Assessing OATP Inhibition DDIs and Gaps in Early Risk Assessment

Owing to the significant role of OATP1B transporter in drug clearance, regulatory agencies, including the US Food and Drug Administration (FDA), the European Medicines Agency (EMA), and the Pharmaceuticals and Medical Devices Agency (PMDA) in Japan mandate investigating each new drug entity for their potential to be a DDI victim or perpetrator of OATP1B1 and 1B3 transporters. Per the prevailing regulatory guidelines, a new chemical entity’s potential for OATP-mediated drug interactions in the clinic can be detected through a DDI study utilizing a probe substrate of OATP such as pitavastatin or rosuvastatin. These guidance documents recommend using validated in vitro systems to explore whether an investigational drug is transported or inhibits OATP1B1 and/or OATP1B3 using clinically relevant substrates. Typically, the decision to conduct the expensive clinical DDI trials is based on an “R-value” obtained by calculating the ratio of unbound maximum portal vein inhibitor concentration in vivo for the new chemical entity against the in vitro OATP IC₅₀. The cutoff value of the “R-value” is conservative to minimize the possibility of a false-negative interaction, which can potentially generate a high rate of false-positive predictions and consequently trigger unnecessary expensive human DDI trials (Prueksaritanont et al., 2013; Tornio et al., 2019). The estimation is further hampered by several uncertainties, including reliable protein binding values for highly protein-bound molecules, the rate and fraction of oral absorption in humans, and the mechanisms of transporter inhibition, etc. In addition, overcoming the limitations of in vitro IC₅₀ assessment using different probe substrates, incubation conditions, and variability when using different expression systems remains a significant challenge. Furthermore, interpreting human DDI data can be difficult due to the lack of clean index substrates, interactions with multiple transporters, and transporter-enzyme interplays (Gopaul et al., 2021). Currently, transporter inhibitors or substrates that fulfill the criteria of DDI index drugs have not been identified (https://www.fda.gov/media/134582/download). The term “index DDI studies” used for most cytochrome P450 (P450)-metabolizing enzymes does not fit transporter DDI assessments. Investigating transporter DDIs are “concomitant-use DDI studies” based on the possibility of both drugs being administered concomitantly (https://www.fda.gov/media/134581/download). In this regard, the discovery of sensitive and specific endogenous biomarkers that can be measured during phase I dose escalation trials becomes attractive as a complementary method for the early assessment of OATP-mediated DDIs (Chu et al., 2018; Müller et al., 2018).

Moreover, it remains of great importance for the pharmaceutical industry to address potential DDI risks during the early stages of drug development, which allows for informed decisions about selecting compounds for clinical development before committing to costly late-phase clinical trials. However, choosing the right dose for DDI studies is not trivial when the clinically effective dose has yet to be established. If the dose selected is too low, there is a risk of false-negative outcomes, and controversially selecting a high dose for a DDI study can cause potential adverse effects. Clinical DDI assessment may also need to include subjects with specific OATP1B1 haplotypes due to the clinically relevant gene polymorphisms associated with this transporter. To mitigate these risks, the industry would benefit from the availability of sensitive and specific endogenous biomarkers that can be measured during phase I dose-escalation trials. These biomarkers could help identify potential PK interactions early on, allowing for better decision-making in drug development.

Coproporphyrins As Biomarkers of OATP1B Inhibition

Heme biosynthesis involves complex processes tightly regulated by various enzymes and transporters. As shown in Fig. 1, condensing glycine with succinyl-CoA produces 5-aminolevulinic acid and subsequently forms the pyrrole derivative porphobilinogen. Four porphobilinogens are connected to form a linear structure of preuroporphyrinogen, and then the cycle is closed to procedure uroporphyrinogen III tetrapyrrolic cycle. Coproporphyrinogen III is created through a few decarboxylation steps and further oxidized to turn into a dead-end product, coproporphyrin III (CP-III). Similarly, uroporphyrinogen I can be decarboxylated into coproporphyrinogen I, which undergoes spontaneous oxidation to form the other dead-end molecule, coproporphyrin I (CP-I) (Fig. 1). No physiologic role and further metabolites of CP-I and CP-III are found, and the produced CPs are eventually eliminated from the bile and urine in their unchanged forms (Kaplowitz et al., 1972). CP-III is the predominated isoform in the urine, and urinal CP-I comprises less than 45% of total CPs (sum of CP-I and III). Patients carrying a loss-of-function OATP1B gene (Rotor syndrome) (Stiel et al., 1982) or impaired multidrug resistance–associated protein (MRP) 2 functions due to genetic variants of ABCC2 gene (Dubin-Johnson syndrome) (Ben-Ezzer et al., 1971) show elevated urinal total CP levels and the abnormal ratio of CP-I (60.2%) over CP-III in the urine, whereas the biliary excretion of CPs is very low in Mrp2-deficient rats compared with normal rats (Moriondo et al., 2009). These clinical and preclinical observations suggest that CP-I and III biliary excretions likely rely on the hepatic OATP1B-MRP2 transporting axis. In theory, the plasma and urinal concentrations of CPs can be endogenous biomarkers for their functions.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Heme biosynthesis is initiated through condensing glycine with succinyl-CoA to produce 5-aminolevulinic acid and subsequently form porphobilinogen. Four porphobilinogens are connected to convert a linear structure of preuroporphyrinogen. Uroporphyrinogens I and III are formed by closing the cycle of preuroporphyrinogen and further decarboxylated to produce coproporphyrinogen I and III. Oxidation of coproporphyrinogen I and III forms CP-I and CP-III, respectively. The produced CPs are eventually eliminated from the bile and urine in their unchanged forms. OATP1B transports CP-I and III into the liver, and the inhibition of OATP1B can increase plasma CPs. Created with BioRender.com.

In in vitro assays, CP-I and CP-III are substrates for human OATP1B1, OATP1B3, MRP2, and MRP3 (Bednarczyk and Boiselle, 2016; Gilibili et al., 2017; Kunze et al., 2018) but not for glycoprotein (P-gp), the breast cancer resistance protein (BCRP), the bile salt export pump (BSEP), MRP4, the multidrug and toxin extrusion proteins (MATE) 1 and 2K, the organic cation transporters (OCT) 1 and 2, the organic anion transporters (OAT) 1–4, and the sodium taurocholate cotransporting polypeptide (NTCP) (Kunze et al., 2018). CP-III, but not CP-I, is a substrate for OATP2B1 (Shen et al., 2017). The data obtained from these in vitro characterizations shed light on the mechanisms involved, indicating that CP-I and III selectively target the OATP-MRP transport axis.

Toward the goal of discovering endogenous biomarkers for OATP1B inhibition, our group first reported in 2016 that levels of CP-I and CP-III were elevated in the plasma and urine of monkeys who had received OATP inhibitors such as cyclosporine and rifampin. The same increase was observed in mice with a knockout of the Oatp1b/2b gene, suggesting that CP-I and CP-III in both plasma and urine could reflect in vivo activity of hepatic OATP transport (Shen et al., 2016). Follow-up clinical trials then confirmed the potential of CP-I and CP-III as biomarkers, as changes in their plasma levels were found to accurately predict the impact of rifampin on rosuvastatin pharmacokinetics in human subjects (Lai et al., 2016). The administration of rifampin resulted in a 5.7- and 4.0-fold increase of CP-I and a 5.4- and 3.3-fold in plasma C_max and AUC_0–24 of CP-III, respectively, which were consistent with the DDI effects of rosuvastatin in combination with rifampin (13.2- and 5.0-fold, respectively) (Lai et al., 2016). These findings provide important insights into using CP-I and CP-III as biomarkers of OATP activity in preclinical and clinical settings.

Followed by our first report that reveals CPs as valuable biomarkers for monitoring OATP DDIs, the investigation of using CPs as DDI biomarkers for OATP1B inhibition has quickly expanded over the past few years. As shown in Fig. 2, in 2018, two years after our first clinical findings, several studies confirmed the utility of CPs, particularly CP-I, as a sensitive biomarker for predicting the DDI impact of OATP1B inhibitors. Shen et al. (2018) provided additional clinical evidence showing that ethnicity has no effect on basal plasma levels of CP-I and CP-III, and human subjects carrying SLCO1B1 c.388AG and c.388GG genotypes (i.e., increased OATP1B1 transport activity) have lower levels of CP-I than those with c.388AA, further supporting their use as biomarkers of OATP1B activities. A comparative study investigating the dose-dependent inhibition of OATP1B by rifampin has also found that changes in CP-I levels are dose dependent, consistent with OATP inhibition on the DDIs with sensitive drugs such as pitavastatin and atorvastatin (Takehara et al., 2018). Model-based simulations and power calculations have confirmed the ability of CP-I to identify moderate and weak OATP1B inhibitors in adequately powered clinical studies (Barnett et al., 2018). In addition, Yoshikado et al. (2018) conducted detailed analyses of the inhibitory effects of rifampin on hepatic OATP1Bs/MRP2-mediated transport of CP-I. The authors demonstrated the accurate prediction of the blood concentration-time profiles of OATP1B probe drugs affected by rifampin at two different dose levels in a physiologically based pharmacokinetic (PBPK) model after optimizing the value of in vivo concentration required to produce half maximum inhibition (Ki) for CP-I and the ratio of in vitro Ki value for each statin to CP-I. Liu and colleagues (2018) conducted a retrospective analysis of plasma CP-I and CP-III concentrations in samples obtained from DDI studies of pitavastatin and GDC-0810. They reported that the measurement of CPs could be useful in detecting weak inhibition of OATP1B by their investigational drug GDC-0810 (Liu et al., 2018). It is worth noting that the first inclusion of CP-I and CP-III as biomarkers of OATP1B inhibition in the 2018 international transporter consortium (ITC) white paper is a significant step toward their widespread use (Chu et al., 2018). The ITC paper further provides decision trees that illustrate the potential utility of endogenous biomarkers, both currently and in the future, for developing new drugs and personalized treatment strategies (Chu et al., 2018).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

The milestones achieved in the validation paths toward identifying CP-I as an endogenous biomarker for OATP1B activities.

As shown in Fig. 2, Barnett and colleagues (2019) conducted a comprehensive evaluation of the utility of 20 endogenous biomarkers in assessing the inhibition DDIs of OATP1B. The data obtained from the study confirmed that none of the biomarkers tested were superior to CP-I in the evaluation of DDIs involving OATP1B inhibition (Barnett et al., 2019). This highlights the crucial role that CP-I plays as a biomarker in such evaluations (Barnett, et al., 2019). Meanwhile, Yee and colleagues (2019) reported the first clinical evidence demonstrating a significant association between the SLCO1B1 c.521T>C genotype (which leads to a reduction of OATP1B1 function) and plasma levels of CP-I but not CP-III. Additionally, the study found that the effects of cyclosporin A inhibition on pravastatin and biomarkers were modulated by the OATP1B1 genotype. Although the baseline levels of CP-I are higher in subjects carrying SLCO1B c.521C, OATP1B inhibitor cyclosporine increases the plasma CP levels in all genotype groups (Yee et al., 2019). These findings provide valuable insights into the mechanisms underlying DDI involving OATP1B and underscore the importance of considering genetic factors when evaluating biomarkers in such studies (Yee et al., 2019).

Also shown in Fig. 2, Jones et al. (2020) conducted clinical DDI studies to investigate the interaction of fenebrutinib with various clearance pathways involving CYP3A, OATP1B, and BCRP. For the first time, the authors used CP-I and CP-III as endogenous biomarkers to elucidate the mechanisms of complex drug-drug interactions involving various clearance pathways. This approach allowed for a comprehensive assessment of the potential risks associated with these interactions and highlighted the importance of considering multiple clearance pathways in DDI studies (Jones et al., 2020). In addition, Mori and colleagues (2020a) reported the utility of CP-I and III as biomarkers in non–small cell lung cancer patients treated with paclitaxel. The use of these biomarkers helped to address practical and ethical issues associated with clinically assessing DDI risks for anticancer drugs that target OATP transporters (Mori et al., 2020a).

As indicated in Fig. 2, Neuvonen and colleagues (2021) conducted a comprehensive performance analysis of plasma CP-I and CP-III as biomarkers for assessing OATP1B activities in humans. They compared the performance of these biomarkers with that of two bile acid conjugates: glycodeoxycholate 3-O-glucuronides and glycochenodeoxycholic acid 3-O-glucuronide (GCDCA-3G). The authors reported that plasma CP-I was superior to CP-III in detecting differences in OATP1B functions across various subject groups with genetically poor, decreased, increased, and highly increased functions. Additionally, the study found that GCDCA-3G was even more sensitive than CP-I as a biomarker for assessing OATP1B1 activity. These findings provide valuable insights into the utility of CPs and bile acid conjugates as biomarkers to evaluate OATP1B activities in humans and underscore the importance of considering multiple biomarkers along with CP-I in such evaluations (Neuvonen et al., 2021). Furthermore, Sane et al. (2021) incorporated the CP-I and CP-III measurements in DDI studies of itraconazole and ipatasertib. They demonstrated that itraconazole or ipatasertib did not affect plasma CP-I and CP-III. The results confirmed the lack of in vivo OATP1B inhibition by either ipatasertib or itraconazole (Sane et al., 2021).

As shown in Fig. 2, Takita and colleagues (2022) were the first to demonstrate the utility of CP-I as an endogenous biomarker for detecting reduced OATP1B activity in patients with chronic kidney diseases. This finding is crucial, as it allows for accurate DDI risk assessment for drugs that are eliminated through both hepatic and renal pathways, thereby avoiding underestimation of OATP DDIs by extrapolating interaction data from healthy volunteers. These results highlight the importance of considering patient-specific factors in DDI studies and the potential value of CP-I as a biomarker for such evaluations (Takita et al., 2022). In another study, Lin and colleagues (2022) measured CP-I plasma levels in patients with hepatic impairments. They demonstrated that in vivo OATP1B activities, as measured by CP-I levels, were strongly associated with the severity of liver impairment (Lin et al., 2022). This finding suggests that CP-I data could be used to capture population variability within specific Child-Pugh classes and aid in individualized DDI risk assessment. These results provide valuable insights into the potential utility of CP-I as a biomarker for assessing hepatic function and informing DDI evaluations in patients with liver disease (Lin et al., 2022).

Additional Endogenous Biomarkers Associated with OATP1B Inhibition DDIs

Primary bile acids are produced through the oxidation of cholesterol, mainly by CYP7A1, and can be further conjugated in the liver before being secreted into the bile. About 95% of bile acids are reabsorbed through the enterohepatic circulation process. Secondary bile acids can be formed through the enzymes in gut microflora and reabsorbed. Cholic acid, chenodeoxycholate acid (CDCA), and deoxycholate acid (DCA), as well as their respective glycine and taurine conjugates such as glycochenodeoxycholate-3-sulfate (GCDCA-S) and chenodeoxycholate 3- or 24-glucuronide (CDCA-3G or -24G), are identified as OATP1B substrates. Several clinical studies explored the utility of bile acids as endogenous biomarkers for assessing DDIs involving OATP1B inhibition. For example, the dose-dependent effects of rifampicin (150, 300, or 600 mg) on several potential biomarkers suggest that several bile acids, including glycochenodeoxycholate- 3- sulfate (GCDCA-3S) and GCDCA-3G (Table 2), could potentially be used to assess OATP1B-mediated DDIs in humans (Mori et al., 2020b). Another study detected a 20.3-fold increase in plasma GCDCA-S after the administration of a single dose of rifampicin (600 mg, oral) (Takehara et al., 2017) (Table 2). Interestingly, these authors also found that the increase of GCDCA-S is associated with the changes of the GCDCA-S precursors GCDCA/glycodeoxycholic acid concentrations, which suggests that the increase of GCDCA-S by the rifampin treatment may involve bile acid homeostasis such as metabolism instead of OATP1B inhibition (Takehara et al., 2017). Many primary and secondary bile acids such as CDCA and GCDCA-S are also substrates for organic anion transporter (OAT) 3, NTCP, and MRP2 transporters (Nigam et al., 2020; Willemin et al., 2021). Probenecid treatment can increase plasma GCDCA-3S (Tsuruya et al., 2016; Bush et al., 2017) (Table 2). These increases in plasma bile acids by probenecid treatment in humans are also found in the changes in the Oat1 or Oat3 gene knockout mice, suggesting the inhibition of OAT1 and 3 (Granados et al., 2022). Thus, it is important to note that bile acids are also substrates for other transporters, which limits their use as a specific biomarker for OATP1B-mediated interactions.

Bilirubin is a byproduct of heme breakdown, and its uptake into liver cells relies at least partially on two transporters: OATP1B1 and OATP1B3. Once inside the liver, bilirubin is converted into bilirubin-glucuronide and bilirubin-di-glucuronide via uridine 5'-diphospho-glucuronosyltransferase 1A1 (UGT1A1)-mediated conjugation before most of it is excreted into bile by MRP2 [ABCC2]. Numerous preclinical and clinical studies have confirmed the vital role of OATP1B1 and OATP1B3 in the uptake of both conjugated and unconjugated bilirubin. In vitro studies using adult hepatocytes showed that OATP1B1 and OATP1B3 play a significant role in the uptake of unconjugated bilirubin. Carriers of impaired function variants in the genes encoding OATP1B1 and OATP1B3 (i.e., Rotor syndrome) have significantly higher levels of unconjugated bilirubin (van de Steeg et al., 2012). Bilirubin and its conjugated forms are investigated to be endogenous biomarkers to evaluate the potential for OATP1B-mediated DDIs. In healthy volunteers, administration of rifampin increased the concentrations of direct bilirubin in a dose-dependent manner, and the changes correlated well with changes in other OATP1B probe substrates (Table 2) such as atorvastatin, pitavastatin, and valsartan (Takehara et al., 2018; Mori et al., 2020b). However, the specificity of bilirubin as a biomarker can be diminished by the involvement of other bilirubin transporters such as MRP2, MRP3, and metabolizing enzymes (i.e., UGT1A1). In addition, the serum bilirubin level is clinically used as a biomarker for liver injury. Careful evaluation is required to distinguish between reversible inhibition of hepatic OATP1B, MRP2, MRP3, and/or UGT1A1 and liver injury.

Certain fatty acid dicarboxylates such as tetradecanedioate (TDA) and hexadecanedioate (HDA) are substrates for OATP1B, indicating that they could also be used as potential endogenous biomarkers of OATP1B (Table 2). HDA and TDA in plasma, along with CP-I, are higher in human subjects harboring SLCO1B1 c521T>C alleles with reduced transporter activity than other genotype groups (e.g., SLCO1B1 c521TT) (Yee et al., 2019). The OATP1B inhibitors rifampin or cyclosporin A can significantly increase the plasma AUC of TDA and HDA (Shen et al., 2017; Yee et al., 2019; Mori et al., 2020b), and the impact on HDA appeared to be dependent on the rifampicin dose (150, 300, or 600 mg) (Mori et al., 2020b). Notably, these dicarboxylic fatty acids are also found to be substrates of OAT1 and OAT3. The treatment of probenecid and furosemide also elevated the plasma HDA and TDA but not by the treatment of probenecid alone (Zhang et al., 2020), suggesting that other mechanisms may be involved in the elimination of HDA and TDA.

Unlocking Insights into Drug-Drug Interactions: The Utility of Plasma CP-I As an Endogenous Biomarker To Assess OATP1B Activity

As of the current writing, the most extensively studied biomarker for SLC is the CP-I, and 18 reports for different laboratories have been published since 2016 (https://www.druginteractionsolutions.org) on eight different clinical perpetrators that cause clinically relevant elevation [i.e., area under the curve ratio > 1.25] (Table 2). Although the initial clinical study on CP-I as an endogenous biomarker for OATP DDIs published in 2016 (Lai et al., 2016) is limited by small sample sizes, numerous subsequent trials have confirmed the reproducibility of the original discovery (Li et al., 2021). CP-I changes have been extensively tested in multiple studies with dose dependence of rifampin from 150 to 600 mg in healthy human subjects and showed agreement in DDI magnitudes of probe substrates (Table 2). CP-I exhibits selectivity toward OATP1B activity in human subjects with genetic variants of OATP1B1. Its sensitivity to a broad spectrum of clinical OATP1B inhibitors has been established from weak to vigorous. PBPK models set for concentration changes of plasma CP-I with OATP inhibitors demonstrate the usefulness of supporting the quantitative translation of the effect of CP-I levels into the DDI risk assessment of potential OATP inhibitors. As the sensitivity, specificity, and selectivity of plasma CP-I as an endogenous biomarker have been validated in humans, the evaluation of CP-I for OATP inhibition DDIs is deemed satisfactory to aid decision making when measured in early drug development. It is now time to adapt the decision tree for validated endogenous biomarkers proposed in the ITC white paper (Chu et al., 2018) for the OATP inhibition DDI to monitor CP-I levels in single and multiple clinical dose-ranging studies (phase I) for early assessment of DDI risks (Fig. 3). The approach could preclude the need for a dedicated DDI study by administering a probe substrate drug to human subjects. The dose-dependent changes of plasma CP-I can be further used to optimize a mechanistic mathematical model and mechanistic insights for complex DDIs involving metabolizing enzymes and/or other transporters to inform further assessment of DDI risks. The plasma CP-I data obtained from dose-escalation trials can warrant the full dose-response of an investigational drug to OATP inhibitions, which can provide the liberty of decision making for dose selection in later phase trials. Additional biomarkers reported thus far that their fidelities to serve as DDI biomarkers are not fully confirmed might also be monitored along with CP-I measurement for continuous efforts on OATP biomarker discovery.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

The utility of CP-I to assess OATP1B inhibition DDIs in early drug development and enable a mechanistic understanding of complex DDIs. The initial risk assessment for OATP1B inhibition DDI involves in vitro characterization using relevant substrates according to regulatory guidance. In single and multiple-dose escalation phase I clinical trials, monitoring plasma CP-I is used to identify potential OATP1B inhibition DDIs. Dose-dependent changes in plasma CP-I level are strong indications of OATP1B inhibition DDIs; In that case, PBPK modeling is used to assess further complex DDIs involving OATP inhibitors, and additional DDI trials may be conducted based on the analysis.

Future Perspectives

As illuminated in Fig. 3, dose-dependent plasma CP-I changes in phase I trials are strong indications of OATP1B inhibition DDIs. However, translating biomarker data directly to drug labels requires more work. The hepatic elimination of CP-I involves multiple transporters, and conditions such as substrate-dependent inhibition, multiple binding sites of interactions, and pharmacogenomic/pharmacokinetic studies in subjects with altered transporter function can be complex factors in translating biomarker data to the drug label. Therefore, quantitative modeling analysis and/or additional dedicated DDI assessment should be followed to define the degree of inhibition with clinically relevant substrates. A biomarker that is more selective to the transporter of interest can better predict DDI; therefore, additional research is warranted to practicality determine the effectiveness of CP-I as the OATP1B inhibition biomarker in different scenarios, and additional endogenous markers in assessing inhibitors with other hepatic uptake transporter activities such as NTCP and/or OATP2B1 are needed.