Risk assessment of drug–drug interactions using hepatocytes suspended in serum during the drug discovery process

Abstract

1. This study optimized the reported approach for the prediction of drug-drug interactions (DDIs) using hepatocytes suspended in serum (HHSS) and provided a practical usage of HHSS in the early and late phases of drug discovery. 2. First, the IC50 was determined using HHSS and evaluated as a qualitative index for DDI risks in the early phase. A retrospective study on clinical DDI cases revealed that inhibitors with IC50 < 100 μmol/L caused clinical DDIs while those with IC50 > 100 μmol/L showed weak or no potential for DDIs. Meanwhile, a pragmatic cutoff value could not be determined using previously reported Ki values of recombinant human cytochrome P450s. 3. Second, for a more substantial DDI risk assessment in the later phase, quantitative predictions of clinical DDI based on a static model were attempted by optimizing the most appropriate inhibitor concentration ([I]). The use of hepatic input plasma concentrations as a surrogate for [I] achieved the most successful predictions of the magnitude of increase in the AUC (within a 2-fold range of the observed values for 93.8% of inhibitors). 4. Through this study, we proposed the practical application of HHSS for an effective workflow to explore and profile candidates with less DDI liability.

Use of HμREL Human Coculture System for Prediction of Intrinsic Clearance and Metabolite Formation for Slowly Metabolized Compounds

“…The formation of 4-OH-diclofenac was also within the formation rates at K m observed in the hepatocytes from the 16 different donors . Paracetamol formation rate in HμREL coculture was close to the median value in freshly thawed hepatocytes from 9 different donors, whereas 1-OH-bufuralol formation rate was 20-fold lower compared to formation rates at K m (calculated as V max /2) in a 10-donor pooled hepatocyte batch . Thus, among the major P450 enzymes tested, merely CYP2D6 activity was lower in HμREL coculture, whereas CYP1A2, CYP2B6, CYP2C9, and CYP3A4 activities were on comparable levels with freshly thawed human hepatocytes published elsewhere.…”

“…29 Paracetamol formation rate in HμREL coculture was close to the median value in freshly thawed hepatocytes from 9 different donors, 30 whereas 1-OH-bufuralol formation rate was 20-fold lower compared to formation rates at K m (calculated as V max /2) in a 10-donor pooled hepatocyte batch. 31 Thus, among the major P450 enzymes tested, merely CYP2D6 activity was lower in HμREL coculture, whereas CYP1A2, CYP2B6, CYP2C9, and CYP3A4 activities were on comparable levels with freshly thawed human hepatocytes published elsewhere. The maintained enzyme activity at high level is crucial for enabling clearance and metabolite formation of slowly metabolized drugs.…”

In vitro and in vivo methods to assess pharmacokinetic drug– drug interactions in drug discovery and development

“…In the study discussed above (Lu, Berg, et al, ), DDI was predicted accurately using a static model without the F g ’ /F g being incorporated (Equation ). Similar conclusions were reported in a few studies from various groups (Kosugi et al, ). This may be attributed by the fact that for high clearance compounds, excluding the GI contribution could be partially compensated for overpredicting hepatic DDI using a static model.…”

“…Predicting F g ’ /F g is not as straight forward as predicting the hepatic compartment due to: (1) GI [I] concentration (especially in the enterocyte) being not measurable due to the solubility of an inhibitor in the complex GI environment including various pH, although regulatory guidelines suggest using the dose/250 ml as the worst case scenario, (2) Since metabolism is the major mechanism for drug elimination in GI, the residence time, which is associated with the permeability of both inhibitor and victim drugs in enterocytes, can influence the degree of metabolism and DDI, and (3) P‐gp/BCRP substrates, which may go through inter‐GI recirculation, i.e., are absorbed into enterocytes and partially metabolized, then pumped out by P‐gp/BCRP, reabsorbed back to enterocytes and subject to another round of metabolism (Gan et al, ). There are some examples showing good success in DDI prediction by including the gut DDI (F g ’ /F g ), while others show that it is less successful (Galetin, Gertz, & Houston, ; Galetin, Hinton, Burt, Obach, & Houston, ; Kosugi et al, ).…”

Prediction of metabolizing enzyme-mediated clinical drug interactions using in vitro information

“…When the basic model does not rule out the DDI potential, mechanistic modeling can be used for further evaluation, and a static model is one of them. The static model approach is relatively simple and useful for preliminary mechanistic evaluation [ 9 10 ]. It was suggested by Fahmi et al [ 11 ] and also called as “Modified Rowland-Matin model” [ 12 13 14 ].…”