Jose Silva

Metabolite Identification in Drug Discovery

Recent developments in the technologies and approaches to identify metabolites in a drug discovery environment are reviewed. Samples may be generated using either in vitro systems--typically, but not exclusively, liver subcellular fractions, such as microsomes, or whole cells, such as hepatocytes. Alternatively, metabolites are generated in vivo using excreta obtained following dosing in preclinical species. Recombinant drug metabolizing enzymes or microorganisms may offer alternate vectors. New techniques, such as the use of solid-phase microextraction, have found application in the isolation of metabolites from biological matrices. However, this is still dominated by the use of preparative chromatography, which has advanced through the use of mass-directed detection. Detection and structural elucidation by mass spectrometry have improved markedly with increases in sensitivity, allowing lower abundance metabolites to be detected, and increases in selectivity, with the use of high-resolution time-of-flight and quadrupole-time-of-flight instruments. Finally, higher field strength magnets coupled with novel probe designs and increased use of liquid chromatographic hyphenation techniques continue to drive the capabilities of nuclear magnetic resonance spectroscopy as the definitive structural elucidation tool.

Tethered spheroids as an in vitro hepatocyte model for drug safety screening

Hepatocyte spheroids mimic many in vivo liver-tissue phenotypes but increase in size during extended culture which limits their application in drug testing applications. We have developed an improved hepatocyte 3D spheroid model, namely tethered spheroids, on RGD and galactose-conjugated membranes using an optimized hybrid ratio of the two bioactive ligands. Cells in the spheroid configuration maintained 3D morphology and uncompromised differentiated hepatocyte functions (urea and albumin production), while the spheroid bottom was firmly tethered to the substratum maintaining the spheroid size in multi-well plates. The oblate shape of the tethered spheroids, with an average height of 32 μm, ensured efficient nutrient, oxygen and drug access to all the cells within the spheroid structure. Cytochrome P450 induction by prototypical inducers was demonstrated in the tethered spheroids and was comparable or better than that observed with hepatocyte sandwich cultures. These data suggested that tethered 3D hepatocyte spheroids may be an excellent alternative to 2D hepatocyte culture models for drug safety applications.

Galactosylated cellulosic sponge for multi-well drug safety testing

Hepatocyte spheroids can maintain mature differentiated functions, but collide to form bulkier structures when in extended culture. When the spheroid diameter exceeds 200 μm, cells in the inner core experience hypoxia and limited access to nutrients and drugs. Here we report the development of a thin galactosylated cellulosic sponge to culture hepatocytes in multi-well plates as 3D spheroids, and constrain them within a macroporous scaffold network to maintain spheroid size and prevent detachment. The hydrogel-based soft sponge conjugated with galactose provided suitable mechanical and chemical cues to support rapid formation of hepatocyte spheroids with a mature hepatocyte phenotype. The spheroids tethered in the sponge showed excellent maintenance of 3D cell morphology, cell-cell interaction, polarity, metabolic and transporter function and/or expression. For example, cytochrome P450 (CYP1A2, CYP2B2 and CYP3A2) activities were significantly elevated in spheroids exposed to β-naphthoflavone, phenobarbital, or pregnenolone-16α-carbonitrile, respectively. The sponge also exhibits minimal drug absorption compared to other commercially available scaffolds. As the cell seeding and culture protocols are similar to various high-throughput 2D cell-based assays, this platform is readily scalable and provides an alternative to current hepatocyte platforms used in drug safety testing applications.

Metabolism and Excretion of Canagliflozin in Mice, Rats, Dogs, and Humans

Canagliflozin is an oral antihyperglycemic agent used for the treatment of type 2 diabetes mellitus. It blocks the reabsorption of glucose in the proximal renal tubule by inhibiting the sodium-glucose cotransporter 2. This article describes the in vivo biotransformation and disposition of canagliflozin after a single oral dose of [14C]canagliflozin to intact and bile duct-cannulated (BDC) mice and rats and to intact dogs and humans. Fecal excretion was the primary route of elimination of drug-derived radioactivity in both animals and humans. In BDC mice and rats, most radioactivity was excreted in bile. The extent of radioactivity excreted in urine as a percentage of the administered [14C]canagliflozin dose was 1.2%–7.6% in animals and approximately 33% in humans. The primary pathways contributing to the metabolic clearance of canagliflozin were oxidation in animals and direct glucuronidation of canagliflozin in humans. Unchanged canagliflozin was the major component in systemic circulation in all species. In human plasma, two pharmacologically inactive O-glucuronide conjugates of canagliflozin, M5 and M7, represented 19% and 14% of total drug-related exposure and were considered major human metabolites. Plasma concentrations of M5 and M7 in mice and rats from repeated dose safety studies were lower than those in humans given canagliflozin at the maximum recommended dose of 300 mg. However, biliary metabolite profiling in rodents indicated that mouse and rat livers had significant exposure to M5 and M7. Pharmacologic inactivity and high water solubility of M5 and M7 support glucuronidation of canagliflozin as a safe detoxification pathway.

LC-ESI-MS/MS quantification of 4β-hydroxycholesterol and cholesterol in plasma samples of limited volume.

A liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) assay was developed and qualified for analyzing 4β-hydroxycholesterol and cholesterol in 5 μl of human and mouse plasma. Stable isotope-labeled d7-analogs of both analytes were used as internal standards and 4.2% (w/v) human serum albumin in phosphate-buffered saline was used as the surrogate matrix for preparation of calibration curves and QCs. The assay is capable of quantification of 4β-hydroxycholesterol and cholesterol from 5 to 500 ng/ml and 50 to 2000 μg/ml, respectively, with acceptable accuracy and precision following evaluation of recovery of analytes, autosampler stability and potential contribution of chemical oxidation to the formation of 4β-hydroxycholesterol. The final reconstituted solution was diluted for quantification of cholesterol typically present at 1000 fold higher concentration than 4β-hydroxycholesterol in the same samples used for 4β-hydroxycholesterol quantification. The successful quantification using a low plasma volume was achieved by quantification of total forms (free and conjugated) of both analytes after alkaline hydrolysis, followed by derivatization to form electrospray ionization-sensitive picolinyl esters, which upon collision-induced dissociation gave high mass precursor-product ion pair for selective detection by multiple reaction monitoring. In addition, chromatographic separation using a 16-min reversed phase gradient elution on a 1.9 μm particle size, C18 column, overcame interference from other isobaric plasma oxysterols during detection by multiple-reaction monitoring. This assay was compared to an orthogonal enzymatic assay for cholesterol and all samples, but one, provided values that were within 10% of each other. In addition, this assay passed the incurred sample tests for both analytes in human and mouse plasma samples according to reported acceptance criteria for incurred sample reanalysis. The quantification of both analytes permitted the determination of 4β-hydroxycholesterol compared to its ratio to cholesterol as an endogenous biomarker for CYP3A4/5 activity. The LC-ESI-MS/MS assay was also successfully applied to quantification of 4β-hydroxycholesterol and cholesterol in plasma samples from untreated human and mice including FRG™ KO C57Bl/6 chimeric mice with humanized livers. The preliminary data indicated that the plasma 4β-hydroxycholesterol concentrations or their ratio to cholesterol from mice including chimeric mice were higher than those from human.

Interleukins-12 and -23 Do Not Alter Expression or Activity of Multiple Cytochrome P450 Enzymes in Cryopreserved Human Hepatocytes

Psoriasis is a T-cell-mediated autoimmune disease involving the skin. Two cytokines, interleukin-12 (IL-12) and IL-23 have been shown to play a pivotal role in the pathogenesis of the disease. Ustekinumab (Stelara) is a therapeutic monoclonal antibody (mAb) targeted against the p40 shared subunit of IL-12 and IL-23. Recently the ability of therapeutic proteins (TP) including mAbs that target either cytokines directly (e.g., Pegasys; peginterferon α-2a) or their respective cell surface receptors [e.g., tocilizumab (Actemra); anti IL-6R] to desuppress cytochrome P450 (P450) enzymes in vitro and in the clinic, has been demonstrated. In the present study the ability of IL-12 and IL-23 to suppress multiple P450 enzymes was investigated in vitro using six separate lots of cultured human hepatocytes. Following exposure of 10 ng/ml IL-12 and IL-23 for 48 hours, either alone or in combination, no change in CYP2B6, 2C9, 2C19, or 3A4 gene expression or functional activity was observed. None of the untreated hepatocyte donors showed appreciable expression of the IL-12 or IL-23 receptors. Similar results were seen with whole human liver samples. Exposure of hepatocytes to IL-12 and/or IL-23, known P450 suppressors (IL-6 and tumor necrosis factor-α) or known P450 inducers (β-naphthoflavone, phenobarbital, and rifampicin) did not appreciably alter the expression of the IL-12 and IL-23 receptors either. Finally, in contrast to the positive control IL-6, expression of the acute phase C-reactive protein was unaltered following IL-12 and/or IL-23 treatment. Together, these data suggest a negligible propensity for IL-12 or IL-23 to directly alter P450 enzymes in human hepatocytes.

Cytochrome P450 induction response in tethered spheroids as a three-dimensional human hepatocyte in vitro model.

Cytochrome P450 (CYP) induction is a key risk factor of clinical drug–drug interactions that has to be mitigated in the early phases of drug discovery. Three‐dimensional (3D) cultures of hepatocytes in vitro have recently emerged as a potentially better platform to recapitulate the in vivo liver structure and to maintain long‐term hepatic functions as compared with conventional two‐dimensional (2D) monolayer cultures. However, the majority of published studies on 3D hepatocyte models use rat hepatocytes and the response to CYP inducers between rodents and humans is distinct. In the present study, we constructed tethered spheroids on RGD/galactose‐conjugated membranes as an in vitro 3D model using cryopreserved human hepatocytes. CYP3A4 mRNA expression in the tethered spheroids was induced to a significantly greater extent than those in the collagen sandwich cultures, indicating the transcriptional regulation was more sensitive to the CYP inducers in the 3D model. Induction of CYP1A2, CYP2B6 and CYP3A4 activities in the tethered spheroids were comparable to, if not higher than that observed in the collagen sandwich cultures. The membrane‐based model is readily integrated into multi‐well plates for higher‐throughput drug testing applications, which might be an alternative model to screen the CYP induction potential in vitro with more physiological relevance. Copyright

Bio-generation of stable isotope labeled internal standards for absolute and relative quantitation of drug metabolites in plasma samples by LC-MS/MS.

In order to achieve a better understanding of the toxicity of drug candidates, quantitative characterization of circulatory drug metabolites has been of increasing interest in current pharmaceutical research. Stable isotope labeled (STIL) internal standards (IS) are ideally used to simplify drug metabolite quantitation via liquid chromatography and tandem mass spectrometry (LC-MS/MS) analysis, primarily due to their capability to compensate matrix effects, thereby leading to faster method establishment by using generic assay conditions. However, chemical synthesis of STIL metabolites can often be resource intensive, requiring lengthy exploratory synthesis route development and/or extensive optimization to achieve the required stability for some metabolites. To overcome these challenges, we developed a general method that could generate STIL metabolites in a matter of hours from STIL parent drugs through the utilization of an appropriate in vitro metabolic incubation. This methodology can potentially save valuable synthesis resources, as well as provide timely availability of STIL IS. The following work demonstrates the proof-of-concept that multiple STIL metabolites can be generated simultaneously to provide satisfactory performance for both absolute quantitation of drug metabolites and for potential use in assessment of relative exposure coverage across species in safety tests of drug metabolites (MIST).

Functional Integrity of the Chimeric (Humanized) Mouse Liver: Enzyme Zonation, Physiologic Spaces, and Hepatic Enzymes and Transporters

Chimeric mouse liver models are useful in vivo tools for human drug metabolism studies; however, liver integrity and the microcirculation remain largely uninvestigated. Hence, we conducted liver perfusion studies to examine these attributes in FRGN [Fah(−/−), Rag2(−/−), and Il2rg(−/−), NOD strain] livers (control) and chimeric livers repopulated with mouse (mFRGN) or human (hFRGN) hepatocytes. In single-pass perfusion studies (2.5 ml/min), outflow dilution profiles of noneliminated reference indicators (51Cr-RBC, 125I-albumin, 14C-sucrose, and 3H-water) revealed preservation of flow-limited distribution and reduced water and albumin spaces in hFRGN livers compared with FRGN livers, a view supported microscopically by tightly packed sinusoids. With prograde and retrograde perfusion of harmol (50 µM) in FRGN livers, an anterior sulfation (Sult1a1) over the posterior distribution of glucuronidation (Ugt1a1) activity was preserved, evidenced by the 42% lower sulfation-to-glucuronidation ratio (HS/HG) and 14% higher harmol extraction ratio (E) upon switching from prograde to retrograde flow. By contrast, zonation was lost in mFRGN and hFRGN livers, with HS/HG and E for both flows remaining unchanged. Remnant mouse genes persisted in hFRGN livers (10%–300% those of FRGN). When hFRGN livers were compared with human liver tissue, higher UGT1A1 and MRP2, lower MRP3, and unchanged SULT1A1 and MRP4 mRNA expression were observed. Total Sult1a1/SULT1A1 protein expression in hFRGN livers was higher than that of FRGN livers, consistent with higher harmol sulfate formation. The composite data on humanized livers suggest a loss of zonation, lack of complete liver humanization, and persistence of murine hepatocyte activities leading to higher sulfation.

A nonradioactive approach to investigate the metabolism of therapeutic peptides by tagging with 127i and using inductively-coupled plasma mass spectrometry analysis.

The metabolic fate of adrenocorticotropic hormone (ACTH) fragment 4–10 (4–10) was evaluated following incorporation of a nonradioactive 127I-tag and with selective detection of I+ at m/z 127 by inductively coupled plasma mass spectrometry (ICP-MS). 127I has all the advantages of radioactive 125I as a metabolite tracer and, together with its detection in the femtogram range, has led to a successful metabolite profiling of 127I-ACTH (4–10) in vitro. The observed metabolic stability of this peptide in tissue preparations from human was plasma > kidney S9 > liver microsomes > liver cytosol, liver S9. Metabolic turnover of 127I-ACTH (4–10) was not NADPH-dependent and, together with inhibition by protease inhibitor cocktail and EDTA, is consistent with metabolism exclusively by proteases. Our preliminary studies using chemical inhibitors suggested the involvement of metalloprotease, serine peptidase, and aminopeptidase in 127I-ACTH (4–10) metabolism. The liver is the primary site of metabolic clearance of 127I-ACTH (4–10), with kidney S9 taking four times longer to produce a metabolite profile comparable to that produced by liver S9. A total of six metabolites retaining the 127I-tag was detected by ICP-MS, and their structures were elucidated using a LTQ/Orbitrap. 127I-ACTH (4–10) underwent both N- and C-terminal proteolysis to produce 127I-Phe as the major metabolite. The 127I-tag had minimal effect on the metabolic turnover and site of proteolysis of ACTH (4–10), which, together with ICP-MS providing essentially equimolar responses, suggests that the use of a 127I-tag may have general utility as an alternative to radioiodination to investigate the metabolism of peptide therapeutics