies have been developed to measure the hepatocyte-to-medium partition coefficient for unbound drug (in vitro Kpuu) in order to translate these data to predict steady-state unbound liver-to-blood partition coefficient (in vivo Kpuu), which is considered a critical and useful parameter for the estimation of unbound drug concentrations in the liver.
This presentation aims to 1) summarize the advantages and limitations of in vitro hepatocyte Kpuu models, 2) discuss critical challenges, knowledge and data gaps which limit the translation of in vitro Kpuu data to in vivo, 3) provide perspectives on the utility and limitation of Kpuu measurement in drug discovery and development, and 4) discuss future considerations to improve the prediction of unbound tissue drug concentrations using in vitro data.
Challenges in Pharma Industry using in vitro and Preclinical Systems to Predict in vivo Clearance of Transporter Substrates | Dr. Yurong Lai
Recorded 03/03/2021
Recorded 03/03/2021
When only animal and in vitro human data are available in the drug discovery stage, human PK prediction can be a significant challenge to drug candidates undergoing transporter-mediated clearance. The extended clearance concept (ECC) derived from the concept of the well-stirred liver model is often used to estimate the total CL by incorporating the processes of hepatic uptake, metabolism and biliary excretion. However, Since the ECC does not denote the liver as a drug distribution organ, it can be misused for compound selection in drug discovery due to the inability in predicting human PK profiles with the right T1/2 and trough concentrations. In order to overcome the shortcomings of the ECC, a PBPK model developed and validated in preclinical animals is necessitated for scaling factors (SFs); The SFs can then be incorporated in a human PBPK model to prospectively predict human PK using in vitro human data.
The presentation will include:
1. Human dose prediction in the drug discovery phase
2. The usefulness and limitations of ECC
3. Strategies to develop and validate a monkey PBPK model for scaling factors (SFs) of IVIVC
4. Examples of prospective human PK prediction in a human PBPK model using human in vitro data and the SFs derived from monkeys will be provided.
Investigations of Compounds Elimination Through Direct Transporter-Mediated Intestinal Secretion | Dr. Laurent Salphati
Recorded 03/03/2021
Recorded 03/03/2021
The Effect of Plasma Protein on Hepatic Drug Clearance in vitro and its Implications for IVIVE | Dr. Laura Francis
Recorded 03/03/2021
Recorded 03/03/2021
Over the past 40 years the effect of plasma protein (PP) on clearance (CL) has been studied in both perfused livers and isolated hepatocytes, leading to much debate on the mechanism of plasma protein mediated uptake (PMU). However, a collective analysis of the impact of PMU on CL parameters from published hepatocyte studies (routinely used for IVIVE) has not previously been performed. The following study collected this literature data, generating a dataset of 26 compounds with a wide variety of physico-chemical, drug and in vivo properties. The presence of PP enhanced CL beyond that conventionally calculated using fu, and the increase in CL in the presence of PP was correlated with the fu, in vitro and in vivo (fup), and absolute unbound intrinsic clearance in vivo (CLint,u in vivo), in rat and human hepatocytes. PMU appeared to be more important for highly bound (fup <0.1) and high CLint,u in vivo drugs, but was independent of species, assay conditions (although the type of PP used in vitro may warrant further research), ionisation and ECCS group, potentially suggesting a more generic mechanism underlying the PMU phenomenon. The quantified relationship between the fup and impact of PMU was used to assess if accounting for PMU could improve IVIVE predictions using the Wood et al. (2017)1 database (n ≥100). Incorporation of predicted PMU improved the IVIVE of hepatic CL, achieving an AFE of 1.17, and >50% of compounds predicted within a 2-fold error, for both rat and human datasets.
1. Wood FL, Houston JB, and Hallifax D (2017) Clearance Prediction Methodology Needs Fundamental Improvement: Trends Common to Rat and Human Hepatocytes/Microsomes and Implications for Experimental Methodology. Drug Metab Dispos 45:1178-1188.
Is the protein-mediated transport effect an artifact or a real phenomenon? | Dr. Mengyue Yin
Recorded 03/03/2021
Recorded 03/03/2021
Successful in vitro to in vivo extrapolation (IVIVE) of transporter-mediated hepatic drug clearance (CLH) is important for drug development. However, based on human hepatocyte studies, considerable underprediction of CLH, especially of highly protein-bound OATP1B1 substrate drugs, has been widely reported (Soars et al., 2007; Zou et al., 2013). However, CLH predictions are much improved (but do not completely overcome the underpredictions), by inclusion of plasma or plasma proteins (e.g. albumin) in human hepatocyte transport studies. This is due to the observed higher intrinsic uptake clearance (passive and active) in the presence of plasma proteins (Kim et al., 2019; Bowman et al., 2020). These data have led to the hypothesis of “protein-mediated transport effect (PMTE)” (Miyauchi et al., 2018). PMTE proposes an interaction between the albumin-drug complex and the cell surface leading to enhanced dissociation of the drug-albumin complex resulting in increased local unbound drug concentration and therefore uptake (Miyauchi et al., 2018). If this hypothesis is correct, the following should be observed in the presence vs. absence of plasma proteins: 1) the active (e.g. by OATP1B1) and passive uptake CL of the drug should increase to the same extent; and 2) the unbound IC50 of the inhibitor of the transporter should decrease, as this IC50 also should also dependent on the local unbound inhibitor concentration; 3) an increase in the slope (i.e. the uptake rate), but not the intercept of the uptake vs. time curve. To determine if the above is true, we determined the total, active (OATP1B1-mediated) and passive uptake intrinsic CL (CLint;) of a cocktail of OATP1B1 substrates [fluvastatin (FLV), pitavastatin (PTV), cerivastatin (CRV), atorvastatin (ATV) and rosuvastatin (RSV)], in the absence and presence of 1%, 2%, or 5% human serum albumin (HSA), by HEK293 MOCK and OATP1B1-transfected cells [incubated without or with unbound 500 µM rifampicin (RIF) to completely inhibit OATP1B1]. Protein binding of each drug was measured in all experiments. In addition, the unbound drug concentration of each drug, in the absence and presence of HSA, was maintained approximately the same. Although the data presented below are for FLV as the substrate, they are similar to those we obtained with the other statins. We found: 1) In the presence of HSA, the increase in the passive CLint (that is of the unbound drug) of FLV was greater (especially for 5% HSA) than that either the total or active uptake CLint of the drug (Table 1). 2) The unbound IC50 of ATV did not decrease in the presence of HSA (e.g. 1.63 µM in HBSS vs. 1.35 µM in 2% HSA). 3) Interestingly, in the presence of HSA, the intercept of the uptake curve increased proportionally to the HSA concentration used in the experiments, in both the MOCK and OATP1B1 cells incubated with RIF [more than 10-fold, consistent with observation by others (Bi et al., 2020)]. In addition, the increase in intercept of the uptake curves was positively correlated with the degree of protein binding. An increase in the intercept is usually interpreted as non-specific binding (NSB) of the drug to the cells; in this case, NSB of the albumin-drug complex to the cells. Indeed, such binding, measured by quantitative targeted proteomics, was found to be proportional to the HSA concentration used in the experiments. We are now conducting studies to determine if this NSB is time-dependent. If it is, the increase in slope of the uptake curves and therefore the increase in CLint in the presence of plasma proteins could be entirely an artifact. This is because when uptake into hepatocytes or cells is determined, one does not distinguish between the drug that is within the cell (reflecting uptake) from the drug that is bound to the cell membrane (NSB). Even if NSB is not time-dependent, the increase in total CLint observed by us and others is primarily due to an increase in passive diffusion CLint of drugs. In conclusion, all our results do not support the PMTE hypothesis and therefore other mechanisms for the effect of plasma proteins on CLint of drugs (including NSB) need to be further explored. In addition, mechanisms other than the presence of plasma proteins should be investigated to explain the underprediction of CLH by IVIVE.
Acknowledgement: The work was supported by Certara's Simcyp Division Grant and Partnership Scheme and University of Washington Research Affiliate Program on Transporters (UWRAPT).
A Novel in vitro Tool for Testing Drug Interactions of Human MDR3 Transporter | Dr. Zsuzsanna Gaborik
Recorded 03/03/2021
Recorded 03/03/2021
Background: MDR3 (ABCB4) is almost exclusively expressed in the liver, where plays an essential role in the process of bile formation. However, MDR3 has been also demonstrated to be highly expressed in leukemic stem cells and blasts in patients with acute myeloid leukemia. Its physiological function is crucial as evidenced by a broad spectrum of phenotypes ranging from progressive familial intrahepatic cholestasis type 3 (PFIC3) to MDR3-related cholestatic liver disorders which develops in adults with MDR3 deficiency. Because MDR3 is responsible for the secretion of phospholipids from hepatocytes into bile, inhibition of its function by drugs and/or their metabolites may be related to cholestasis and drug-induced liver injury (DILI). At present, compared to other drug transporters there are only few identified drug substrates and inhibitors of MDR3. Therefore, we aimed to develop an assay system for studying the interaction of several hepatotoxic, anticancer, as well as MDR1 interactor drugs with MDR3. Methods: Since MDR3 shares up to 76% identity and 86% similarity in the amino acid sequence with MDR1, we chose the approach of using an MDCKII cell line completely lacking endogenous Abcb1, generated by CRISPR-Cas9 gene editing technology, as parental cell. Using this Abcb1-knock out MDCKII cells a stable cell line was generated which express the human MDR3 and can be grown as monolayer, therefore applicable for transcellular transport assays. Transcellular transport of digoxin as a known MDR3 substrate drug can be investigated in this experimental setup. This in vitro system applicable for screening MDR3 specific drug substrates, as well as inhibitors of MDR3-mediated digoxin or other drug transport independently of MDR1 activity. The assay system can also be applicable for the investigation of the transport of endogenous substrates of MDR3. Results: Indicative examples including, anticancer tyrosine kinase inhibitors (erlotinib, imatinib, sorafenib); the antiretroviral therapeutics (darunavir, lopinavir, ritonavir and saquinavir); known MDR1 inhibitors and other drugs such as ivermectin, daunomycin, which all significantly inhibited the MDR3-mediated digoxin transport. Conclusions: the use of this in vitro model to study drug interactions with MDR3 can be considered a reproducible, conclusive and easy to use system. Our results are also consistent with the prior hepatotoxicity causality knowledge and provide new findings with identifying drugs as potential MDR3-inhibitors.