Human Liver Microsomes Immobilized on Magnetizable Beads: A Novel Approach to Study In Vitro Drug Metabolism

Abstract

Human liver microsomes (HLM) are a commonly used tool to study drug metabolism in vitro. Typical experiments conducted using suspensions of HLM can be challenging to separate from the incubation solution without lengthy ultracentrifugation steps. Magnetizable beads coated with silica (MGBS) were found to bind strongly to HLM, which could then be isolated and purified using a magnet. Binding of HLM to the MGBS (HLM-MGBS) was demonstrated to be mediated by strong interactions between microsomal phospholipids and MGBS, as artificially prepared phosphatidylcholine (PC) liposomes could be more efficiently captured by the MGBS. HLM-MGBS complexes retained functional cytochrome P450 and uridine-diphosphate-glucuronosyltransferase (UGT) activity as indicated by CYP2C8-mediated amodiaquine de-ethylation, CYP3A4-mediated midazolam 1'hydroxylation, UGT1A1-mediated glucuronidation of estradiol, UGT1A9-mediated glucuronidation of propofol, and UGT2B7-mediated glucuronidation of zidovudine. When comparing suspension HLM alone with HLM-MGBS complexes containing equivalent amounts of HLM, the intrinsic clearance (CL_int) of CYP450 substrates was comparable; however, CL_int of UGT1A1, UGT1A9, and UGT2B7 was increased in the HLM-MGBS system between 1.5- and 6-fold. HLM-MGBS used in an incubation could also be readily replaced with fresh HLM-MGBS to maintain the presence of active enzymes. Thus, HLM-MGBS demonstrate increased in vitro metabolic efficiency and manipulability, providing a new platform for determination of accurate metabolic parameters. SIGNIFICANCE STATEMENT: The following work describes the strong binding of HLM to magnetizable beads. In addition, the preservation of enzyme activity on the bound HLM provides a novel means to conduct preclinical metabolism studies.

Fig. 1.

General procedure for preparing and using HLM-MGBS.

Fig. 2.

Relationship between binding of HLM to magnetic silica beads. Binding was conducted over a period of 30 minutes, with constant mixing at a temperature of 4°C. Quantitation of HLM bound was conducted by measuring protein content remaining in the media after the incubation period. Each point represents the mean ± S.D. from N = 3 separate experiments.

Fig. 3.

Visualization of silica beads treated with and without HLM: (A) Bright-field image of beads alone; (B) HLM-MGBS after binding to fluorophore-glibenclamide drug conjugate; (C) Z-stacked three-dimensional image of HLM-MGBS treated with fluorophore-glibenclamide drug conjugate; (D) Bright-field image of HLM-MGBS merged with an extended-depth focused confocal fluorescence image produced from labeling HLM with fluorophore-glibenclamide drug conjugate. Scale bars, 10 μm.

Fig. 4.

Protein content in the media before and after incubation with 53 × 10⁹ beads per milliliter. The nominal concentration of HLM added to beads was 1 mg/ml. Protein content in the wash solutions was below the lower limit of quantitation (LLOQ, lower dotted line).

Fig. 5.

Binding of PC liposomes, HLM (0.5 mg/ml), and HSA (0.5 mg/ml) to silica magnetizable beads. A total of 11 × 10⁹ beads per milliliter were used across treatments. The content of PC used was set to equal the amount of phospholipids (PL) measured in 0.5 mg/ml HLM. Values represent the means (±S.D.) from N = 3 separate experiments. The amount of PL measureable in the buffer was below the limit of quantitation in PC samples treated with beads. As such, the percentage of phospholipids bound to beads was set to a value of 100.

Fig. 6.

Effect of preloading MGBS with PC on binding of HLM to MGBS. A total of 5.3 × 10⁹ beads per milliliter were used across treatments. Each bar represents the mean ± S.D. from triplicate measurements.

Fig. 7.

Amodiaquine de-ethylation and midazolam 1′-hydroxylation in HLM vs. HLM-MGBS. The HLM concentration used was 0.25 mg/ml. (A) Hyperbolic curve for CYP2C8-mediated amodiaquine de-ethylation. (B) Hyperbolic curves for CYP3A4-mediated midazolam hydroxylation. Values represent the means ± S.D. from N = 3 replicates for each substrate concentration evaluated.

Fig. 8.

Estradiol, propofol, and AZT glucuronidation in HLM vs. HLM-MGBS. (A) Sigmoidal curve of UGT1A1-catalyzed estradiol 3-glucuronidation. (B) Hyperbolic curve of UGT1A9-mediated propofol glucuronidation. (C) Hyperbolic curves of UGT2B7-mediated AZT glucuronidation. Data points represent the means ± S.D. from three replications for each substrate concentration evaluated.

Fig. 9.

HLM-MGBS relay strategy to deplete midazolam (A) and form 1′OH midazolam (B). When HLM-MGBS were used, each time point represents one exchange of used HLM-MGBS with new HLM beads in a new incubation container. With HLM suspensions, each time point represents one exchange of the entire incubation into a new incubation container. The HLM protein content in both reactions was adjusted to 0.005 mg/ml.