In vitrometabolic mapping of neobavaisoflavone in human cytochromes P450 and UDP-glucuronosyltransferase enzymes by ultra high-performance liquid chromatography coupled with quadrupole time-of-flight tandem mass spectrometry

Abstract

Neobavaisoflavone (NBIF), a phenolic compound isolated from Psoralea corylifolia L., possesses several significant biological properties. However, the pharmacokinetic behaviors of NBIF have been characterized as rapid oral absorption, high clearance, and poor oral bioavailability. We found that NBIF underwent massive glucuronidation and oxidation by human liver microsomes (HLM) in this study with the intrinsic clearance (CL_int) values of 12.43, 10.04, 2.01, and 6.99 μL/min/mg for M2, M3, M4, and M5, respectively. Additionally, the CL_int values of G1 and G2 by HLM were 271.90 and 651.38 μL/min/mg, respectively, whereas their respective parameters were 59.96 and 949.01 μL/min/mg by human intestine microsomes (HIM). Reaction phenotyping results indicated that CYP1A1, 1A2, 2C8, and 2C19 were the main contributors to M4 (34.96 μL/min/mg), M3 (29.45 μL/min/mg), M3 (13.16 μL/min/mg), and M2 (63.42 μL/min/mg), respectively. UGT1A1, 1A7, 1A8, and 1A9 mainly catalyzed the formation of G1 (250.87 μL/min/mg), G2 (438.15 μL/min/mg), G1 (92.68 μL/min/mg), and G2 (1073.25 μL/min/mg), respectively. Activity correlation analysis assays showed that phenacetin-N-deacetylation was strongly correlated to M3 (r = 0.860, p = 0.003) and M4 (r = 0.775, p = 0.014) in nine individual HLMs, while significant activity correlations were detected between paclitaxel-6-hydroxylation and M2 (r = 0.675, p = 0.046) and M3 (r = 0.829, p = 0.006). There was a strong correlation between β-estradiol-3-O-glucuronide and G1 (r = 0.822, p = 0.007) and G2 (r = 0.689, p = 0.040), as well as between propofol-O-glucuronidation and G1 (r = 0.768, p = 0.016) and G2 (r = 0.860, p = 0.003). Moreover, the phase I metabolism and glucuronidation of NBIF revealed marked species differences, and mice are the best animal model for investigating the metabolism of NBIF in humans. Taken together, characterization of NBIF-related metabolic pathways involving in CYP1A1, 1A2, 2C8, 2C19, and UGT1A1, 1A7, 1A8, 1A9 are helpful for understanding the pharmacokinetic behaviors and conducting in-depth pharmacological studies.

1. Introduction

Neobavaisoflavone (NBIF) is one of the most abundant bioactive compounds isolated from the seeds of Psoralea corylifolia L., which are widely used to invigorate the kidneys and strengthen Yang in clinical practices using herbal medicines and dietary supplements [1]. NBIF accounts for 0.1–0.9% of the dried seed weight [2,3]. Numerous studies have examined NBIF in the fields of pharmacology to study the inhibition of platelet aggregation (IC₅₀ = 2.5 μM) [4], anti-inflammatory activity [5], enhancing tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-mediated apoptosis in prostate cancer cells [6], activation of both estrogen receptor (ER) subtypes [7], and α-glucosidase inhibitory activities (IC₅₀ = 27.7 μM) [8]. Additionally, NBIF is considered a potential anabolic agent for treating bone loss-associated diseases [9]. A combination of NBIF and TRAIL may be a new therapeutic strategy for treating TRAIL-resistant glioma cells [10].

These important biological activities have attracted attention in related metabolism and pharmacokinetics studies. Our previous studies indicated that NBIF mainly underwent oxidization, hydroxylation, hydration, cyclization, epoxidation, sulfation, and glucuronidation reactions in rats [11], while hydroxylation, hydration and epoxidation reactions should be the main biotransformations in rat intestinal microflora [12]. Additionally, pharmacokinetic studies reported that NBIF was quickly absorbed into the rat plasma with a maximal plasma concentration of less than 18.37 ng/mL after oral administration of Psoralea corylifolia extracts or P. corylifolia-containing prescriptions [13-15]. NBIF also readily penetrated the blood brain barrier and accumulates in the brain [15].

In contrast, NBIF exhibited significant inhibitory effects on drug-metabolizing enzymes and nuclear receptors. For example, NBIF displayed obvious inhibitory effects on UDP-glucuronosyltransferase 1A1 (UGT1A1) and human liver microsomes (HLM) with IC₅₀ values of 2.42 and 2.25 μM, respectively [16]. Additionally, NBIF noncompetitively inhibits the activities of human carboxylesterase 1 [17] and human carboxylesterase 2 (IC₅₀ = 6.39 μM) [18]. NBIF showed weak inhibition and selectivity of human monoamine oxidase-A and -B (IC₅₀ = 66.52 and 77.29 μM, respectively) [19]. Furthermore, NBIF showed the peroxisome proliferator activated receptor-γ (PPAR-γ) agonist activity [20]. However, the cytochromes P450 (CYP) and UGT metabolizing enzymes actively involved in metabolizing NBIF remain unknown.

Therefore, we investigated the metabolic pathways of NBIF in the present study to identify the corresponding main CYP and UGT enzymes and characterize species differences by HLM and animal liver microsomes. Metabolic rates were determined by incubating NBIF with nicotinamide adenine dinucleotide phosphate (NADPH)− and uridine diphosphoglucuronic acid (UDPGA)-supplemented microsomes. Kinetic parameters were derived by fitting an appropriate model to the data. A series of independent assays including reaction phenotyping and activity correlation analysis were performed by ultra high-performance liquid chromatography coupled with quadrupole time-of-flight tandem mass spectrometry (UHPLC/Q-TOF-MS) to identify the main CYPs and UGTs contributing to the metabolism of NBIF. Our results would improve the understanding of the metabolic fates of NBIF.

2. Experimental

2.1. Chemicals and reagents

Neobavaisoflavone (purity > 98%) was purchased from Shanghai Winherb Medical Technology Co., Ltd. (Shanghai, China). Alamethicin, D-saccharic-1, 4-lactone, MgCl₂, NADPH, and UDPGA were from Sigma-Aldrich (St. Louis, MO, USA). Human liver microsomes (HLM) and human intestine microsomes (HIM) were pooled from 20 and 50 mixed gender donors, respectively, and were both purchased form Corning Biosciences (Corning, NY, USA). Dogs liver microsomes (DLM), expressed human CYP enzymes (CYP1A1, 1A2, 1B1, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, and 3A5), individual pooled human liver microsomes (iHLM), mice liver microsomes (MLM), monkeys liver microsomes (MkLM), rabbits liver microsomes (RaLM), rats liver microsomes (RLM), and recombinant UGT enzymes (UGT1A1, 1A3, 1A4, 1A6, 1A7, 1A8, 1A9, 1A10, 2B4, 2B7, 2B10, 2B15, and 2B17) were obtained from Corning Biosciences. Phenacetin, paracetamol, paclitaxel, 6α-hydroxy-paclitaxel, β-estradiol, and propofol and were purchased from Aladdin Chemicals (Shanghai, China). Propofol-O-glucuronide and β-estradiol-3-O-glucuronide were obtained from Toronto Research Chemicals (Ontario, Canada). All other chemicals and reagents were of analytical grade or the highest grade commercially available.

2.2. In vitro metabolism assay

In this study, the incubation system (100 μL) for phase I metabolism contained Tris-HCl buffer solution (50 mM, pH 7.4), drug-metabolizing enzyme solutions (0.5 mg/mL), MgCl₂ (5 mM), a series of NBIF solutions (0.16–80 μM), and NADPH solution (1 mM) as described previously [21]. The reactions were terminated by adding cold acetonitrile (100 μL) after incubation at 37 °C for 1 h. Furthermore, after centrifugation at 13,800 × g for 10 min, the supernatant (8 μL) was injected into an UHPLC/Q-TOF-MS system. A sample incubated without NADPH served as the negative control to confirm that the metabolites produced were NADPH-dependent. Similarly, NBIF was incubated with HIM, 12 iHLM samples, animal liver microsomes, and recombinant CYP enzymes in the incubation system. All experiments were performed in triplicate.

For glucuronidation assays, a series of NBIF solutions (0.16–80 μM) was incubated in Tris-HCl buffer (50 mM, pH 7.4), MgCl₂ (4.0 mM), alamethicin (22 μg/mL), saccharolactone (4.4mM), and UDPGA solutions (3.5 mM) [21]. After 1 h, equal volumes of ice-cold acetonitrile (100 μL) were added to the incubation mixture to terminate the reactions. Additionally, the supernatant (8 μL) was subjected to UHPLC/Q-TOF-MS analysis after centrifugation at 13,800 × g for 10 min. Incubation without UDPGA served as the negative control to confirm that the metabolites produced were UDPGA-dependent. Similarly, the incubation system for HIM, 12 iHLM samples, animal liver microsomes, and each expressed CYP enzyme was similar to the microsomal incubation system described above. All experiments were performed in triplicate.

2.3. UHPLC/Q-TOF-MS analysis

UHPLC analysis was performed on an Acquity™ UHPLC I-Class system equipped with PDA detector (Waters Corporation, Milford, MA, USA). Chromatographic separation was performed on a BEH C₁₈ column (2.1 × 50mm, 1.7 μm, Waters, Part No. 186002350) maintained at 35 °C and with water (A) and acetonitrile (B) (both containing 0.1% formic acid, V/V) as the mobile phase. For phase I metabolism, the gradient elution program was 10–50% B from 0 to 2.0 min, 50–100% B from 2.0 to 3.0 min, maintained at 100% B from 3.0 to 3.2 min, 100–10% B from 3.2 to 3.5 min, and maintained at 10% B from 3.5 to 4.0 min at a flow rate of 0.5 mL/min. For glucuronidation assays, the gradient elution program was 10–30% B from 0 to 2.0 min, 30–33% B from 2.0 to 4.0 min, 33–90% B from 4.0 to 4.7 min, 90–100% B from 4.7 to 4.8 min, maintained at 100% B from 4.8 to 5.0 min, 100–10% B from 5.0 to 5.2 min, and maintained at 10% from 5.2 to 5.5 min at a flow rate of 0.4 mL/min. The detection wavelength was 254 nm.

The UHPLC system was coupled to a hybrid quadrupole orthogonal time-of-flight (Q-TOF) tandem mass spectrometer (SYNAPT™ G2 HDMS; Waters) equipped with electrospray ionization (ESI). The operating parameters were as follows: capillary voltage, 3 kV (ESI+); sample cone voltage, 35 V; extraction cone voltage, 4 V; source temperature, 100 °C; desolvation temperature, 300 °C; cone gas flow, 50 L/h and desolvation gas flow, 800 L/h. The full scan mass range was 50–1500 Da. Leucine enkephalin was dissolved in acetonitrile-water (50:50, v/v, containing 0.1% formic acid) at 200pg/mL (m/z 556.2771 in positive ion mode), and was used as an external reference of LockSpray™ infused at a constant flow of 5 μL/min to ensure mass accuracy.

2.4. Enzyme kinetics evaluation

A series of NBIF concentrations (0.16–80 μM) was incubated with pooled HLM, HIM, each iHLM, expressed CYP enzymes, and recombinant UGT enzymes to determine the metabolic rates. Two kinetic models, the Michaelis-Menten equation and substrate inhibition equation, were fitted to the data of metabolic activities versus substrate concentrations and displayed in Eqs. (1) and (2), respectively. Appropriate models were selected by visual inspection of the Eadie-Hofstee plot [21]. Model fitting and parameter estimation were performed using GraphPad Prism V5 software (GraphPad, Inc., San Diego, CA, USA).

The corresponding kinetic parameters were as follows. V is the formation rate of NBIF-related metabolites, V_max is the maximal velocity. K_m is the Michaelis constant, [S] is the substrate concentration, and K_si is the substrate inhibition constant. The intrinsic clearance (CL_int) values were derived from V_max/K_m for Michaelis-Menten and substrate inhibition models [21].

$$V=\frac{V\max \times [S]}{Km+[S]}$$ (1)

$$V=\frac{V\max \times [S]}{Km+[S]\left(1+\frac{[S]}{Ksi}\right)}$$ (2)

2.5. Activity correlation analysis

According to a previously described protocol, the metabolic rates of NBIF, phenacetin (probe substrate for CYP1A2) [21], paclitaxel (selective substrate for CYP2C8) [22], β-estradiol (probe substrate for UGT1A1) [23], and propofol (specific substrate for UGT1A9) [23] were determined for individual HLMs (n = 9). NBIF (5 μM), phenacetin (200 μM), and paclitaxel (20 μM) were incubated with NADPH-supplemented individual HLMs (n = 9), while NBIF (5 μM), β-estradiol (50 μM), and propofol (500 μM) were incubated with UDPGA-supplemented individual HLMs (n = 9). Furthermore, correlation analysis was performed between NBIF-oxidation (M1–M5) and phenacetin-N-deacetylation (paracetamol) and paclitaxel-6-hydroxylation (6α-hydroxy-paclitaxel), respectively. Correlation analysis was performed between NBIF-glucuronidation (G1 and G2) and β-estradiol-3-O-glucuronidation (β-estradiol-3-O-glucuronide) and propofol-O-glucuronidation (propofol-O-glucuronide), respectively. These correlation (Pearson) analyses were performed using GraphPad Prism V5 software.

2.6. Species difference analysis

In this study, a series of NBIF solutions (0.16–80 μM) was incubated with five animal liver microsomes to determine the oxidation (M1–M5) and glucuronidation (G1 and G2) rates, respectively. Kinetic parameters were derived from the appropriate model fitting. Additionally, the CL_int values for NBIF oxidation and glucuronidation by HLM and different animal liver microsomes were used as the evaluation parameters to estimate species diversity as published previously [21,23].

2.7. Statistical analysis

Data are displayed as the mean ± SD (standard deviation). Mean differences between the treatment and control groups were analyzed by two-tailed Student’s t test. The level of significance was set to p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***).

3. Results

3.1. Structural elucidation of NBIF-related metabolites by CYP and UGT enzymes

After incubation with NBIF by NADPH-supplemented HLM, five additional obvious peaks were detected by UHPLC/Q-TOF-MS with extracted ion chromatograms as shown in Fig. 1A. M0 was unambiguously identified as NBIF based on the reference standard. The mother ion at m/z 323.1286 (C₂₀H₁₉O₄, 0.9 ppm, Fig. S1a) showed daughter ions at 267.066, 255.065, 239.075, and 137.023, which agreed with the results of our previous study [11 ]. Additionally, we characterized the oxidation sites of NBIF according to the presence or absence of the diagnostic ion at m/z 137.023 [11]. When a fragment ion was observed at m/z 137.023, the oxidation substituent was linked at the isopentenyl group of NBIF, whereas the oxidation position was at the A ring of NBIF when the characteristic ion at m/z 137.023 was absent [11]. Therefore, M1 (C₂₀H₁₉O₆, −0.8 ppm, Fig. S1b) was characterized as di-oxidated NBIF with two oxidation units at the isopentenyl group of NBIF. Similarly, M2 (C₂₀H₁₉O₅, 1.8 ppm, Fig. S1c) and M3 (C₂₀H₁₉O₅, −2.1 ppm, Fig. S1d) were both considered as mono-oxidated NBIFs with the oxidation position at the isopentenyl unit, while the oxidation units of M4 (C₂₀H₁₉O₅, −1.5 ppm, Fig. S1e) and M5 (C₂₀H₁₉O₅, −2.9 ppm, Fig. S1f) were both on the A ring of NBIF.

Fig. 1.

Extracted ion chromatograms of NBIF, di-oxidated NBIF, mono-oxidated NBIF after incubation of NBIF with NADPH-supplemented HLM (A), and mono-glucuronidated NBIF after incubation of NBIF with UDPGA-supplemented HLM (B).

In contrast, when NBIF was incubated with UDPGA-supplemented HLM, two metabolites were detected by UHPLC/Q-TOF-MS (Fig. 1B). G1 (C₂₆H₂₇O₁₀, 1.4 ppm, Fig. S1g) and G2 (C₂₆H₂₇O₁₀, −0.8 ppm, Fig. S1h) both displayed the [M+H]⁺ ion at m/z 499.1611 which were 176.0321 Da higher than NBIF (M0) and similar to that of NBIF. Additionally, the fragment ion at m/z 113.026 suggested that G1 and G2 were mono-glucuronidated NBIF. According to the ClogP values [11], G1 (CLogP = 1.841) and G2 (CLogP = 2.084) were considered as NBIF-7-O-glucuronide and NBIF-4′-O-glucuronide, respectively. The UHPLC/Q-TOF-MS data for NBIF and its related metabolites are shown in Table 1.

Table 1.

UHPLC/Q-TOF-MS data of NBIF and its related metabolites.

NO.	RT (min)	[M+H]+ ion	Formula	Error (ppm)	(+) ESI-MS/MS	Identification
M0	2.65	323.1286	C20H18O4	0.9	267.066 255.065 239.075 137.023	NBIF
M1	1.55	355.1179	C20H18O6	−0.8	279.063 137.024	di-oxidated-NBIF
M2	1.89	339.1238	C20H18O5	1.8	321.113 279.066 137.024	mono-oxidated-NBIF
M3	2.07	339.1225	C20H18O5	−2.1	321.113 279.067 137.025	mono-oxidated-NBIF
M4	2.24	339.1227	C20H18O5	−1.5	283.060 271.060 255.064 181.069	mono-oxidated-NBIF
M5	2.36	339.1222	C20H18O5	−2.9	283.062 271.061 255.065 181.070	mono-oxidated-NBIF
G1	2.70	499.1611	C26H26O10	1.4	323.131 267.065 255.066 113.026	mono-glucuronidated-NBIF
G2	2.83	499.1600	C26H26O10	−0.8	323.131 267.065 255.066 113.026	mono-glucuronidated-NBIF

Note: NBIF, neobavaisoflavone.

3.2. Semi-quantification of NBIF and NBIF-related metabolites

Because of the lack of reference standards, semi-quantification of NBIF-related metabolites was based on the standard curve of NBIF according to the assumption that NBIF and its metabolites have similar UV absorbance maxima [21,23]. Thus, a practical UHPLC method was developed and validated for the semi-quantification of NBIF and its related metabolites. The limit of detection and limit of quantification were 0.005 and 0.01 μM, respectively. Additionally, an acceptable linear correlation (Y = 15392X) was confirmed by the correlation coefficient (r²) of 0.9994 between 0.01 and 20 μM.

3.3. Phase I metabolism and glucuronidation of NBIF in HLM and HIM

Kinetic profiling indicated that the mono-oxidation activities of NBIF (M2, M3, M4, and M5) in HLM followed classical Michaelis-Menten kinetics as shown in Fig. S2a and b. The intrinsic clearance (CL_int) values of M2, M3, M4, and M5 were 12.43, 10.04, 2.01, and 6.99 μL/min/mg, respectively (Table 2). Because the concentration was under the limit of quantification, we could not determine the kinetic parameters of M1 in HLM. Additionally, the phase I metabolism of NBIF in HIM was clearly weaker than that in HLM. Only M3 was detected after incubation with NBIF in HIM. Similarly, we could not obtain the complete kinetic profile of M3 in HIM.

Table 2.

Kinetic parameters derived for NBIF-related oxidated metabolites (M1~M5) by HLM, HIM, expressed CYP enzymes, MkLM, RLM, MLM, DLM and RaLM (mean±SD). All experiments were performed in triplicate.

	Meta.	Vmax (pmol/min/mg)	Km (μM)	Ki (μM)	CLint (μL/min/mg)	Model
HLM	M1	++	++	++	++
	M2	67.28 ± 3.68	5.41 ± 0.91	N.A.	12.43 ± 2.19	MM
	M3	33.60 ± 1.39	3.35 ± 0.48	N.A.	10.04 ± 1.49	MM
	M4	9.31 ± 0.40	4.64 ± 0.64	N.A.	2.01 ± 0.29	MM
	M5	15.53 ± 0.30	2.22 ± 0.16	N.A.	6.99 ± 0.53	MM
HIM	M3	+	+	+	+
CYP1A1	M1	30.63 ± 1.10	5.98 ± 0.78	N.A.	5.12 ± 0.67	MM
	M2	22.12 ± 3.13	2.62 ± 0.73	30.85 ± 10.56	8.43 ± 2.62	SI
	M3	32.29 ± 2.38	1.06 ± 0.20	38.0 ± 8.78	30.46 ± 6.09	SI
	M4	130.0 ± 32.38	3.72 ± 1.53	14.23 ± 6.25	34.96 ± 16.84	SI
	M5	26.1 ± 2.33	1.38 ± 0.29	26.54 ± 5.97	18.89 ± 4.29	SI
CYP1A2	M2	36.27 ± 1.07	23.73 ± 1.72	N.A.	1.53 ± 0.12	MM
	M3	601.9 ± 20.48	20.44 ± 1.78	N.A.	29.45 ± 2.75	MM
	M4	266.6 ± 9.92	34.94 ± 2.81	N.A.	7.63 ± 0.68	MM
	M5	36.83 ± 1.63	24.02 ± 2.60	N.A.	1.53 ± 0.18	MM
CYP1B1	M1	+	+	+	+
	M2	+	+	+	+
	M3	+	+	+	+
	M4	+	+	+	+
CYP2A6	M3	+	+	+	+
CYP2B6	M1	+	+	+	+
	M3	+	+	+	+
CYP2C8	M1	+	+	+	+
	M2	+	+	+	+
	M3	121.8 ± 2.51	9.25 ± 0.61	N.A.	13.16 ± 0.90	MM
	M4	+	+	+	+
	M5	+	+	+	+
CYP2C9	M3	+	+	+	+
	M4	+	+	+	+
CYP2C19	M1	+	+	+	+
	M2	912.0 ± 276.0	14.38 ± 5.79	15.94 ± 7.30	63.42 ± 31.94	SI
	M3	197.5 ± 33.18	12.63 ± 2.82	12.77 ± 3.133	15.64 ± 4.37	SI
	M4	52.38 ± 20.68	39.66 ± 19.72	24.84 ± 13.56	1.32 ± 0.84	SI
	M5	+	+	+	+
CYP2D6	M3	+	+	+	+
	M4	++	++	++	++
	M5	++	++	++	++
CYP2E1	M3	+	+	+	+
CYP3A5	M3	++	++	++	++
MkLM	M1	+	+	+	+	+
	M2	84.31 ± 1.83	2.81 ± 0.22	N.A.	29.99 ± 2.41	MM
	M3	32.01 ± 1.19	2.29 ± 0.23	138.6 ± 21.68	13.98 ± 1.47	SI
	M4	72.37 ± 1.75	2.24 ± 0.20	N.A.	32.38 ± 3.03	MM
	M5	38.28 ± 5.80	12.22 ± 3.12	86.0 ± 31.3	3.13 ± 0.93	SI
RLM	M1	+	+	+	+
	M2	100.5 ± 2.36	1.43 ± 0.14	N.A.	70.28 ± 6.93	MM
	M3	134.0 ± 5.23	2.66 ± 0.22	57.83 ± 7.34	50.41 ± 4.55	SI
	M4	295.9 ± 7.60	15.51 ± 1.10	N.A.	19.08 ± 1.44	MM
MLM	M1	+	+	+	+
	M2	34.93 ± 1.01	3.42 ± 0.39	N.A.	10.23 ± 1.21	MM
	M3	248.2 ± 7.12	22.75 ± 1.62	N.A.	10.91 ± 0.84	MM
DLM	M3	141.3 ± 9.61	21.0 ± 3.63	N.A.	6.73 ± 1.25	MM
	M4	15.71 ± 0.51	2.17 ± 0.31	N.A.	7.25 ± 1.07	MM
RaLM	M2	50.27 ± 2.31	3.47 ± 0.39	168.6 ± 33.94	14.49 ± 1.74	SI
	M3	311.5 ± 7.64	20.87 ± 1.30	N.A.	14.93 ± 1.00	MM
	M4	27.95 ± 0.51	1.22 ± 0.11	N.A.	22.83 ± 2.13	MM

Note: Meta., metabolites; N.A., not available; +, under the limit of quantification; ++, unable to determine the kinetic parameters in the absence of a full kinetic profile; HLM, human liver microsomes; HIM, human intestine microsomes; MkLM, monkeys liver microsomes; RLM, rats liver microsomes; MLM, mice liver microsomes; DLM, dogs liver microsomes; RaLM, rabbits liver microsomes; SI, substrate inhibition model; MM, Michaelis-Menten model.

Compared to phase I metabolism, glucuronidation of NBIF was more efficient in both HLM and HIM. The formation of G1 in HLM was well-modeled by the substrate inhibition equation (Fig. S2c), while G2 followed the classical Michaelis-Menten equation (Fig. S2c). In contrast, the metabolic rates of G1 and G2 in HIM followed the Michaelis-Menten kinetic and substrate inhibition equation, respectively (Fig. S2d). Additionally, the CL_int values of G1 and G2 in HLM were 271.90 and 651.38 μL/min/mg (Table 3), respectively, whereas their corresponding parameters of G1 and G2 were 59.96 and 949.01 μL/min/mg in HIM, respectively (Table 3). The glucuronidation activities of NBIF far outweighed NBIF-related phase I metabolism.

Table 3.

Kinetic parameters derived for NBIF-related glucuronides (G1 and G2) by HLM, HIM, expressed UGT enzymes, MkLM, RLM, MLM, DLM and RaLM (mean ± SD). All experiments were performed in triplicate.

Enzyme	Meta.	Vmax (pmol/min/mg)	Km (μM)	Ki (μM)	CLint (μL/min/mg)	Model
HLM	G1	1788.0 ± 203.6	6.58 ± 1.25	48.45 ± 13.86	271.90 ± 60.39	SI
	G2	919.1 ± 20.23	1.41 ± 0.13	N.A.	651.38 ± 60.31	MM
HIM	G1	409.3 ± 18.9	6.83 ± 0.91	N.A.	59.96 ± 8.42	MM
	G2	8655.0 ± 1021.0	9.12 ± 1.64	33.32 ± 7.95	949.01 ± 203.75	SI
UGT1A1	G1	912.4 ± 49.36	3.64 ± 0.48	291.1 ± 100.1	250.87 ± 35.80	SI
	G2	48.33 ± 1.59	1.03 ± 0.11	401.2 ± 135.7	47.01 ± 5.35	SI
UGT1A7	G1	11.93 ± 0.55	0.41 ± 0.06	61.82 ± 13.5	28.87 ± 4.57	SI
	G2	135.3 ± 4.36	0.31 ± 0.04	154.5 ± 44.85	438.15 ± 56.41	SI
UGT1A8	G1	419.1 ± 18.69	4.52 ± 0.54	N.A.	92.68 ± 11.73	MM
	G2	147.0 ± 6.84	4.0 ± 0.51	N.A.	36.75 ± 5.00	MM
UGT1A9	G1	70.01 ± 1.37	1.77 ± 0.14	N.A.	39.49 ± 3.11	MM
	G2	805.9 ± 17.46	0.75 ± 0.07	N.A.	1073.25 ± 108.93	MM
UGT2B4	G2	+	+	+	+
MkLM	G1	4129.0 ± 182.8	11.39 ± 1.25	N.A.	362.51 ± 42.87	MM
	G2	7072.0 ± 228.4	5.62 ± 0.55	N.A.	1257.92 ± 129.47	MM
RLM	G1	2490.0 ± 249.5	6.40 ± 1.01	24.73 ± 4.75	388.88 ± 72.83	SI
	G2	826.6 ± 99.48	2.95 ± 0.57	9.09 ± 1.82	280.11 ± 63.39	SI
MLM	G1	3589.0 ± 432.9	7.41 ± 1.43	38.19 ± 10.18	484.67 ± 110.41	SI
	G2	3878.0 ± 563.1	8.84 ± 1.96	31.68 ± 9.16	438.64 ± 116.08	SI
DLM	G1	1807.0 ± 48.14	9.31 ± 0.65	N.A.	194.09 ± 14.55	MM
	G2	3363.0 ± 85.47	3.48 ± 0.30	N.A.	965.27 ± 86.90	MM
RaLM	G1	3006.0 ± 103.8	4.82 ± 0.43	N.A.	623.26 ± 60.06	MM
	G2	3411.0 ± 155.5	4.13 ± 0.51	N.A.	826.61 ± 109.47	MM

Note: Meta., metabolites; N.A., not available; +, under the limit of quantification; HLM, human liver microsomes; HIM, human intestine microsomes; MkLM, monkeys liver microsomes; RLM, rats liver microsomes; DLM, dogs liver microsomes; MLM, mice liver microsomes; RaLM, rabbits liver microsomes; SI, substrate inhibition model; MM, Michaelis-Menten model.

3.4. Reaction phenotyping by CYP and UGT enzymes

To determine which enzymes have the greatest contribution to phase I metabolism and glucuronidation, two test concentrations of NBIF were incubated with a series of expressed CYPs and UGTs enzymes to examine their catalysis activities (expressed as pmol/min/mg protein). The formation of M1 was mainly attributed to CYP1A1 (Fig. 2A and B), while M3 was produced by CYP1A1, 1A2, 2C8, and 2C19 (Fig. 2A and B). NBIF was catalyzed by CYP1A1, 1A2, and 2C19 to form M2 and M4 (Fig. 2A and B). Additionally, M5 was mainly formed by CYP1A1 and 1A2 (Fig. 2A and B). Moreover, CYP1B1, 2A6, 2B6, 2C9, 2D6, 2E1, and 3A5 contributed to the formation of these five oxidated products of NBIF (Table 2). However, because the concentration of NBIF-related metabolites was lesser than the limit of quantification, we were unable to determine the complete kinetic parameters of these CYP enzymes.

Fig. 2.

Comparisons of phase I metabolism rates (A: 2.5 μM; B: 20 μM) and glucuronidation rates (C: 0.625 μM; D: 10 μM) of NBIF by expressed CYP and UGT enzymes. All experiments were performed in triplicate.

As shown in Fig. 2C and D, the formation of G1 and G2 mainly contributed to expression of the UGT1A1, 1A7, 1A8, and 1A9 enzymes. UGT2B4 also catalyzed the production of G2 (Table 3) but the concentration was under the limit of quantification. Furthermore, other test UGT enzymes could not produce G1 and G2.

3.5. Phase I metabolism and glucuronidation kinetics by active CYP and UGT enzymes

According to the reaction phenotyping results, CYP1A1, 1A2, 2C8, and 2C19 were the main active CYP enzymes involved in the phase I metabolism of NBIF (Fig. 2A and B). The formation of M1 by CYP1A1 followed classical Michaelis-Menten kinetics (Fig. S3a), whereas the metabolic rates of M2, M3, M4, and M5 were all modeled by the substrate inhibition equation (Fig. S3b and c), which did not always follow the same kinetics as in HLM. The corresponding CL_int values were 5.12, 8.43, 30.46, 34.96, and 18.89 μL/min/mg, respectively (Table 2). Additionally, CYP1A2 catalyzed the formation of M2, M3, M4, and M5 following Michaelis-Menten kinetics (Fig. S3d and e) with CL_int values of 1.53, 29.45, 7.63, and 1.53 μL/min/mg, respectively (Table 2), which agreed with the glucuronidation profiles in HLM. Notably, M3 was formed by CYP2C8 following Michaelis-Menten kinetics (Fig. S3f), while the relative CL_int value was 13.16 μL/min/mg (Table 2). Furthermore, the formation of M2, M3, and M4 all followed the substrate inhibition equation (Fig. S3g and h) with CL_int values of 63.42, 15.64, and 1.32 μL/min/mg, respectively (Table 2). Notably, M1, M2, M3, M4, and M5 were mainly formed by CYP1A1 (5.12 μL/min/mg), 2C19 (63.42 μL/min/mg), 1A1 (30.46 μL/min/mg), 1A1 (34.96 μL/min/mg), and 1A1 (18.89 μL/min/mg), respectively (Fig. 3A). Therefore, CYP1A1, 1A2, 2C8, and 2C19 were the main contributors to NBIF oxidation.

Fig. 3.

Intrinsic clearance (CL_int) values of NBIF-related metabolites by expressed CYPs (A), recombinant UGTs (B), HLM and five animal liver microsomes for phase I metabolism (C) and glucuronidation (D). N.D., not detected. N.A., under the limit of quantification or unable to determine the kinetic parameters in the absence of a full kinetic profile. ^{a, b, c, d} compared with the CL_int values of NBIF-related metabolites in HLM, respectively. (^{a, b, c, d} p < 0.05, ^{aa, bb, cc, dd} p < 0.01, ^{aaa, bbb, ccc, ddd} p < 0.001).

As clearly shown in Fig. 2C and D, UGT1A1, 1A7, 1A8, and 1A9 were active contributors to the glucuronidation of NBIF. Glucuronidation reactions (250.87 μL/min/mg for G1 and 47.01 μL/min/mg for G2) mediated by UGT1A1 both followed substrate inhibition kinetics (Fig. 4a), which agreed with the glucuronidation profiles (G1) in HLM (Fig. S2c). Similarly, the production of G1 and G2 by UGT1A7 followed the substrate inhibition equation (Fig. 4b) with CL_int values of 28.87 and 438.15 μL/min/mg, respectively (Table 3). Additionally, glucuronidation of NBIF by UGT1A8 and 1A9 all were modeled by classical Michaelis-Menten kinetics (Fig. 4c and d). The CL_int values for G1 and G2 by UGT1A8 were 92.68 and 36.75 μL/min/mg (Table 3), respectively, and the CL_int values of 39.49 and 1073.25 μL/min/mg (Table 3) by UGT1A9, respectively. In general, UGT1A1, 1A7, 1A8, and 1A9 were mainly responsible for the glucuronidation of NBIF (Fig. 3B).

Fig. 4.

Metabolic pathways of NBIF involving in CYPs and UGTs.

3.6. Activity correlation analysis by active CYP and UGT enzymes

The phase I metabolism and glucuronidation activities of individual HLMs (n = 9) toward NBIF and probe substrates of CYP1A2, 2C8 [21,22] and UGT1A1, 1A9 [23] were evaluated. We found that phenacetin-N-deacetylation was strongly correlated with M3 (r = 0.860, p = 0.003, Fig. S5b) and M4 (r = 0.775, p = 0.014, Fig. S5c), indicating that CYP1A2 plays a critical role in the formation of M3 and M4. In contrast, there were no significant differences between phenacetin-N-deacetylation and M2 (r = 0.525, p = 0.147, Fig. S5a), M5 (r = 0.382, p = 0.310, Fig. S5d). Additionally, significant activity correlation was observed between paclitaxel-6-hydroxylation and M2 (r = 0.675, p = 0.046, Fig. S6a) and M3 (r = 0.829, p = 0.006, Fig. S6b), while no obvious differences (p > 0.05) were detected between paclitaxel-6-hydroxylation and M4 (r = 0.577, p = 0.104, Fig. S6c) and M5 (r = 0.284, p = 0.460, Fig. S6d). The detailed correlation factors are shown in Table 4. These finding suggest that CYP1A2 is the greatest contributor to M3 and M4 production, whereas CYP2C8 plays an important role in M2 and M3 formation.

Table 4.

Metabolic activities correlation analysis of individual HLMs (n = 9) toward NBIF, phenacetin, paclitaxel, β-estradiol and propofol.

Enzyme	Probe substrate	Metabolite	r value	p value
CYP1A2	phenacetin-N-deacetylation	NBIF-oxidation (M2)	0.525	0.147
CYP1A2	phenacetin-N-deacetylation	NBIF-oxidation (M3)	0.860	0.003, **
CYP1A2	phenacetin-N-deacetylation	NBIF-oxidation (M4)	0.775	0.014, *
CYP1A2	phenacetin-N-deacetylation	NBIF-oxidation (M5)	0.382	0.310
CYP2C8	paclitaxel-6-hydroxylation	NBIF-oxidation (M2)	0.675	0.046, *
CYP2C8	paclitaxel-6-hydroxylation	NBIF-oxidation (M3)	0.829	0.006, **
CYP2C8	paclitaxel-6-hydroxylation	NBIF-oxidation (M4)	0.577	0.104
CYP2C8	paclitaxel-6-hydroxylation	NBIF-oxidation (M5)	0.284	0.460
UGT1A1	β-estradiol-3-O-glucuronidation	NBIF-O-glucuronidation (G1)	0.822	0.007, **
UGT1A1	β-estradiol-3-O-glucuronidation	NBIF-O-glucuronidation (G2)	0.689	0.040, *
UGT1A9	propofol-O-glucuronidation	NBIF-O-glucuronidation (G1)	0.768	0.016, *
UGT1A9	propofol-O-glucuronidation	NBIF-O-glucuronidation (G2)	0.860	0.003, **

Note: NBIF, neobavaisoflavone. All experiments were performed in triplicate.

^*p < 0.05

^**p < 0.01

^***p < 0.001.

Notably, there was strong correlation between β-estradiol-3-O-glucuronide and G1 (r = 0.822, p = 0.007, Fig. S7a) and G2 (r = 0.689, p = 0.040, Fig. S7b) in the bank of individual HLMs (n = 9), respectively. The formation of G1 (r = 0.768, p = 0.016, Fig. S7c) and G2 (r = 0.860, p = 0.003, Fig. S7d) was also significantly correlated with propofol-O-glucuronidation. These results indicate that UGT1A1 and 1A9 catalyzed massive glucuronidation to form G1 and G2, respectively.

3.7. Species differences

As shown in Fig. 3C, M2, M3, and M4 were the main metabolites of NBIF produced by animal liver microsomes. The kinetic profiles of M2 and M4 in MkLM (Fig. S8a), M2 and M4 in RLM (Fig. S8c and d), M2 and M3 in MLM (Fig. S8e), M3 and M4 in DLM (Fig. S8f), and M3 and M4 in RaLM (Fig. S8g and h) followed the classical Michaelis-Menten equation, while M3 in MkLM (Fig. S8b), M3 in RLM (Fig. S8c), and M2 in RaLM (Fig. S8g) were modeled to the substrate inhibition equation. The catalytic efficiencies (reflected by CL_int values, Fig. 3C) for M2 by HLM and animal liver microsomes followed the order of RLM (70.28 μL/min/mg) > MkLM (29.99 μL/min/mg) > RaLM (14.49 μL/min/mg) > HLM (12.43 μL/min/mg) > MLM (10.23 μL/min/mg) (Table 2). Similarly, the CL_int values of M3 in RLM, RaLM, MkLM, MLM, HLM, and DLM were 50.41, 14.93, 13.98, 10.91, 10.04, and 6.73 μL/min/mg, respectively (Table 2). Similarly, the CL_int value order for M4 was MkLM (32.38 μL/min/mg) RaLM (22.83 μL/min/mg) > RLM (19.08 μL/min/mg) > DLM (7.25 μL/min/mg) > HLM (2.01 μL/min/mg) (Table 2).

As in the glucuronidation of NBIF in HLM, NBIF was rapidly metabolized into two glucuronides by animal liver microsomes (Fig. 3D). Additionally, the Michaelis-Menten equation fit best to the glucuronidation (G1 and G2) of NBIF by MkLM (Fig. S9a), DLM (Fig. S9d), and RaLM (Fig. S9e), whereas glucuronidation in RLM (Fig. S9b) and MLM (Fig. S9c) followed substrate inhibition kinetics. For G1, the CL_int value order was RaLM (623.26 μL/min/mg) > MLM (484.67 μL/min/mg) > RLM (388.88 μL/min/mg) > MkLM (362.51 μL/min/mg) > HLM (271.90 μL/min/mg) > DLM (194.09 μL/min/mg) (Table 3). Similarly, the CL_int values for G2 were 1257.92 μL/min/mg (MkLM), 965.27 μL/min/mg (DLM), 826.61 μL/min/mg (DLM), 651.38 μL/min/mg (HLM), 438.64 μL/min/mg (MLM), and 280.11 μL/min/mg (RLM) (Table 3).

Taken together, our results clearly revealed marked species differences in the phase I metabolism (approximately 16.1-fold) and glucuronidation (approximately 2.3-fold) of NBIF. Additionally, mice are the best model for NBIF-related oxidation and glucuronidation studies in humans because of their appropriate CL_int values.

4. Discussion

NBIF, as a major bioactive compound in P. corylifolia, has gained increasing attentions in recent years [4-10]. NBIF undergoes massive phase I and II metabolism in vivo and in the intestinal microflora [11,12], explaining why the pharmacokinetic behaviors of NBIF have been characterized as rapid oral absorption, high clearance, and poor absolute bioavailability [13-15]. We further examined the phase I and glucuronidation pathways of NBIF in the human liver and intestine.

Our results indicated that NBIF could be rapidly oxidized and glucuronidated in both HLM and HIM (Fig. 1). Notably, the oxidation rates (CL_int values less than 12.43 μL/min/mg) were clearly weaker than the glucuronidation rates (CL_int values between 59.96 and 949.01 μL/min/mg) of NBIF in HLM (Table 2) and HIM (Table 3). These results agree with those of our previous study showing that glucuronidation was the major metabolic pathway of corylin in HLM and HIM [21]. Additionally, CYP1A1, 1A2, and 2C19 clearly exhibited significant catalytic activities toward the A ring of prenylflavonoid, while CYP1A1, 1A2, 2C8, and 2C19 showed strong oxidation activities towards the isopentene group of prenylflavonoid, which agrees with previous results for corylin [21]. CYP1A1 also played the most important role in the dioxidation activities of the isopentene group of NBIF. The oxidation activities toward the isopentene group by CYP enzymes were the major oxidation reaction of NBIF. In addition, NBIF, also known as an individual prenylflavonoid, was efficiently glucuronidated, in addition to wushanicaritin (CL_int values >160 μL/min/mg) [23,24] and icaritin (CL_int values more than 270 μL/min/mg) [25]. Therefore, the first-pass glucuronidation reaction played a vital role compared to phase I metabolism (mainly oxidation) for NBIF clearance and determining bioavailability.

Because NBIF-containing herbal preparations are widely used in clinics, it is important to clarify the metabolic clearance of NBIF involving CYPs and UGTs in human. In the present study, CYP1A1, 1A2, 2C8, and 2C19 (Fig. S3) and UGT1A1, 1A7, 1A8, and 1A9 (Fig. S4) mainly participated in the oxidation and glucuronidation of NBIF, while CYP2C19 and UGT1A9 displayed the best catalytic activity for NBIF-related mono-oxidation (CL_int value = 63.42 μL/min/mg, Table 2) and glucuronidation (CL_int value = 1073.25 μL/min/mg, Table 3). CYP1A1, UGT1A7, and 1A8 were absent from the human liver and were mainly expressed in the intestine, stomach, and kidney [26,27], while CYP1A2, 2C8, and 2C19 and UGT1A1, 1A9 were mainly expressed in the human liver [26,27]. These findings showed that NBIF-related metabolism in the human liver and intestine for determining its oral bioavailability should not be underestimated, particularly the glucuronidation reactions.

Activity correlation analysis and reaction phenotyping experiments are also important for determining the contribution of individual CYP or UGT enzymes to respective phase I and II metabolism [21,23,25]. As shown in Table 4, phenacetin-N-deacetylation (Fig. S5), paclitaxel-6-hydroxylation (Fig. S6), β-estradiol-3-O-glucuronidation (Fig. S7), and propofol-O-glucuronidation (Fig. S5) were all significantly correlated with the formation of NBIF-related metabolites in a bank of individual HLMs (n = 9). The reason we did not determine the roles of CYP1A1, UGT1A7, and 1A8 in hepatic metabolism of NBIF was that they were not detected in the human liver or are not widely accepted probe substrates, although they exhibited high metabolic activities towards NBIF. These results (Table 4) support that CYP1A2 and 2C8 and UGT1A1 and 1A9 play key roles in the hepatic metabolism of NBIF.

Notably, because of the presence of bovine serum albumin, uncorrected parameters may lead to erroneous estimation of in vivo metabolic activities, namely poor in vivo-in vitro extrapolation, significantly weakening the ability to predict pharmacokinetics properties [28]. Hence, it is important to identify the protein binding factors affecting the outcome of in vitro metabolism assays. The Hallifax and Houston model, a widely recognized prediction approach, provides accurate predictions of free unbound (f_u) values based on the log P values of compounds with intermediate lipophilicity (log P values between 2.5 and 5.0) [29]. Therefore, binding between NBIF (log P = 3.73) and microsomal proteins may be negligible according to the Hallifax and Houston model, while the kinetic parameters (K_m, V_max, CL_int, etc.) did not require correction in this study.

Among the active individual enzymes (Fig. 3A and B), CYP1A2, 2C19, and UGT1A1 catalyzed approximately 9%, 12%, and 15% of marketed drugs, respectively [30], while these active isoforms also catalyzed the metabolism of endogenous agents [26]. For example, CYP1A1 is involved in the bioactivation of foreign agents [e.g., benzo(a)pyrene] to carcinogens and has been implicated in the formation of various types of human cancer [31]. Additionally, CYP1A1 and 1A2 metabolize polyunsaturated fatty acids into signaling molecules with physiological and pathological activities [32]. UGT1A1 also appeared to be of particular importance in maintaining human health because UGT1A1 is the sole physiologically relevant enzyme participating in the clearance of bilirubin, and plays a key role in the metabolism of many therapeutic drugs (e.g., irinotecan, morphine) [33]. The competition of these active CYP or UGT enzymes substrates [e.g., phenacetin, paclitaxel, (S)-mephenytoin, β-estradiol, bilirubin, propofol] may greatly decrease enzyme function and reduce the metabolism rate of NBIF in vivo. Common genetic polymorphisms among different ethnicities are important factors influencing the expression levels and activities of CYPs and UGTs. In clinics, before several therapeutic drugs (e.g., omeprazole, sertraline, clopidogrel, warfarin, voriconazole) are orally administered, the polymorphisms in CYP2C19 (CYP2C19*2, CYP2C19*3, CYP2C19*17) must be analyzed using gene detection approaches [34]. Similarly, the expression of UGT1A1 polymorphisms (UGT1A1*6, UGT1A1*28) should be detected to avoid or reduce the dose-limiting toxicities in irinotecan chemotherapy [35]. Therefore, humans with CYP or UGT enzyme dysfunction or genetic polymorphisms likely exhibit altered metabolism of NBIF, which alters its bioavailability in vivo.

5. Conclusion

This is the first study to characterize NBIF-related oxidation and glucuronidation pathways in human (Fig. 4). CYP and UGT isoforms involved in NBIF-oxidation and NBIF-O-glucuronidation, and their respective catalytic activities, were expressed as CL_int values by HLM, HIM, recombinant CYPs, and expressed UGTs (Tables 2 and 3). The results indicated that CYP1A1, 1A2, 2C8, 2C19, and UGT1A1, 1A7, 1A8, 1A9 were the main contributors to the phase I metabolism and glucuronidation of NBIF in HLM and HIM (Fig. 3A and B). Activity correlation analysis assays showed that the formation of several NBIF-related metabolites was significantly correlated with the probe substrates of active CYP or UGT enzyme-related metabolism in a bank of individual HLM (n = 9) (Table 4), suggesting that CYP1A2 and 2C8 and UGT1A1 and 1A9 play important roles in the hepatic metabolism of NBIF. Moreover, there were marked species differences in the oxidation and glucuronidation of NBIF between HLM and animal liver microsomes from monkeys, rats, mice, dogs, and rabbits (Fig. 3C and D), and mice may be the best animal model for evaluating the oxidation and glucuronidation of NBIF in humans. Taken altogether, NBIF underwent massive glucuronidation in both HLM and HIM compared to oxidation, indicating that CYP1A1, 1A2, 2C8, and 2C19 and UGT1A1, 1A7, 1A8, and 1A9 are the most active isoform enzymes. Thus, characterization of NBIF-related metabolic pathways involving the key CYP and UGT enzymes (Fig. 4) would reveal the pharmacokinetic behaviors and metabolic fates of NBIF in the human liver and intestine.