MMF (Product No. 100923-202003), MPA (Product No. 100924-202004), prednisolone (Product No. 100153-201405), and prednisone (Product No. 100199-201503) were purchased from the China National Institutes for Food and Drug Control (NIFDC). MPAG (Product No. M831520) was purchased from Toronto Research Chemicals, Inc. (Toronto, Canada).

Cefoperazone (Product No. R36425), pantoprazole (Product No. 313481), meropenem (Product No. 9160629101), alprostadil (Product No. 3B096K), and ganciclovir (Product No. 160805-1) were purchased from Pfizer (New York, NY, USA), Takeda Bio (Tokyo, Japan), Haibin Pharmaceutical Company (Shenzhen, China), Tide Pharmaceutical Company (Beijing, China), and China Meheco Keyi Pharma Co., Ltd. (Wuhan, China), respectively. Salvia miltiorrhiza injection (Product No. 1603283) and olaparib (Product No. S50443) were purchased from Chiatai Qingchunbao Pharmaceutical Co., Ltd. (Hangzhou, China), and Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China), respectively.

Omeprazole (Product No. PHR1059) and β-estradiol (Product No. E8875) and its metabolite estradiol-3-glucuronide (Product No. E1127) were purchased from Sigma-Aldrich (Burlington, MA, USA). Additionally, 4-trifluoromethyl-7-hydroxycoumarin (Product No. 36852) and its metabolite, 4-trifluoromethyl-7-hydroxycoumarin glucuronide (Product No. T792035), were obtained from Sigma-Aldrich and Shanghai ZZBIO Co., Ltd. (Shanghai, China), respectively. Cryopreserved primary human hepatocytes (Lot No. HVN), as well as plating and incubation media, were purchased from BioIVT (Westbury, NY, USA).

Recombinant human embryonic kidney (HEK) 293 cells expressing human OATP1B1 (HEK293-OATP1B1, Product No. GM1102G), OATP1B3 (HEK293-OATP1B3, Product No. GM1106G), OAT1 (HEK293- OAT1, Product No. GM1103G), OAT3 (HEK293-OAT3, Product No. GM1104G), and HEK293 cells transfected with an empty vector (HEK293-MOCK, Product No. GM1101G) were supplied by GenoMembrane (Tokyo, Japan). Estradiol-17β-D-glucuronide (Product No. E1127), p-aminohippuric acid (Product No. A1422), estrone-3-sulfate (Product No. E0251), probenecid (Product No. P8761), and rifampicin (Product No. R3501) were purchased from Sigma-Aldrich. Dulbecco’s modified eagle’s medium (DMEM, Product No. 11885-084), penicillin-streptomycin (Product No. 15140-122), and fetal bovine serum (FBS, Product No. 10099-141) were obtained from Gibco (Grand Island, NY, USA). The bicinchoninic acid (BCA) protein assay kit (Product No. 23225) was provided by Thermo Fisher Scientific (Waltham, MA, USA).

Vesicle membranes expressing human MRP2 (Product No. 0001) were supplied by GenoMembrane (Tokyo, Japan). The transporter kit (GM3010RD), including Reaction Buffer A2 (containing 50 mM of MOPS-Tris, 70 mM of KCl, and 7.5 mM of MgCl₂), Stop/Wash Buffer B (400 mM of MOPS-Tris, and 700 mM of KCl), 10 mM of MgATP and MgAMP solution and 200 mM of glutathione, used in the vesicle studies was provided by Research Institute for Liver Diseases (China). Benzbromarone (Product No. B5774) was purchased from Sigma-Aldrich.

Cryopreserved human hepatocytes were recovered in a pre-warmed incubation medium, diluted to a viable cell density (∼0.7 × 10⁶ cells/mL) using a plating medium, followed by the addition of 100 μL hepatocytes into each well of a collagen I-coated 96-well plate. The cell monolayer was formed after an overnight incubation period at 37°C under 5%CO₂. The plating medium was then replaced with an incubation medium for further culture. On the second day, an incubation medium containing different concentrations of either compound, the inducer omeprazole (final concentration of 50 μM), or vehicle (0.1% [v/v] dimethyl sulfoxide [DMSO]). The cells were incubated for 96 h and additives were changed every 24 h. At the end of the incubation period, the medium was replaced with a substrate for UGT and MPA. Following a further incubation of 15 or 30 min at 37°C under 5% CO₂, 100 μL of the reaction medium was transferred to 400 μL of ice-cold methanol containing internal standard (IS) to terminate the reaction. Finally, the supernatant was collected after centrifugation at 4,000 rpm for 10 min and analyzed for glucuronide formation using liquid chromatography-mass spectrometry (LC-MS).

Briefly, cells were removed from liquid nitrogen and recovered at 37°C in a water bath and diluted to 0.8 × 10⁶ cells/mL with pre-warmed DMEM medium containing 10% (v/v) FBS. After seeding in 96-well plates pre-coated with poly-L-lysine at a density of 100 μL/well, the cells were pre-incubated at 37°C and 5% CO₂ for 3–4 h to form a monolayer. Subsequently, the medium was removed and replaced with a fresh DMEM culture medium containing 10% FBS and penicillin-streptomycin solution (100 U/mL and 0.1 mg/mL) and incubated overnight.

Stock solutions of the compounds, substrates, and inhibitors were prepared using DMSO and further diluted to a dosing solution using a transporter buffer (TB). Before the substrate assay, DMEM was aspirated, and the cells were washed twice with 100 μL of pre-warmed TB. During the second wash, the buffer remained in the wells and was pre-incubated at 37°C for 5 min. Subsequently, TB was removed and replaced with 50 μL of TB-containing compounds at different concentrations or probe substrates with or without inhibitors, and the cells were incubated at 37°C for 10 min. Notably, MOCK cells were included in the compound and substrate groups to calculate the uptake ratio (UR) and its change with or without inhibitors. Following incubation, TB was discarded, and the cells were washed three times with 100 μL of ice-cold TB for each well to completely remove the residual incubation solution. After washing, 100 μL of distilled water was added to each well, and the cells were completely lysed by repeatedly freezing and thawing with liquid nitrogen for three cycles (−196°C to 37°C). Cell lysate (30 μL) was obtained and mixed with ice-cold methanol containing IS to precipitate proteins at a volume ratio of 1:4. The mixture was centrifuged at 12,000 rpm at 4°C for 5 min, and the supernatant was collected and used to determine the amount of the compounds and substrates using LC-MS. An additional 25 μL of cell lysate was used to determine the total protein concentration using a BCA protein assay kit.

Briefly, BSA standards were prepared and diluted to obtain concentrations of 25–2000 μg/mL. BCA working reagent (WR) was prepared by mixing 50 parts of BCA reagent A with 1 part of BCA reagent B (50:1, Reagent A: B). Subsequently, 25 μL cell lysate or BSA standard was added to 200 μL WR, followed by incubation for 30 min at 37°C. Finally, the absorbance of these samples was measured at 570 nm using a plate reader.

The procedures for cell culture, dosing solution preparation, cell lysate preparation, and sampling were consistent with those in the substrate study described above, except for the pre-incubation and incubation steps. For the OATP1B1 and OATP1B3 inhibition experiments, after the cells were washed twice with TB and incubated with TB for 5 min, 50 μL of the pre-dosing solution of the compounds or inhibitors at different concentrations were added into each assigned well for a 30-min incubation. The pre-dosing solution was subsequently discarded and replaced with 50 μL of the probe substrate and compounds or known inhibitor mixture incubated for 10 min at 37°C. Finally, the amount of substrate was quantified using LC-MS.

The substrate study was performed at 37°C with a water bath using a mixture comprising of 10 μL of vesicle membranes (5 mg/mL) and 10 μL of Buffer A2 (replaced with 5 μL Reagent G and 5 μL inhibitors for the inhibitor groups). The mixture was pre-incubated at 37°C with a water bath for 5 min. Meanwhile, 10 μL of dosing solution of the compounds or probe substrates and 10 μL of Reagent C2 (ATP) or D2 (AMP, as the non-ATP dependent control) were mixed and pre-incubated at 37°C for 5 min. Following pre-incubation, the two mixtures were mixed and incubated for an additional 5 min.

After incubation, the reactions were stopped with ice-cold Buffer B, filtered through a 96-well filter plate to separate vesicle-associated ligands from the matrix, and washed five times with ice-cold Buffer B. The vesicles on the filer plate were dissolved using 50 μL of 80% methanol (v/v), and the filtrate was collected by centrifugation at 2000 rpm for 2 min. The dissolution and centrifugation procedures were repeated. The filtrates were combined up to a total of 100 μL and mixed with 200 μL of ice-cold methanol containing IS, followed by centrifugation at 12,000 rpm for 5 min. The supernatant was collected to quantify the amount of compound or substrate using LC-MS.

The incubation, termination, washing, and sampling steps were similar to those in the substrate study described above, except for the mixture content. Briefly, 10 μL of vesicle membranes and 5 μL of Reagent G were mixed and pre-incubated with 10 μL of the compounds or Buffer A2 (used as no compound-treatment control [NC]) at 37°C for 5 min. Meanwhile, 5 μL of MPAG and 20 μL of Reagent C2 (ATP) or D2 (AMP, used for non-ATP dependent control) were mixed and pre-incubated at 37°C for 5 min. The two mixtures were mixed before the incubation and then incubated for an additional 5 min. Finally, the supernatant was collected to analyze the amount of MPAG using LC-MS.

The LC-MS system used in this study contained a LEAP CTC auto sampler (HTS PAL), liquid chromatograph (LC 20AD), and triple-quadrupole mass spectrometer (API 4000), and was programmed using the Analyst^®1.6.3 software. The detailed conditions for the LC-MS analysis of glucuronide formation, transporter substrates, and MPAG are shown in Supplementary Texts S1–S3 and Supplementary Table S1.

The glucuronidation of MPA is primarily known to occur in the liver and, to a certain extent, in the gastrointestinal tract and kidney. UGT1A8, UGT1A9, UGT1A10, and UGT2B7 are involved in the metabolism of MPA to MPAG. UGT1A9 and UGT2B7 are considered the major metabolizing enzymes of MPA owing to their relatively high hepatic and renal expression.

After treating human hepatocytes with prednisolone, the active metabolite of prednisone, and CsA at concentrations ranging from 0.5 to 100 μM for 96 h, the incubation medium was replaced with 100 μM 4-trifluoromethyl-7-hydroxycoumarin and 100 μM β-estradiol, which are probe substrates of UGT1A9/2B7 and UGT1A1/1A3, respectively, for an additional incubation of 30 min. The formation of 4-trifluoromethyl-7-hydroxycoumarin glucuronide and estradiol-3-glucuronide was detected to evaluate the induction of UGT enzyme activity compared with that in the NC group. In this study, omeprazole, which is a known UGT inducer, was used as the positive control (Schaefer et al., 2012).

After treatment with prednisolone at concentrations of 0.5, 1, 5, 10, 50, and 100 μM, the glucuronidation activity of 4-trifluoromethyl-7-hydroxycoumarin in hepatocytes was 139%, 157%, 154%, 154%, 173%, and 155% of that in the NC group, respectively (p-value <0.05), and the β-estradiol glucuronidation activity in hepatocytes was 119%, 122%, 124%, 123%, 140%, and 134% of that in the NC group, respectively (Figure 1 and Supplementary Table S2, p-value <0.05). After treatment with omeprazole, the glucuronidation activities of 4-trifluoromethyl-7-hydroxycoumarin and β-estradiol were 145% and 243%, respectively (p-value <0.05). Furthermore, the conversion of MPA to MPAG in human hepatocytes was evaluated after incubation with prednisolone for 96 h. A slight increase in MPAG formation was observed in the prednisolone-treated groups.

These results suggest that prednisolone can significantly induce UGT activity, particularly the UGT1A9 and UGT2B7 isoforms, resulting in increased metabolism of MPA to MPAG. However, this induction potential was not observed in CsA-treated human hepatocytes (data not shown).

First, we examined the UR of MPAG at concentrations of 0.5, 1, and 10 μM after incubation with HEK293 cells expressing human OATP1B1 or OATP1B3 to determine whether MPAG uptake is mediated by OATP1B1 and OATP1B3 transporters. The inhibition rate of MPAG was determined after treatment with rifampicin, which is a selective inhibitor of OATP1B1 and OATP1B3.

At a concentration of 10 μM, the UR of MPAG mediated by OATP1B1 and OATP1B3 was 5.54 and 15.1, respectively, which decreased by 100% and 99.7% after being treated with 100 μM rifampicin (Figure 2; Supplementary Table S3). In contrast, the URs of estradiol-17β-glucuronide (E₂17βG), used as a probe substrate of OATP1B1 and OATP1B3, were 17.9 and 12.1, respectively, and were inhibited by 97.7% and 92.7%, respectively, after rifampicin treatment. Therefore, these results suggest the potential of MPAG as a substrate for OATP1B1 and OATP1B3, consistent with observations reported by previous studies (Matsunaga et al., 2014).

We then investigated the concentration-dependent inhibitory effects of prednisolone and CsA on the uptake activity of 10 μM MPAG mediated by OATP transporters. CsA, rather than prednisolone, was shown to have potential inhibitory effects on the OATP-mediated transport of MPAG. The IC₅₀ of CsA was estimated to be 2.25 and 0.906 μM for OATP1B1 and OATP1B3, respectively. However, the IC₅₀ of rifampicin on MPAG uptake was 0.528 and 0.224 μM for OATP1B1 and OATP1B3, respectively. CsA inhibited OATP-mediated MPAG uptake, and its inhibitory potency was comparable to that of rifampicin.

Finally, the R value was calculated to assess the clinical risk of CsA inhibiting OATP-mediated MPAG uptake. At 300 mg dosage, the maximum plasma concentration of CsA at the liver inlet was 1924 ng/mL. The R value of CsA inhibiting OATP1B1 and OATP1B3-mediated transport of MPAG was 1.05 and 1.13, respectively, indicating that CsA has the potential to inhibit OATP1B3-mediated transport of MPAG. The detailed parameters are shown in Supplementary Table S4.

Several studies have demonstrated that MRP2 is involved in the renal and biliary excretion of MPAG in vivo. We evaluated the inhibitory potential of prednisolone and CsA on human MRP2 transporters using membrane vesicles. The UR of 2 μM MPAG was 18.8, and a known inhibitor, benzbromarone, demonstrated apparent inhibitory effects on MPAG uptake mediated by MRP2 with an IC₅₀ of 17.6 μM (Figure 3). However, such inhibition was not detected after treatment with prednisolone and CsA.

MPAG has been shown to be transported by OAT3; however, the study only evaluated the transport rate in human OAT3-expressing cells compared with empty vector-expressing cells, as well as the changes following incubation with inhibitors. Therefore, we further determined the UR of MPAG mediated by human OAT1 and OAT3 as well as the change in the UR following treatment with probenecid, a known selective inhibitor of OAT1 and OAT3.

The UR of p-aminohippuric acid (PAH), used as a probe substrate for OAT1, in OAT1-expressing cells compared to that in MOCK cells was 159, which decreased to 14.6 (91.4% inhibition) following probenecid treatment (Figure 4). At concentrations of 0.5, 1, and 10 μM, the URs of MPAG mediated by OAT1 were all less than 2.0, indicating that MPAG is not a substrate of OAT1.

In the OAT3 substrate assay, the UR of the probe substrate estrone-3-sulfate (ES) was 7.37, which decreased to 3.26 (69.4% reduction) after probenecid treatment. At a concentration of 10 μM, the UR of MPAG mediated by OAT3 was 2.47, which is close to the reference value and thereby considered a substrate. Following incubation with probenecid, the UR of MPAG decreased to 1.19 (90.8% reduction). Therefore, this result suggests that MPAG is a weak substrate for OAT3.

The results of our study indicated that MRP2, OAT3, OATP1B1, and OATP1B3 are responsible for MPAG efflux or influx. Based on clinical practice in China, the inhibitory effects of seven drugs that are usually co-administered with MMF (cefoperazone, pantoprazole, meropenem, with probenecid, alprostadil, ganciclovir, S. miltiorrhiza, and olaparib) were evaluated for their impact on the activity of the abovementioned transporters.

No apparent inhibitory effect on MRP2 was observed following treatment with 50 μM or 50 μg/mL of the seven drugs (Figure 5 and Supplementary Table S5), suggesting that these drugs did not inhibit the activity of MRP2. For OATP1B1, following incubation with 50 μM S. miltiorrhiza and alprostadil, its activity was found to be reduced to 8.67% and 33.6% relative to NC, respectively. Regarding OATP1B3, following treatment with 50 μM S. miltiorrhiza, its activity decreased to 8.33% of that in NC. Following treatment with 50 μM of cefoperazone, S. miltiorrhiza, and meropenem, the transport activity of OAT3 decreased to 13.3%, 26.5%, and 26.6% of that observed in NC, respectively.

Therefore, these results indicate a potential DDI risk between S. miltiorrhiza and MMF. Salvia miltiorrhiza showed apparent inhibitory effects on OATP and OAT3 transporters, which are involved in hepatic uptake and renal excretion of MPAG. Furthermore, S. miltiorrhiza reportedly has an inhibitory effect on OAT1 and OAT3 and is considered a substrate of OATP1B1 and OATP1B3 (Wang and Sweet, 2012; Lu et al., 2022). However, no reports exist on its inhibitory effect on OATP transporters.