Abstract

The organic anion transporting polypeptide (OATP)2B1 [(gene: solute carrier organic anion transporter family member 2B1 (SLCO2B1)] is an uptake transporter that facilitates cellular accumulation of its substrates. Comparison of SLCO2B1+/+ knockin and rSlco2b1−/− knockout rats showed a higher expression of rCYP3A1 in the humanized animals. We hypothesize that humanization of OATP2B1 not only affects cellular uptake but also metabolic activity. To further investigate this hypothesis, we used SLCO2B1+/+ and rSlco2b1−/− rats and the OATP2B1 and rCYP3A1 substrate erlotinib, which is metabolized to OSI-420, for in vivo and ex vivo experiments. One hour after administration of a single dose of erlotinib, the knockin rats exhibited significantly lower erlotinib serum levels, but no change was observed in metabolite concentration or the OSI-420/erlotinib ratio. Similar results were obtained for liver tissue levels comparing SLCO2B1+/+ and rSlco2b1−/− rats. Liver microsomes isolated from the erlotinib-treated animals were characterized ex vivo for rCYP3A activity using testosterone, showing higher activity in the knockin rats. The contrary was observed when microsomes isolated from treatment-naïve animals were assessed for the metabolism of erlotinib to OSI-420. The latter is in contrast to the higher rCYP3A1 protein amount observed by western blot analysis in rat liver lysates and liver microsomes isolated from untreated rats. In summary, rats humanized for OATP2B1 showed higher expression of rCYP3A1 in liver and reduced serum levels of erlotinib but no change in the OSI-420/erlotinib ratio despite a lower OSI-420 formation in isolated liver microsomes. Studies with CYP3A-specific substrates are warranted to evaluate whether humanization affects not only rCYP3A1 expression but also metabolic activity in vivo.

SIGNIFICANCE STATEMENT Humanization of rats for the organic anion transporting polypeptide (OATP)2B1 increases rCYP3A1 expression and activity in liver. Using the OATP2B1/CYP3A-substrate erlotinib to assess the resulting phenotype, we observed lower erlotinib serum and liver concentrations but no impact on the liver/serum ratio. Moreover, there was no difference in the OSI-420/erlotinib ratio comparing humanized and knockout rats, suggesting that OSI-420 is not applicable to monitor differences in rCYP3A1 expression as supported by data from ex vivo experiments with rat liver microsomes.

Introduction

The organic anion transporting polypeptide (OATP)2B1 is a transporter that facilitates cellular uptake and is known to be expressed in pharmacokinetically relevant organs such as the intestine (Kobayashi et al., 2003), the kidney (Ferreira et al., 2018), and the liver (Tamai et al., 2000). Despite various reports indicating that various drugs are substrates of OATP2B1, there is still limited understanding of the transporter’s involvement in drug disposition (Kinzi et al., 2021). To investigate the transporter’s function in vivo, we have generated the rSlco2b1−/− rats, which are lacking the expression of rOATP2B1, and the SLCO2B1+/+ rats, which are humanized for OATP2B1 (Kinzi et al., 2024a). The latter humanized model was established as we had previously shown that there are pronounced differences in substrate recognition when comparing the transport function of the rat and the human ortholog in vitro (Hussner et al., 2021). In an initial phenotyping of the rat models with atorvastatin, a high affinity substrate of OATP2B1 (Grube et al., 2006), it was shown, that humanization of the transporter is associated with a 40% reduction in systemic exposure [area under the curve (AUC)] and a 57% increase in clearance of the statin (Kinzi et al., 2022, 2024b). Hepatic tissue content 1 hour after tail vein injection was significantly lower in the SLCO2B1 knockin than in the rSlco2b1 knockout rats (Kinzi et al., 2022), despite the expected role of OATP2B1 as a hepatic drug uptake mechanism.

In humans, atorvastatin is not only a substrate of transporters but also metabolized by CYP3A4 (Chong et al., 2001). This phase I metabolizing enzyme is involved in the metabolism of about 30% of the drugs used in clinics (Zanger and Schwab, 2013). In rats, atorvastatin also undergoes hydroxylation (Black et al., 1999) most likely mediated by the rat orthologs of CYP3A4 of which especially rCYP3A1 and rCYP3A2 are considered to be expressed in the rat’s liver (Takara et al., 2003). Accordingly, it was proposed that one explanation for the reduced hepatic content of atorvastatin in the livers of SLCO2B1+/+ rats might be the observed higher rCYP3A1 protein abundance in the liver of the humanized animals (Kinzi et al., 2022, 2024b). Unfortunately, this previous phenotyping study with atorvastatin in SLCO2B1+/+ and rSlco2b1−/− rats was not designed to monitor the metabolism of atorvastatin in the treated rats.

As mentioned before, rCYP3A1 (gene CYP3A23-3A1) is one of the orthologs of CYP3A4 in rats. Rat CYP3A1 and CYP3A2 are expressed in the liver of Wistar rats (Kishida et al., 2008) and both are inducible, which means that their expression and activity is transcriptionally regulated (Gibson et al., 2002). Changes in CYP3A expression have been shown to translate into higher 6β-hydroxylation of testosterone, 3-hydroxylation of diazepam, and hydroxylation of midazolam in rats (Eeckhoudt et al., 2002; Kishida et al., 2008), supporting the notion that the rCYP3A enzymes exhibit a comparable substrate spectrum as CYP3A4 (Zanger and Schwab, 2013).

Another drug whose metabolism involves OATP2B1 and CYP3A4 is erlotinib. Indeed, erlotinib is extensively metabolized (Ling et al., 2006), and the formation of its active metabolite OSI-420 (desmethyl-erlotinib) is catalyzed by CYP3A4 (Hidalgo and Bloedow, 2003; Ling et al., 2006; Abdelgalil et al., 2020). This metabolite is also formed in rats (Thappali et al., 2012). Moreover, the tyrosine kinase inhibitor is a substrate of OATP2B1 (Bauer et al., 2018) and has previously been investigated in genetically modified mouse models. In a PET-study with [11C]-erlotinib, mSlco2b1−/− mice showed a reduced hepatic uptake clearance, however, with unchanged concentrations in blood (Marie et al., 2021). In another work, Li et al. (2023) studied the pharmacokinetics of erlotinib and OSI-420 after oral administration in genetically modified mice. Comparing systemic exposure (erlotinib AUC0–4h) in wild-type and mSlco2b1−/− knockout animals, they observed no significant difference after oral administration of neither the parent compound nor the metabolite. However, in mice with hepatoselective humanization of SLCO2B1, the authors observed a higher liver/plasma ratio of the metabolite after oral administration of erlotinib (Li et al., 2023).

In the herein presented study, we first intended to verify our previous observation on changes in hepatic rCYP3A1 expression comparing SLCO2B1+/+ and rSlco2b1−/− rats using additional animals. Then, we conducted a tissue distribution study after a single intravenous dose of erlotinib and measured the serum and tissue content of erlotinib and its metabolite OSI-420. Microsomes isolated from livers of the erlotinib-treated animals were tested for rCYP3A1 expression and activity using testosterone as a probe substrate. Moreover, we quantified the rCYP3A1 content in liver microsomes isolated from treatment naïve rats and assessed the OSI-420 formation rate.

Materials and MethodsAnimals.

All procedures applied to animals were reviewed and approved by the cantonal veterinary authority of Basel-Stadt, Switzerland. The livers that were used in this study were harvested under the licenses 32054-3056 and 35459-3184. The erlotinib distribution was performed under the license 33885-3124. The outbred Wistar rats used in the study were generated from heterozygous knockout rSlco2b1−/wt and humanized SLCO2B1+/wt as described elsewhere (Kinzi et al., 2024a). The rats were bred at the Animal Facility of the University of Basel. Experiments were conducted following the ARRIVE 2.0 guideline for care and use of laboratory animals (Percie du Sert et al., 2020) and the CRUS Policy for Animal Research principles by Swiss universities (CRUS, 2013: https://www.swissuniversities.ch/fileadmin/swissuniversities/Dokumente/Kammern/Kammer_UH/Empfehlungen/CRUS_Grundsatze_tierexpForschung_170113_e.pdf). Animals from the in-house breeding were identified by a continuous number independent of the genotype. Animal assignment to the experimental procedures was randomized. The assignment of the number for identification to the genotypes was not available to the experimenters. Group sizes in the experiments were first estimated using the resource equation approach with a degree of freedom between 10 (minimum) and 20 (maximum) for the error term in an analysis of variance (ANOVA) (Arifin and Zahiruddin, 2017). The calculation resulted in an expected number of 6 to 11 animals per group. We selected six animals per group for the pharmacokinetic study, as this group size was expected to allow us to detect differences of around 50%, assuming a standard deviation of 30% with the type 1 error (p) at the level of 5% with a power of the study at 80%.

Erlotinib Distribution Study.

rSlco2b1−/− and SLCO2B1+/+ rats were housed until the age of 10–12 weeks in a 12-hour light/dark cycle with controlled temperature (21–22°C) and water and food ad libitum. On the day of the experiment, the homozygous male rats were dosed intravenously with 12.4 mg/kg erlotinib (diluted in 0.9%-NaCl/DMSO/Tween 80; v/v/v: 85/10/5). One hour after tail vein injection, the animals were euthanized using CO2, immediately followed by blood and organ harvest. Cardiac blood was incubated for 20 minutes at room temperature (RT) and subsequently centrifuged at 3000 × g for 15 minutes. The supernatant was transferred to a new tube, and the serum was stored at −80°C until analysis. Tissues were harvested, washed with sterile 0.9% NaCl, weighed, and frozen in liquid nitrogen. Tissue samples were stored at −80°C until further use.

Total RNA Isolation and Real-Time Quantitative Polymerase Chain Reaction.

For total RNA extraction frozen liver samples of SLCO2B1+/+ and rSlco2b1−/− rats were milled using the Mixer Mill MM400 (Retsch GmbH, Düsseldorf, Germany), followed by RNA extraction applying the TRI Reagent Protocol (Sigma-Aldrich, Buchs, Switzerland) according to the manufacturer’s instruction. After quality and content assessment using the NanoDrop OneC (Thermo Fisher Scientific, Reinach, Switzerland), the mRNA of the tissue (1000 ng/reaction) was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). Ten nanograms of the resulting cDNA was applied to quantitative real-time polymerase chain reaction (qPCR). For quantification of the copy number, a standard curve was generated by serial dilution of a plasmid containing the subcloned qPCR-amplicon of rCyp3a1 [(108 base pairs (bp)]. Briefly, after amplification the polymerase chain reaction (PCR) amplicon was ligated into pDrive (Qiagen, Hilden, Germany). After amplification in Escherichia coli and isolation of the plasmid, the insert’s sequence was controlled by Sanger Sequencing (Microsynth, Balgach, Switzerland). The standard curve used during the qPCR ranged from 4.68 × 102 to 4.68 × 106 plasmid copies per reaction. The qPCR reaction was performed using gene-specific primers for rCyp3a1 (forward: 5′-TCTGTGCAGAAGCATCGAGT-3′; reverse: 5′-GGCTGTGATCTCCATATCG-3′), the Power SYBR Green PCR Master Mix, and the QuantStudio 5 (Thermo Fisher Scientific). Linear regression analysis of the standards was applied to evaluate the copy number of the respective transcripts.

Membrane Fraction Enrichment from Rat Liver.

To prepare the protein lysate, approximately 50 μg of the liver tissue was mixed with 1 ml subcellular fractionation (SF) buffer (250 mM sucrose, 20 mM HEPES pH 7.4, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 1 mM PMSF, 5 μg/ml leupeptin, 2 μg/ml aprotinin). Livers were cut into pieces with scissors, and samples were placed on ice for 15 minutes and vortexed every 5 minutes. For homogenization, 30 strokes were performed with the Dounce tissue grinder (Wheaton, Millville, NJ), followed by a 15-minute agitation on ice with a vortex every 5 minutes and a final sonication (three cycles, on ice) with the Ultrasonic Processor UP200S (Hielscher Ultrasonics, Teltow, Germany). After centrifugation at 720 × g and 4°C for 5 minutes, the supernatant was collected and centrifuged at 1000 × g and 4°C for 10 minutes. Then, the supernatant containing the cell lysate was centrifuged at 100,000 × g and 4°C for 1 hour to enrich the cellular membrane fraction. The pellet was resuspended in 500 μl SF buffer, and after a second ultracentrifugation, the pellet was dissolved in 50 μl Lysis buffer containing protease inhibitors (50 mM Tris HCl pH 8.0, 150 mM NaCl, 0.5 sodium deoxycholate 0.1% SDS, 1% Triton X-100, 10% glycerol, 1 mM PMSF, 5 μg/ml leupeptin, 2 μg/ml aprotinin) and sonicated three times on ice. The protein content was measured using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) and the TECAN Infinite M200 Pro (Tecan, Männedorf, Switzerland) according to the manufacturer’s instructions.

Preparation of Microsomes from Rat Liver Tissue.

Liver tissue (∼5 g) of either untreated rats or erlotinib-treated animals of the distribution study was thawed on ice, cut into small pieces, and rinsed in Tris/KCl buffer (20 mM Tris base/0.15 M KCl, pH 7.4) before homogenization in 10 ml Tris/KCl buffer using the Homogenizer Potter S (Sartorius, Göttingen, Germany). The homogenate was centrifuged at 10,000 × g and 4°C for 20 minutes. After removal of the fat layer, the supernatant was centrifuged at 100,000 × g and 4°C for 1 hour. Subsequently, the pellet was suspended in 0.15 M KCl buffer and homogenized using the Polytron PT1200E (Kinematica AG, Malters, Switzerland) before an additional centrifugation at 100,000 × g and 4°C for 1 hour. The resulting pellet containing the liver microsomal fraction was suspended in 20 mM Tris/HCl pH 7.4/0.25 M sucrose using the Polytron prior to the storage at −80°C until further experiments. Protein content was quantified as described above.

Western Blot Analysis.

Twenty micrograms of protein was supplemented with Laemmli buffer and incubated for 1 hour at 37°C with gentle agitation prior to separation by SDS-10% PAGE using the Mini-PROTEAN Tetra cell system (Bio-Rad Laboratories AG, Cressier, Switzerland). After separation, the proteins were transferred to a nitrocellulose membrane using the eBlot L1 system (GenScript, Rijswijk, Netherlands). After blocking with 5% Blotto in TBST (140 mM NaCl, 2.6 mM KCl, 25 mM Tris, 0.1% Tween 20, pH 7.4) for 1 hour at RT, the membranes were incubated with the anti-CYP3A1 (1:10,000, ab1253; Merck Millipore, Burlington, MA) or the anti-GAPDH (1:1,000; STJ96417; St John’s Laboratory, London, UK) antibody diluted in Signal Boost Immunoreaction Enhancer Kit (Merck Millipore) overnight at 4°C, followed by three washings with TBST. The subsequent incubation with the horseradish peroxidase (HRP)-labeled anti-rabbit antibody (Bio-Rad Laboratories) was performed for 1 hour at RT. For visualization of the immobilized HRP-labeled secondary antibody, the Clarity Western ECL substrate (Bio-Rad Laboratories) was used. Chemiluminescence was detected using the ChemiDoc MP Imaging System equipped with the Image Laboratory 5.0 software (Bio-Rad Laboratories AG).

CYP3A1 Activity Measurement.

For the determination of the metabolic activity in rat liver microsomes, we quantified the formation of OSI-420 from erlotinib or of 2β- and 6β-hydroxytestosterone from testosterone. For in vitro activity measurement, an incubation mixture consisting of 100 μl of the microsomal suspension (1 mg protein/ml) and 430 μl incubation buffer was prepared. The incubation buffer was phosphate buffer (50 mM Na2HPO4 adjusted to pH 7.4 with 50 mM KH2PO4) supplemented with MgCl2 (5.0 mM), EDTA (1 mM), NADP (1 mM), glucose-6-phosphate (5 mM), and glucose-6-dehydrogenase (1.7 U/ml). The reaction was initiated by adding 20 μl erlotinib (25 μM) or testosterone (100 μM) and placing the samples in a shaking water bath at 37°C and 160 rpm for 15 minutes.

For erlotinib, the reaction was terminated by adding an equal volume of acetonitrile/methanol (ACN/MeOH) (50:50, v/v) containing 0.5 ng/ml d6-erlotinib as an internal standard (IS). The resulting mixture was left on ice for 1 hour, followed by a centrifugation at 15,000 × g and 4°C for 10 minutes, and the supernatant was sonicated for 3 minutes and was then transferred to high-performance liquid chromatography (HPLC) vials for injection. The high-performance liquid chromatography tandem mass spectrometry (HPLC-MS/MS) calibration standards were prepared by mixing 50 μl ACN/MeOH (50:50, v/v) working solutions containing 5 ng/ml d6-erlotinib, and erlotinib and OSI-420 in concentrations ranging from 10 to 40,000 ng/ml with 430 μl incubation buffer, 50 μl mixed microsomes, and 20 μl MeOH. For precipitation 450 μl ACN/MeOH (50:50, v/v) was added, and then the calibrants were treated as described for the samples above. Erlotinib and OSI-420 were detected by HPLC-MS/MS as described below.

For testosterone, the reaction was stopped by adding 200 μl methanol and transferring the samples on ice. Testosterone and its metabolites 2β- and 6β-hydroxytestosterone were extracted from the incubation mixture using chloroform (6 ml). After evaporation under nitrogen flow, the residue was dissolved in 100 μl MeOH/water (50:50, v/v) for the HPLC measurement. Ten microliters was injected for chromatographic separation on a C18 SunFire (Waters) column (3.5 μm, 3.0 × 150 mm) with a precolumn (VanGuard; Waters) at a column temperature of 40°C and a gradient of the mobile phase at a flow rate of 0.5 ml/min. The gradient started after 2 minutes. The mobile phase consisted of MeOH/water/ACN (39:60:1, v/v/v) (A) and MeOH/water/ACN (80:18:2, v/v/v) (B) with a gradient of 0% to 70% B (2–6.5 minutes) from 70% to 80% B (from 6.5–9.5 minutes), then kept at 80% B (from 9.5–18 minutes) and back to 100% A (from 18.01 to 20 minutes). Absorbance of testosterone and its metabolites was measured at a wavelength of 254 nm. Retention times for the reaction products were 8.6, 10.4, and 11.6 minutes for 6β-hydroxytestosterone, 2β-hydroxytestosterone, and testosterone, respectively. Substrate and metabolite concentrations were calculated applying a standard curve ranging from 5 to 200 μM.

Quantification of Erlotinib in Rat Serum and Tissue by HPLC-MS/MS.

For quantification of erlotinib, a validated HPLC-MS/MS method was applied (Rysz et al., 2023). To obtain calibration standards, 20 μl blank rat serum (Lucerna-Chem, Lucerne, Switzerland) was supplemented with 10 μl of the working solution and precipitated with 90 μl ACN/MeOH (50:50, v/v) vortexed for 10 seconds, followed by centrifugation for 10 minutes at 16,000 × g and 4°C. The final range of the calibration standards was between 5 ng/ml and 20 μg/ml. Frozen tissue samples were pulverized using the Retsch MM400 Mixer Mill, and 20 mg tissue powder was supplemented with 100 μl ACN/MeOH (50:50, v/v) containing IS and immediately vortexed for 10 seconds. Then, the tissue samples were centrifuged as described above. Serum samples obtained from animals were prepared in the same manner: 20 μl serum was precipitated with 100 μl ACN/MeOH (50:50, containing IS), which was followed by vortexing for 10 seconds and centrifugation for 10 minutes at 16,000 × g and 4°C. The resulting supernatant was transferred to micro-HPLC vials for injection.

Statistical Analysis.

Data were analyzed using GraphPad Prism Version 9.3.1 for Windows (GraphPad Software, San Diego, CA). Prior to statistical analysis, the data linked to the animal IDs were assigned to the different genotype groups; however, the identity of groups was blinded during statistical analysis. For data with Gaussian distribution with unequal variances, statistical analysis was conducted using the Welch’s adjusted t test. For data that were normalized or did not have a normal distribution, the nonparametric Mann-Whitney U test was applied. The respective statistical test used is specified in the context of the presentation of the results. P values <0.05 were considered statistically significant (marked with *). Data are presented as mean ± S.D. if not otherwise stated.

ResultsComparison of rCYP3A1 Expression in the Liver of SLCO2B1+/+ and rSlco2b1−/− Rats.

To validate the previous findings showing alterations in rCyp3a1 mRNA expression and rCYP3A1 abundance in the liver of humanized rats, we performed quantitative real-time PCR and western blot analysis. Despite our earlier finding on significantly higher levels of rCyp3a1 mRNA (Kinzi et al., 2022), we are now reporting comparable numbers of rCyp3a1 mRNA copies in rat liver of both genotypes [rCyp3a1 copies/10 ng cDNA; median (95% confidence interval, CI); SLCO2B1+/+ vs. rSlco2b1−/−; 1.20 × 105 (0.96 × 105 to 1.49 × 105) vs. 1.49 × 105 (1.17 × 105 to 2.19 × 105); n = 5; Mann-Whitney U test; Fig. 1A]. Probing the enriched liver membrane fraction of five treatment-naïve animals of each genotype for the presence of rCYP3A1, we again observed significantly higher levels of the enzyme in the livers of SLCO2B1+/+ rats [rCYP3A1/GAPDH ratio; median (95% CI); SLCO2B1+/+ vs. rSlco2b1−/−; 1.021 (0.59 to 1.63) vs. 0.439 (0.40 to 0.82); Fig. 1, B and C].

Fig. 1.

Detection of rCyp3a1 mRNA and rCYP3A1 protein in the livers of Slco2b1−/− and SLCO2B1+/+ rats. The number of rCyp3a1 mRNA copies was quantified by real-time PCR applying a serial dilution of the cloned PCR amplicon and analyzing the data by linear regression (A). rCYP3A1 protein expression is reported as rCYP3A1/GAPDH ratio calculated normalizing the band intensity of rCYP3A1 to that of GAPDH. Band intensity was determined by densitometry (B). The rCYP3A1 and the GAPDH protein were detected by western blot analysis (C). Data of each individual rat (n = 5 animals per group) is shown, and the median and 95% confidence interval is indicated. For statistical analysis the Mann-Whitney U test was applied. *P < 0.05.

Serum and Tissue Levels after a Single Intravenous Dose of Erlotinib.

Measuring serum concentrations of the OATP2B1 substrate erlotinib 1 hour after intravenous administration, we observed significantly lower serum levels of erlotinib in the humanized animals [mean ± S.D. (ng/ml); SLCO2B1+/+ vs. rSlco2b1−/−; 1866.11 ± 1577.23 vs. 4044.54 ± 955.78; Fig. 2B]. This finding was in parallel with our expectations of OATP2B1 functioning as hepatic uptake mechanism. Surprisingly, we also observed a tendency toward lower serum concentrations of OSI-420 [mean ± S.D. (ng/ml); 502.107 ± 539.27 vs. 1028.82 ± 94.66; Fig. 2C], translating into comparable OSI-420/erlotinib ratios in both groups [median (95% CI); 0.160 (0.120 to 0.440) vs. 0.240 (0.200 to 0.360); Fig. 2D]. Considering that OSI-420 formation is mediated by CYP3A, this result suggested that there is no difference in metabolic capacity in rSlco2b1−/− and SLCO2B1+/+ rats, despite the higher CYP3A1 protein content in livers of SLCO2B1+/+ rats. Irrespective of that notion, we quantified the amount of erlotinib and OSI-420 in tissues shown to express the human transporter (Kinzi et al., 2024a). Of the organs analyzed, the liver exhibited the highest content of erlotinib or OSI-420, respectively (compare Fig. 3 and Table 1). In the livers, we also observed significantly lower levels of erlotinib (Fig. 3A) and OSI-420 (Fig. 3D) in animals expressing the human transporter compared with knockout animals. This difference remained when we considered the total organ weight (Fig. 3, B and E). No statistically significant difference was observed when we calculated the liver/serum ratio for erlotinib or OSI-420 (Fig. 3, C and F), suggesting that hOATP2B1 does not affect hepatic erlotinib or OSI-420 uptake. Moreover, there was no difference in the metabolism as indicated by the OSI-420/erlotinib ratio in the rats’ livers [median (95% CI); SLCO2B1+/+ vs. rSlco2b1−/−; 0.190 (0.120 to 0.520) vs. 0.270 (0.240 to 0.340)]. For all other tissues analyzed, we observed no difference in the content of erlotinib or OSI-420 comparing rSlco2b1−/− and SLCO2B1+/+ rats (Table 1). In the kidney samples, we found slightly higher serum-to-tissue ratios for both molecules in SLCO2B1+/+ rats compared with the knockout animals.

Fig. 2.

Quantification of erlotinib and its major metabolite OSI-420 in serum of rSlco2b1−/− and SLCO2B1+/+ rats after a single-dose application of erlotinib. Animals were administered a single intravenous dose of erlotinib, and samples were collected after 1 hour for analysis by HPLC-MS/MS (A). The image in (A) was created with BioRender.com. Quantification of erlotinib (B) and OSI-420 (C) in serum samples collected from SLCO2B1+/+ and rSlco2b1−/− rats. Data of each individual rat is shown, and the mean ± S.D. is indicated. *P < 0.05 unpaired t test with Welch’s correction. (D) The OSI-420/erlotinib ratio was calculated for each animal. The median with 95% CI (D) is indicated. For statistical analysis the Mann-Whitney U test was applied.

Fig. 3.

Erlotinib and OSI-420 liver content in SLCO2B1+/+ and rSlco2b1−/− rats. Erlotinib (A) or OSI-420 (D) were measured in the liver 1 hour after intravenous application of 12.4 mg/kg erlotinib. The animal’s wet liver tissue weight was used to calculate the total erlotinib (B) or OSI-420 (E) amount in each animal. Each individual data point is shown, and the mean ± S.D. is indicated. For statistical analysis an unpaired t test with Welch’s correction was applied. *P < 0.05. (C and F) show the liver-to-serum ratios of erlotinib (C) and OSI-420 (F), respectively. Next to each individual data point, the median and 95% CI are displayed. For statistical analysis the Mann-Whitney U test was used.

TABLE 1

Summary of the erlotinib and OSI-420 content in the different organs The tissue content was determined by HPLC-MS/MS in pulverized tissue samples. The tissue content was normalized to that observed in serum of the same animal. Data on tissue content are reported as mean ± S.D. of n = 5 to 6 animals per group; for statistical analysis an unpaired t test with Welch’s correction was used. Ratios are reported as median with 95% confidence interval] of n = 5 to 6 animals per group; for statistical analysis a Mann-Whitney U test was performed.

rCYP3A1 Assessment in Erlotinib-Treated Animals.

Measuring tissue content, we found comparable OSI-420/erlotinib ratios in all organs irrespective of genotype despite our observation on higher levels of the metabolizing enzyme in the livers of untreated SLCO2B1+/+ rats. To evaluate rCYP3A1 expression and activity in the rats treated with erlotinib, we isolated liver microsomes from the treated animals of the distribution study and used them to quantify the amount of rCYP3A1 protein (Fig. 4A). As shown in Fig. 4C, rCYP3A1 was readily detected in all samples. However, densitometric analysis revealed no difference in rCYP3A1 protein amount in the erlotinib-treated animals, even if there was a trend toward lower levels in the knockout rats [Fig. 4B; rCYP3A1/GAPDH ratio; median (95% CI); SLCO2B1+/+ vs. rSlco2b1−/−; 0.972 (0.437 to 1.857) vs. 0.621 (0.213 to 1.325); n = 5 animals per group; P = 0.111; Mann-Whitney U test]. Finally, we determined the metabolic activity of CYP3A enzymes in the liver microsomes using testosterone as a substrate. By quantifying the formation of 6β-hydroxytestosterone, a higher rCYP3A activity was detected in microsomes isolated from the erlotinib-treated SLCO2B1+/+ rats (Fig. 4D). Similar results were obtained for 2β-hydroxytestosterone (Fig. 4E). Data on protein amount and metabolic activity were not correlated (Supplemental Fig. 1).

Fig. 4.

Rat CYP3A1 expression and activity in erlotinib-treated animals. Microsomes were isolated from the livers collected of animals intravenously dosed with erlotinib as shown in (A). The image was created with BioRender.com. The samples were first applied to a western blot analysis where band intensity was analyzed by densitometry. The rCYP3A1 protein expression is reported as rCYP3A1/GAPDH ratio showing the data of each individual rat and indicating the median and 95% CI; Mann-Whitney U test (B). In (C), we show the results of the western blot where microsomes were probed for rCYP3A1 or GAPDH, respectively. Liver microsomes were used to assess the formation of 6β-hydroxytestosterone (6β-OH-TST, D) and 2β-hydroxytestosterone (2β-OH-TST, E) using testosterone at a concentration of 100 μM. Data are reported as reaction rate, which is calculated considering the amount of microsomal protein and time of reaction. Shown is the reaction rate for each individual rat and indicated is the mean ± S.D. For statistical analysis the unpaired t test with Welch’s correction was used. *P < 0.05.

OSI-420 Formation in Microsomes Isolated from Livers of Treatment-Naïve Rats.

The ex vivo activity of rCYP3A1 measured with testosterone in the microsomes isolated from erlotinib-treated rats may be affected by erlotinib in the sample. Accordingly, we collected livers from treatment-naïve rSlco2b1−/− and SLCO2B1+/+ rats for preparation of microsomes. Probing these microsomes for rCYP3A1 content by western blot analysis revealed again significantly higher levels of the enzyme in the animals expressing the human transporter (Fig. 5, A and B). Subsequently, we determined the erlotinib demethylating activity in the liver microsomes (Fig. 5C). Surprisingly, we learned that microsomes isolated from treatment-naïve knockout animals exhibit significantly higher metabolic activity for erlotinib as determined by assessing the formation of OSI-420 (desmethyl-erlotinib) (Fig. 5D). This translated into a higher OSI-420/erlotinib ratio in the samples isolated from rSlco2b1−/− rats (Fig. 5E) and may be explained by contribution of another enzyme.

Fig. 5.

rCYP3A1 protein amount in microsomes isolated from untreated rSlco2b1−/− and SLCO2B1+/+ animals and their metabolic activity for erlotinib. (A) Microsomes were prepared from the liver of SLCO2B1+/+ and rSlco2b1−/− rats and probed for rCYP3A1 by western blot analysis. GAPDH served as loading control. (B) Band intensity was analyzed by densitometry. Expression is reported as rCYP3A1/GAPDH ratio showing the data of each individual rat and indicating the median and 95% CI. *P < 0.05; Mann-Whitney U test. (C) Erlotinib is demethylated to OSI-420; this reaction is catalyzed by CYP3A enzymes. (D) OSI-420 reaction rate was determined using the rat liver microsomes. *P < 0.05; unpaired t test with Welch’s correction. (E) Quantities measured by HPLC-MS/MS were used to calculate the OSI-420/erlotinib ratio. Data points of each individual rat are shown, and indicated is the median and 95% CI. *P < 0.05; Mann-Whitney U test. Chemical structures in C were generated using Marvin JS.

Discussion

In this study, we showed that the amount of rCYP3A1 protein is higher in the livers of animals expressing the human transporter compared with rSlco2b1 knockout rats. This was shown in liver tissue lysates (Fig. 1) and in liver microsomes isolated from treatment-naïve rats (Fig. 5), thereby extending the finding from a previous study (Kinzi et al., 2022). In this particular study, it was shown that the increase in rCYP3A1 protein in liver lysate is limited to humanized animals, suggesting that only the human OATP2B1 governs entry of an endogenous rCYP3A1 regulating molecule. Considering that there are pronounced differences in substrate recognition where the rodent ortholog does not transport the endogenous sulfated steroid conjugates (Hussner et al., 2021), this notion appears plausible.

We also quantified the number of mRNA copies of rCyp3a1 in untreated animals and observed no difference, even if we previously reported significantly higher levels of rCyp3a1 in another group of animals of the same outbred strain [mean normalized levels ± S.E.M.; SLCO2B1+/+ vs. rSlco2b1−/− 3.557 ± 0.928 vs. 1.120 ± 0.347; Kruskal-Wallis test (Kinzi et al., 2022)], performing a real-time PCR experiment analyzed using the 2−ΔΔCT method and three housekeeping genes (Livak and Schmittgen, 2001). Comparing mRNA expression of mCyp3a enzymes detected by RNA sequencing analysis in wild-type mice and mice with liver specific expression of OATP2B1 (Slco12B1-Apo), Li et al. (2023) reported downregulation of most isoforms in mouse liver. However, compared with the knockout strain (mSlco1a/1b/2b1−/−), no change in hepatic mRNA expression of the mCyp3a isoforms was observed. For the Slco12B1-Vil mice expressing OATP2B1 specifically in the intestine, the authors report upregulation of the liver mCyp3a mRNA in comparison with wild-type animals (Li et al., 2023). Whether this supports or contradicts our findings in humanized rats is challenging to evaluate, as their humanized mouse model is based on the mSlco1a/1b/2b1−/− mouse strain.

Hence, testing activity in liver microsomes isolated from the erlotinib-treated animals, we observed a slightly higher formation of 6β-hydroxytestosterone and 2β-hydroxytestosterone, even though there was only a trend toward higher protein expression in the liver microsomes isolated from the erlotinib-treated SLCO2B1+/+ rats from our distribution study (Fig. 4). Especially 6β-hydroxytestosterone formation is applied when assessing CYP3A activity in rat microsomal studies (Ibrahim et al., 2000; Baati et al., 2012), and higher formation of this metabolite is considered indicative for a higher CYP3A activity and content, which would support our initial observation in the SLCO2B1+/+ rats. However, one limitation of the experiments with the liver microsomes is that we did not control for losses during fractionation in the preparation of the liver microsomes.

In this study, we isolated the liver microsomes, which we showed to exhibit differential testosterone hydroxylation rates, from rats that had been treated with erlotinib, and the tyrosine kinase inhibitor was still present in the liver at the time of organ collection (Fig. 3). Importantly, there is a report suggesting that erlotinib is a time-dependent inhibitor of human CYP3A4 (Dong et al., 2011). Whether this also applies to the rat isoforms remains to be determined. Especially for mechanism-based inhibitors, which are time-dependent, the inhibitory effect is considered irreversible or quasi-irreversible (Deodhar et al., 2020). Hence, considering that we detected higher erlotinib concentrations in the livers of the rSlco2b1−/− rats (Fig. 3, A and B), one could argue that the lower rCYP3A activity observed in the liver microsomes isolated from those rSlco2b1−/− rats is due to the higher exposure of the enzyme to the potential inhibitor. Nevertheless, there is a conflicting report on the function of erlotinib as a cytochrome P450 (P450) inhibitor. Indeed, Li et al. (2007) showed that erlotinib stimulated the CYP3A4-mediated formation of 1’-hydroxymidazolam in human microsomes. However, it remains speculative whether presence of erlotinib modulates the hydroxylation of testosterone by rodent P450 enzymes, but it could have affected the result of the content and activity measurement.

As mentioned before, we had speculated that atorvastatin liver content is reduced in SLCO2B1+/+ rats due to the higher hepatic activity of rCYP3A1 (Kinzi et al., 2022). To test whether the increase in rCYP3A1 protein indeed affects hepatic content of a parent OATP2B1 substrate, we determined serum levels and hepatic liver content of erlotinib, a molecule that shares the profile in absorption, distribution, metabolism, and excretion (ADME) mechanisms in terms of metabolism by CYP3A and of the involved drug transporters (OATP2B1, ABCB1, ABCG2), but we integrated the monitoring of the metabolite in the herein reported study.

Quantifying the erlotinib serum levels in SLCO2B1+/+ and rSlco2b1−/− animals, we saw higher levels in animals lacking the transporter (Fig. 2). This particular finding is in accordance with the assumption that erlotinib is a substrate of the human transporter (Bauer et al., 2018) and that OATP2B1 contributes to the hepatic uptake of the tyrosine kinase inhibitor in rodents and humans (Amor et al., 2018; Bauer et al., 2018; Marie et al., 2021). However, there was no change in the OSI-420/erlotinib ratio in the serum samples, suggesting that there is no difference in metabolism of erlotinib in our rat models. The observed ratio was about 9-fold higher than that observed at that time-point in orally dosed mSlco2b1−/− mice (Li et al., 2023), but it was comparable to that previously reported by Thappali et al. (2012) measuring plasma samples of Wistar rats orally dosed with 20 mg/kg erlotinib. Irrespective of this in vivo finding, we analyzed the tissue levels of erlotinib and OSI-420. Here, we found significantly lower levels of the parent molecule in the liver of SLCO2B1+/+ animals compared with the knockout rats. The later finding was in accordance with our previous observation for atorvastatin (Kinzi et al., 2022), but we also found lower levels of the erlotinib metabolite OSI-420. One may argue that the genetic modification of the rats also affects expression of efflux transporters that reduce the hepatocellular content. Of note, none of the efflux transporters mentioned above exhibited differences in expression comparing SLCO2B1+/+ and rSlco2b1−/−, as reported before (Kinzi et al., 2022).

There was no impact of the genotype on the OSI-420 to erlotinib ratio in serum or liver, even if we report that OSI-420 formation in microsomes isolated from livers of untreated rSlco2b1−/− rats was 25% higher in reaction rate when exposed to 25 μM erlotinib (Fig. 5). At least for human P450s, it has been shown that CYP1A1 and CYP1A2 are capable of catalyzing the demethylation of the site chain of erlotinib at least to some extent (Li et al., 2007; Luong et al., 2021). Whether this is of relevance for our finding warrants further investigation, which appears beyond the scope of our report.

In accordance with our understanding of erlotinib distribution and pharmacokinetics (Hidalgo and Bloedow, 2003; Amor et al., 2018), the livers showed the highest content of erlotinib and OSI-420 compared with all other tissues measured within our study, but it showed a low liver-to-serum ratio, at least compared with that reported in mice (Li et al., 2023). In this particular publication, Li et al. (2023) report on a comparison of pharmacokinetics of erlotinib in various genetically modified mouse strains. For none of the genetic modifications (mSlco2b1−/−, mSlco1a/1b−/−, or mSlco1/1b/2b−/−) did they observe an impact on the erlotinib exposure (AUC0–4h) or the liver-to-plasma ratio. However, when they compared the mSlco1a/1b−/− and the mSlco1a/1b/2b1−/− animals, they found a significantly reduce