Over the last 15 years, the importance of drug transporters has become paramount for understanding drug absorption, distribution into tissues, and in particular, drug-drug interactions (DDIs). In vitro-in vivo extrapolation (IVIVE) is an approach to link in vitro systems to the human in vivo situation using algorithms and physiologically-based scaling factors. In this blog, I will discuss the expansion of gut transporters and IVIVE techniques in the Simcyp Simulator’s Advanced Dissolution, Absorption and Metabolism (ADAM) and multi-layer ADAM (M-ADAM) models.
Transporter-mediated DDIs: A regulatory and industry concern
Modeling and simulation approaches are increasingly being used to inform the risk of DDIs for investigational drugs. For example, a 2016 paper by Pan and his FDA colleagues highlighted some of the regulatory submissions that evaluated the role of transporters.1 In particular, they highlighted four submissions that utilized the Simcyp Simulator to develop physiologically-based pharmacokinetic (PBPK) models to inform drug labels.
The FDA also commented on the use of PBPK to predict transporter and metabolic DDIs in a 2015 paper by Wagner et al. With respect to transporter-mediated DDIs, the FDA’s opinion on the current status of this approach is that IVIVE is not mature due to an inadequate body of information and that its predictive performance has yet to be adequately demonstrated. In addition, authors from the pharma industry commented that the IVIVE scaling factors were poorly understood.
In light of these findings, the Simcyp Consortium members are keen to improve their mechanistic understanding of transporter function for improving IVIVE predictability for transporter-mediated disposition and DDIs.
Expansion of ADAM/M-ADAM models gut transporters
To understand the role of gut transporters in drug disposition, you first need to understand the cell biology of enterocytes—the intestinal absorptive cells. These cells have a polarized morphology where the apical brush border membrane faces the intestinal lumen, and the basolateral membrane is in close proximity with the circulation. Efflux and update transporters transport drugs across both the apical and basolateral membranes.
The Simcyp Simulator Version 17 includes a substantial increase in the number of gut transporters. For example, the apical membrane uptake transporters now include the Organic Anion Transporting Polypeptide 2B1 (OATP2B1), which transports various statins and sulfasalazine, while the highly abundant oligopeptide transporter 1 (PepT1) is often associated with transporting pro-drugs like valacyclovir. On the basolateral membrane, we’ve added three efflux transporters: the Multidrug Resistance-Associated Proteins (MRPs)—MRP1, MRP3 and MRP4, which have been implicated to transport methotrexate, retroviral drugs, and oncologic drugs. The basolateral membrane uptake transporter OATP4C1, which is postulated to transport digoxin, has been added as well.
Scaling transporters: Relative vs. absolute abundance
The Simcyp Simulator ADAM model consists of the following nine segments:
- Jejunum, which is further subdivided into 2 segments
- Ileum, which is further subdivided into 4 segments
The abundance of enzymes and transporters varies spatially in the gastrointestinal (GI) tract. Western Blot and mRNA data can be used to scale the abundance of transporters in vitro to the in vivo abundance in a gut segment.
For years, we’ve relied on scaling in vitro transporters via a relative abundance or expression approach in the intestine. This approach has typically utilized “relative” quantification approaches where suitable standards for specific transporter moieties are not available, ie, methodologies such as Western Blotting and RT-PCR. However, these quantification techniques can still be valuable for gut transporter IVIVE-PBPK in that they are employed to distinguish the region-specific expression of given transporter along the length of the GI tract, once normalized to the proximal jejunum segment (jejunum I) in the ADAM/M-ADAM models.3 For example, for P-glycoprotein (P-gp), the jejunum I has an average value of 1 in a population, and the average value in the ileum is typically approximately 50% higher. Therefore, a scaling factor of 1.5 for P-gp is employed in ileum segments relative to the jejunum I.
This is in contrast to the way we scale for the cytochrome P-450 (CYP) enzymes. For the last dozen years, we have had the capability of scaling CYP450-mediated activity via absolute abundances for CYPs in the gut. Therefore we scale via absolute abundances where in vitro CYP450 data from recombinant expression systems, for example, are provided as intrinsic clearance per pmol of CYP450 (µL/min/pmol CYP450), which are scaled in vivo according to the amount of CYP450 in pmol for each virtual individual in each GI segment for CYPs. In recent years, absolute abundance data have become available for transporters in the intestine. So, we can utilize these data to not only assess the region-specific expression of several transporters via meta-analyses but also enable scaling of gut transporter activity per pmol (an “absolute approach”) using IVIVE-PBPK. Thus, this provides an alternative strategy to the relative expression approach.
Option 1: Relative scaling approach for transporters
There are two major approaches for scaling data from in vitro cell monolayer assays to estimate the intrinsic clearance due to active transport in a GI segment. The first option is the relative scaling approach for transporters.4 First, the user needs to generate a Jmax (the maximal flux capacity of the transporter protein) and Km (the Michaelis-Menten constant—the concentration of drug which permits the transporter to achieve half Jmax). Alternatively, an intrinsic transporter clearance (CLint,T) can be entered by the user. The Jmax or CLint,T is then corrected to the filter surface area of their tissue culture transwell in which the transporter activity across the cell monolayer was derived. If a Km is assigned, then the potential for saturation in transporter activity can be assessed. The driving concentration for the saturation will be dependent on the location (ie, apical or basolateral membrane) and its function (ie, uptake or efflux). A scaling factor to bridge any mechanistic gap in terms of activity or expression for a given transporter between the in vitro and the in vivo system is required. When using the relative scaling approach, this scaling factor is called the Relative Expression Factor (REF)—a unit-less scaler that relies on the ratio of the in vivo transporter abundance (in the proximal jejunum) to the in vitro abundance.
We can then scale through the various GI segments according to the relative expression of a given transporter using a fixed REF for each segment. A coefficient of variance is associated with each transporter, and this propagates variability in transporter expression/activity to virtual individuals.
Option 2: Absolute abundance scaling strategy
For Version 17 of the Simcyp Simulator, we’ve incorporated a new approach—the absolute abundance (ISEF,T) scaling approach.4 It requires the user to consider the derivation of their transporter activity (ie, Jmax and CLint,T) for assignment in the Simulator. The user still needs to measure Jmax and KM, but they also need to measure the transporter’s absolute abundance across their filter monolayers using a suitable proteomic method.
In the first stage, the user corrects their Jmax or CLint,T for the absolute abundance of the transporter per filter. Thus the user needs to measure the pmol abundance of their transporter across the cellular monolayer. You can then scale via the Inter-System Extrapolation Factor for Transporters (ISEF,T), which is a scalar that, like the REF, bridges any mechanistic gap between the in vitro and in vivo system. But the difference to the REF is that the ISEF,T corrects implicitly for activity differences per pmol of transporter in the in vivo versus the in vitro system. Therefore, stage 1 deals with defining the in vitro CLint,T (µL/min/pmol transporter) while considering the potential for saturation (when Km is assigned) and bridging any activity difference to the in vivo system.
Stage 2 of the absolute scaling procedure scales the in vitro transporter activity for the absolute transporter abundance (pmol) in each given GI segment to give the segmental intrinsic clearance in L/hr. The absolute abundance (pmol transporter per GI segment) is derived from assigning the transporter abundance in pmol/mg total membrane protein in the proximal jejunum (ie, jejunum I segment in the ADAM/M-ADAM models) a parameter value obtained from literature meta-analysis (see below). The conversion to a pmol abundance per GI segment is achieved via multiplication of the yield of total membrane protein (mg) in each GI segment and via the region-specific abundance, ie, this would be approximately 1.5-fold higher for P-gp in the ileal segments versus the jejunum I for P-gp.
Gut transporter meta-analysis
We conducted a literature meta-analysis of relative and absolute gut transporter expression in healthy Caucasians. Our final database had greater than 1,750 transporter measurements for 14 transporters. We used the meta-analysis to simulate the transporter abundance for 2000 virtual Caucasians. PepT1 was easily the most abundant transporter in this virtual population. The second most abundant transporter was MRP2.
We also examined regional transporter expression. For example, we normalized OATP2B1 expression to the jejunum I for all of our segments in the ADAM and M-ADAM models. The majority of transporters have a relatively uniform regional expression. Their average expression is within 2-fold of the proximal jejunum expression. The significant exception is the Ileal Bile Acid Transporter (IBAT), which has approximately 100-fold higher expression in the terminal ileum segments than proximal jejunum.
Summary and future directions
This project significantly expanded the number of available transporters and in many instances provided their region-specific expression along the GI tract and their absolute abundances in the proximal jejunum. The incorporation of an absolute scaling approach in addition to the established relative approach represents an alternative mechanistic method to perform IVIVE-PBPK using gut transporter absolute abundances. Ideally, we need more relevant in vitro data to verify this approach. Further quantification of transporter abundances and membrane protein scalars are needed to enhance the robustness of this scaling approach.
 Pan Y, Hsu V, Grimstein M, Zhang L, Arya V, Sinha V, Grillo JA, & Zhao P. (2016). The application of physiologically based pharmacokinetic modeling to predict the role of drug transporters: Scientific and regulatory perspectives. The Journal of Clinical Pharmacology, 56, 122–131. doi:1002/jcph.740
 Wagner C, Zhao P, Pan Y, Hsu V, Grillo J, Huang S, & Sinha V. (2015). Application of physiologically based pharmacokinetic (PBPK) modeling to support dose selection: Report of an FDA public workshop on PBPK. CPT Pharmacometrics Syst. Pharmacol., 4, 226–230. doi:1002/psp4.33
 Harwood MD, Neuhoff S, Carlson GL, Warhurst G, & Rostami‐Hodjegan A. (2013). Absolute abundance and function of intestinal drug transporters: A prerequisite for fully mechanistic in vitro–in vivo extrapolation of oral drug absorption. Biopharm. Drug Dispos., 34, 2–28. doi:1002/bdd.1810
 Neuhoff S, Yeo KR, Barter Z, Jamei M, Turner DB, & Rostami‐Hodjegan A. (2013), Application of permeability-limited physiologically-based pharmacokinetic models: Part I—digoxin pharmacokinetics incorporating P-glycoprotein-mediated efflux. J. Pharm. Sci., 102, 3145–3160. doi:1002/jps.23594
Should you be interested in learning more, please watch the webinar we presented on this topic.