Materials.

Irinotecan, SN-38, and SN-38-G were purchased from PerkinElmer (Waltham, MA). Bovine serum albumin (BSA), dithiothreitol, iodoacetamide, trypsin protease (mass spectroscopy grade), PBS, Hanks’ balanced salt solution, hepatocyte maintenance supplement pack (serum-free), membrane protein extraction kit (Mem-PER Plus kit), and bicinchoninic acid assay (BCA) kit were purchased from Thermo Fisher Scientific (Rockford, IL). Recombinant UGT and CES enzymes were procured from Corning (Riverfront, NY). Human liver S9 (pool of n = 10) and intestinal S9 (pool of n = 15) fractions were purchased from Sekisui XenoTech LLC (Kansas City, KS). Membrane vesicle (MV) or cell lines overexpressing P-gp, breast cancer resistant protein (BCRP), multidrug resistance–associated proteins (MRP2, MRP3, and MRP4), and organic anion transporting polypeptide (OATP1B1 and OATP2B1) transporters were provided by Solvo Biotechnology (Budapest, Hungary). The multiscreen HTS Vacuum Manifold and 96-well filter plates with class B glass fiber filters were obtained from EMD Millipore (Billerica, MA). Dulbecco’s modified Eagle’s medium, penicillin-streptomycin, alamethicin, uridine 5′-diphosphoglucuronic acid trisodium salt (UDPGA), ATP disodium salt, AMP monohydrate, magnesium chloride, glutathione, Tris [hydroxymethyl] aminomethane (Tris-Base), NaCl, sucrose, 3-[N-morpholino] propanesulfonic acid (MOPS), epitestosterone, and epitestosterone glucuronide were purchased from Sigma-Aldrich (St. Louis, MO). Poly(d-lysin)–coated 24-well tissue culture plates were from BD biosciences, (Franklin Lakes, NJ). Stable isotope–labeled (heavy) peptides and synthetic unlabeled (light) peptides for targeted proteomics assay were purchased from Thermo Fisher Scientific (Rockford, IL) and New England Peptides (Boston, MA), respectively. All other reagents and chemicals were purchased from commercial suppliers offering the highest purity.

Outline of Experimental Workflow.

A systematic workflow used for the quantitative characterization of irinotecan disposition involved the following five steps (Supplemental Fig. 1).

1. Determination of the metabolic intrinsic clearance of irinotecan and SN-38 by recombinant CESs and UGTs, respectively.

2. Confirmation of the rate-determining steps in irinotecan metabolism in S9 fractions and hepatocytes.

3. Determination of the uptake and efflux transport intrinsic clearance of irinotecan, SN-38, and SN-38-G using in vitro cellular uptake or vesicular assay.

4. Estimation of the relative contributions of individual enzymes and transporters [fractional metabolism (f_m) and fractional transport (f_t)] by quantifying and comparing the protein abundance data in in vitro systems versus liver, intestine, and kidney tissues.

5. Investigation of SN-38 reactivation by gut bacterial β-glucuronidases in human fecal homogenate.

Irinotecan and SN-38 Metabolism in Recombinant Enzymes, S9 Fractions, and Hepatocytes.

Metabolic intrinsic clearances of irinotecan and SN-38 were determined by incubation in individual recombinant enzymes using an established method (Hanioka et al., 2001). Sequential metabolism of irinotecan in human liver and intestinal S9 fractions was investigated as done previously (Zhang et al., 2018). The relative contributions of CESs and UGTs on irinotecan and SN-38 metabolism were characterized in cryopreserved individual human hepatocytes (n = 3 donors) (Lau et al., 2002). Briefly, the potential of irinotecan metabolism by CES1 and CES2 was estimated by incubating irinotecan (5 and 10 µM) with 10 µg of recombinant CES1 and CES2 in 50 mM Tris buffer (pH 7.0) in 100 µl incubation volume. The reaction was performed for 15 minutes at 37°C with gentle shaking and quenched by the addition of 200 μl of 1:1 ice-cold acetonitrile:formic acid (0.2%) containing internal standard (10 ng/ml epitestosterone). The reaction mix was centrifuged at 10,000g for 5 minutes at 4°C to remove the cell debris, and the supernatant was collected to quantify the metabolite formation by LC-MS/MS. Since the screening assay indicated the involvement of CES1 and CES2 in irinotecan hydrolysis, a detailed irinotecan metabolism kinetics assay was performed following the same protocol across a wide concentration range (1–400 µM). SN-38 formation was monitored in the incubation buffer to estimate the nonenzymatic hydrolysis of irinotecan.

SN-38 (5 and 10 µM) was incubated with recombinant UGT1A1, UGT1A4, UGT1A6, UGT1A9, UGT2B7, UGT2B15, and UGT2B17 enzymes (20 µg per reaction) to identify UGTs involved in SN-38 glucuronidation. The assay was performed at pH 7.4 in 100 µl of buffer consisting of 5 mM MgCl_2, 100 mM KPO₄, alamethicin (0.2 mg/ml, final concentration), and BSA (0.2%). After addition of the substrate and the enzyme, the buffer system was preincubated on ice for 15 minutes to allow microsomal pore formation by alamethicin. The glucuronidation was initiated by adding 2.5 mM UDPGA. After a 30-minute incubation at 37°C with gentle shaking, the reaction was stopped by addition of 200 μl ice-cold acetonitrile:0.2% formic acid (1:1; v/v) containing internal standard (10 ng/ml epitestosterone glucuronide). In parallel, the reaction was also performed in the absence of UDPGA (negative control). The reaction mix was centrifuged at 10,000g for 5 minutes at 4°C to remove precipitated proteins, and the supernatant was collected to quantify the metabolite formed. The metabolism kinetic experiment was conducted for the shortlisted UGTs. Briefly, SN-38 was incubated across 1–120 µM concentration range with and without UDPGA using the protocol described above. All reactions were carried out in triplicate. The concentration-dependent SN-38 and SN-38-G formation was measured by a validated LC-MS/MS method (Supplemental Table 1). The V_max and the substrate concentration at half-maximum velocity (K_m) were estimated, and the V_max was normalized by tissue enzyme abundance (pmol/mg of total protein) (Supplemental Table 2).

Irinotecan (1–400 µM) and SN-38 (1–120 µM) were incubated with the human liver (a pool of n = 10) and intestinal (a pool of n = 15) S9 fractions (20 µg per reaction). UDPGA (2.5 mM) was added to the SN-38 metabolism reaction and incubated for 30 minutes in a buffer containing 5 mM MgCl₂, 100 mM KH₂PO₄, alamethicin (0.2 mg/ml), and BSA (0.2%) at pH 7.4. The metabolism was quenched by addition of 200 μl ice-cold 1:1 acetonitrile:0.2% formic acid containing corresponding epitestosterone and epitestosterone glucuronide (i.e., the internal standards for SN-38 and SN-38-G, respectively). The reaction mix was centrifuged at 10,000g for 5 minutes at 4°C to remove the precipitated proteins. The supernatant was subjected to metabolite quantification by LC-MS/MS (Supplemental Table 2).

The liver and intestinal S9 fraction incubations were performed to quantify the sequential formation of SN-38 and SN-38-G by CESs and by UGTs, respectively. Briefly, irinotecan (1–400 µM) was incubated for 30 minutes with the liver and intestinal S9 fractions (20 µg/reaction) with and without 2.5 mM UDPGA in triplicate using the protocol described above. The reaction was quenched, and irinotecan, SN-38, and SN-38-G were quantified by LC-MS/MS (Supplemental Table 1). The rate of metabolite formation was estimated and normalized by the tissue protein abundance (pmol/mg of S9 protein).

For irinotecan and SN-38 metabolism in human hepatocytes, the cryopreserved cells were thawed at 37°C in a water bath and transferred to a 50-ml tube containing 15 ml of suspension medium at 4°C. The cells were centrifuged at 600 rpm at 4°C for 5 minutes and washed twice. The pellet was gently resuspended in the medium to a final density of 2 million cells per milliliter. The viability of individual hepatocyte lots was determined by the trypan blue staining method immediately after thawing and centrifugation. Both irinotecan and SN-38 were tested at 1, 10, and 100 µM final concentrations. The incubations were carried out with 0.5 × 10⁶ hepatocytes per milliliter in 96-well plates at 37°C for 30 minutes. The reaction was stopped by the addition of 200 μl ice-cold 1:1 acetonitrile:0.2% formic acid containing epitestosterone and epitestosterone glucuronide (internal standards for SN-38 and SN-38-G, respectively), and the samples were centrifuged to collect the supernatants for LC-MS/MS analysis of irinotecan, SN-38, and SN-38-G (Supplemental Table 1). UGT and CES enzymes were quantified in the same hepatocytes using the method described below.

Cellular and Vesicular Uptake Transport Assay of Irinotecan, SN-38, and SN-38-G.

First, active uptake of irinotecan, SN-38, and SN-38-G by hepatic and intestinal OATP transporters was investigated using a cellular transport assay as described previously (Izumi et al., 2015). We used HEK-293 and Madin-Darby canine kidney-II cells stably expressing OATP1B1 and OATP2B1, respectively. Briefly, transporter or mock-transfected HEK-293 and Madin-Darby canine kidney-II cells were grown in tissue culture flasks at 37°C in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 1% nonessential amino acids, 2 mM l-glutamine, and 100 U/ml penicillin-streptomycin supplemented with 5% CO₂ and 3 µg/ml puromycin. Approximately 4 × 10⁵ cells were seeded per well in a poly(d-lysine)–coated 24-well plate. After 24 hours, 5 mM sodium butyrate was added to the cells for another 24 hours to induce OATP expression. Cells were washed twice with 1× PBS and preincubated with 300 µl Hanks’ balanced salt solution for 10 minutes at 37°C before incubation with medium containing irinotecan, SN-38, and SN-38-G for 5 minutes. Cells were washed three times with ice-cold 1× PBS, and the cell pellet was lysed by adding acetonitrile containing the internal standards (epitestosterone for irinotecan and SN-38, epitestosterone glucuronide for SN-38-G). Cell lysate was centrifuged at 10,000g for 10 minutes, and the supernatant was subjected to LC-MS/MS analysis. SN-38 and SN-38-G transport kinetics experiments were conducted for the shortlisted transporters (final concentrations, 1–200 µM). Further, total protein quantification was performed by BCA assay and transporter quantification by LC-MS/MS quantitative proteomics. K_m and V_max were calculated, and the V_max was normalized by the transporter protein abundance in the cell system. Second, we characterized the role of hepatic and intestinal efflux transporters (P-gp, BCRP, and MRPs) in the transport of irinotecan and its metabolites by vesicular uptake assay using ABC efflux transporter–expressing MVs (Li et al., 2019). The MVs were diluted in the transport buffer [40 mM MOPS-Tris (pH 7.0), 70 mM KCl, and 7.5 mM MgCl₂ for MRP2 vesicles and 10 mM Tris-HCl, 10 mM MgCl2, and 250 mM sucrose for BCRP, MDR1, MRP3, and MRP4 vesicles] and added (50 µl/well) to a 96-well plate on ice. The transport of irinotecan, SN-38, and SN-38-G (1 µM) was initiated by adding 25 µl AMP or ATP (4 mM) at 37°C for 20 (irinotecan and SN-38) or 30 seconds (SN-38-G). The reaction was stopped by the addition of 200 μl ice-cold washing buffer (40 mM MOPS-Tris, pH 7.0, 70 mM KCl), which was washed for five times with 200 μl of ice-cold wash buffer. The MVs were eluted with 100 μl of 1:1 acetonitrile:0.2% formic acid containing corresponding deuterated internal standards and kept 1 hour at room temperature for vesicle lysis. The lysed vesicles containing substrates were collected through the vacuum filtration into a 96-well collector plate. The plate was centrifuged at 3000g for 2 minutes to collect the supernatant and subjected to substrate quantification by LC-MS/MS. Transport kinetics data were obtained for the shortlisted transporters with a concentration range of 1–200 μM. Transport kinetic parameters (V_max and K_m) were estimated and normalized by the protein abundance (pmol/mg of total protein) in the in vitro system.

Quantification of CESs, UGTs, and Transporter Proteins.

The protein abundance of enzymes and transporters was quantified in the recombinant systems, S9 fractions, transporter-expressed cell system, cryopreserved hepatocytes, and tissues (liver, intestine, kidney, lung) using optimized LC-MS/MS methods (Supplemental Table 1) (Xu et al., 2017). The crude membranes were isolated from the cell pellets using a membrane protein extraction kit (Thermo Fisher Scientific, Rockford, IL). Approximately 10 million cells were suspended with the permeabilization buffer (250 μl) and gently mixed and kept on a compact digital waving rotator (Thermo Fisher Scientific, Rockford, IL) for 30 minutes (4°C) at 300 rpm. The permeabilized cell suspension was centrifuged at 16,000g for 15 minutes (4°C). The pellet was resuspended with solubilization buffer (250 μl), gently mixed, incubated for 60 minutes at 300 rpm (4°C), and centrifuged at 16,000g for 15 minutes (4°C). The supernatant containing membrane proteins was collected, and the total protein concentration was measured by BCA assay. In total, 80 μl of 2 mg/ml protein sample was mixed with 30 μl ammonium bicarbonate buffer (100 mM, pH 7.8), 10 μl dithiothreitol (250 mM), and 20 μl BSA (0.02 mg/ml), followed by heat denaturation and 10 minutes of reduction at 95°C. In all, 10 μl of iodoacetamide (100 mM) was added after cooling at room temperature and kept in the dark for 30 minutes. Ice-cold acetone (500 μl) was added to the sample, which was kept at −20°C for 30 minutes before centrifugation at 16,000g for 5 minutes at 4°C. Protein pellet was washed with 500 µl methanol, the sample was centrifuged at 8000g for 5 minutes at 4°C, and the pellet was collected after air drying. The pellet was resuspended with 60 μl ammonium bicarbonate (50 mM, pH 7.8) and 20 μl trypsin (0.16 μg/μl) for 16 hours at 300 rpm (37°C). Trypsin digestion was quenched on ice with addition of 20 μl of stable isotope–labeled peptide cocktail (internal standard) and 10 μl acetonitrile: water 80:20 (v/v) with 0.5% formic acid. The sample was prepared by centrifuging at 8000g for 5 minutes (4°C), and the supernatants (50 μl) were subjected to LC-MS/MS quantifications. Previously characterized pooled human liver microsome and liver membrane samples were digested and used as calibrators for the quantification of enzymes and transporters, respectively (Zhang et al., 2018).

SN-38-G to SN-38 Reactivation by Gut Bacterial β-d-Glucuronidase in Human Fecal Samples.

Human fecal samples were collected and stored at −80°C until further use. Characterization of the bacterial flora and ethical concerns were previously reported by us in Jariwala et al. (2020), in which the total fecal protein was extracted to quantify SN-38 reactivation rate by the bacterial β-d-glucuronidase. In brief, approximately 5 g of thawed fecal material in a solution containing 25 ml of cold extraction buffer (pH 6.5, 25 mM HEPES, 25 mM NaCl with Roche cOmplete protease inhibitor cocktail) and 500 mg of autoclaved garnet beads was vortexed vigorously to break up dense, fibrous material. The suspended sample was centrifuged at a low speed (300g, 5 minutes, 4°C) to separate out any insoluble fecal material. After decanting the microbial supernatant from the fecal homogenate, an additional 25 ml of cold extraction buffer was added to the remaining fibrous material, and the extraction process was repeated. The combined microbial supernatant (∼40–45 ml) was centrifuged at a low speed to remove any remaining insoluble debris. This process was repeated with the decanted microbial supernatant. The microbial supernatant was ultrasonicated for 1.5 minutes while on ice. The lysate was mixed by inversion, and the sonication was repeated. The lysed cells were centrifuged at high speed (17,000g, 20 minutes, 4°C) to remove cellular debris. The decanted lysate was concentrated, and metabolites were removed by buffer exchanging with fresh extraction buffer. The concentration of total protein in the fecal extract was calculated using a standard Bradford assay protocol. The human fecal extract was aliquoted and snap-frozen using liquid nitrogen. The aliquots were stored at –80°C until further use. Relative quantifications of bacterial β-d-glucuronidase enzymes from the fecal extracts were obtained following a previously described protocol (Jariwala et al., 2020). Relative quantification values were reported as label-free quantification (LFQ) intensities, which were normalized, and combined peptide signal intensities as determined by the MaxLFQ algorithm in MaxQuant (Jariwala et al., 2020).

Data Analysis of Drug Metabolism, Transport, and Microbial Reactivation.

Irinotecan and SN-38 metabolite formation kinetic parameters were estimated by fitting the Michaelis-Menten equation (eq. 1) in GraphPad Prism (version 5.0) (La Jolla, CA).

where Y is the metabolite formation rate of SN-38 and SN-38-G (pmol/min/pmol of protein), S is the substrate concentration in the reaction (μM), K_m is the Michaelis-Menten constant (μM), and V_max is the maximum rate. The metabolism rate was normalized by the enzyme abundance in the recombinant versus human tissues (liver, intestine, and kidney). To address the nonenzymatic degradation, SN-38 formation was subtracted from the value in the incubation buffer. The vesicular and cellular uptake kinetic parameters were estimated using Michaelis-Menten equation (eq. 1) after subtracting the passive uptake in the mock-transfected system. Transport data were normalized by the protein abundance value (pmol per mg protein). The percentage of inside-out data of the MVs from our previous study (Li et al., 2019) was used to normalize the transport kinetic parameters.

Two-sided Student’s t test was used to determine the statistical significance of the differences between the control (e.g., mock-transfected cells, vesicles, or without UDPGA incubations) verses metabolism/transport data in recombinant systems and with cofactors. The results are expressed as means ± S.D. The Pearson correlation between SN-38-G deglucuronidation and β-glucuronidase (LFQ) was obtained using GraphPad Prism (version 5.0). P < 0.05 was considered statistically significant.

In Vitro to In Vivo Scaling of Fractional Contribution of Individual Enzyme and Transporter.

A summary of stepwise scaling approach to determine fractional contributions of individual enzymes and transporters in irinotecan disposition is provided below and elaborated in the Supplemental Material.

1. The CES and UGT-mediated intrinsic clearance (CL) data were normalized by the abundance of individual enzymes in the recombinant system (eq. 2):

where E_recombinant is the enzyme abundance in the recombinant system and CL_recombinant, is recombinant clearance (Supplemental Table 2).

2. The tissue intrinsic clearance (CL,_tissue) of individual enzymes in whole organ was estimated using eq. 3.

3. The tissue intrinsic clearance for CES or UGT was used to calculate the total intrinsic clearance using eqs. 4 and 5.

4. The scaled total clearance (CL_total) allowed the calculation of fractional metabolism (ƒ_m) of irinotecan and SN-38 by individual CES and UGT in vivo using eq. 6:

where CL_int,DME is the tissue intrinsic clearance of individual enzymes.

Similarly, in vitro transport kinetics were used to estimate the fractional contribution (ƒ_t) of each transporter (Li et al., 2019), where the intrinsic clearance for vesicular uptake was calculated using in vitro transport kinetics data normalized by percentage of inside-out of the transporters in the vesicular system (eq. 7).

The vesicular and cellular intrinsic clearance was scaled to tissue intrinsic CL using transporter expression data, where ƒ_t was estimated similar to approaches described for metabolism using eqs. 8–13:

where CL_{int,transporter} is the tissue intrinsic clearance of individual transporters.