2.1 Chemicals and reagents

The compounds including toosendanin (42), cucurbitacin B (43), hemslecin A (44, HSA), cucurbitacin Ⅱb (45), pachymic acid (46), cucurbitacin E (47, CBE) were purchased from Herbest Biotechnology Co., Ltd (Baoji, China). (24E)-coccinic acid (23), kadpolysperin I (24), schisanlactone B (25), kadcoccinone N (26), schisandronic acid (27), heteroclitalactone B (28), heteroclitalactone E (29), were kindly provided by Prof. Wei Wang (Hunan University of Chinese Medicine, Wang et al., 2006). The other compounds were isolated and identified in our laboratory (Mai et al., 2015; Wei et al., 2017; Zhang et al., 2017; Wei et al., 2018; Sun et al., 2020; Luan et al., 2019). NADPNa₂, glucose-6-phosphate dehydrogenase, glucose-6-phosphate, and midazolam (MDZ) were obtained from Sigma-Aldrich (St. Louis, MO). The human liver microsomes were obtained from Bioreclamation IVT (MD, USA) and stored at −80 °C until use.

2.2 Screening of the inhibitory effects of 47 tetracyclic triterpenoids against CYP3A4

The inhibitory effects of 47 TT against CYP3A4 mediated midazolam 1′-hydroxylation were evaluated. The NADPH-generating incubation system (1 mM NADPNa₂, 4 mM MgCl₂, 1 unit/mL glucose-6-phosphate dehydrogenase, 10 mM glucose-6-phosphate), HLM (0.05 mg/mL), and substrate MDZ incubated in 100 mM potassium phosphate buffer (pH 7.4) in the presence (inhibition samples) or absence (control samples) of various TT compounds (10 µM). The incubation samples were initiated by adding NADPNa₂ after 3 min preincubation at 37 °C. Continuous incubation at 37 °C for 7.5 min, adding 100 μL acetonitrile to stop the enzymatic reaction. The incubation samples were centrifugated at 20,000 g for 20 min, and the supernatants were analyzed by LC-MS. The residual activity was calculated by the relative value of the formation rate of 1′-hydroxylated metabolite of MDZ of inhibition samples and control samples.

2.3 Evaluation of the inhibitory effects of 47 tetracyclic triterpenoids toward various CYP subtypes

The inhibition of 47 TT toward other mainly drug metabolizing CYP subtypes distributed in HLM, including CYP1A2, 2B6, 2C8, 2C9, 2C19 and 2D6, was evaluated using respective probe substrate. The functionality of each CYP isoforms was assessed by determining the catalytic activities, that is, phenacetin O-deethylation for CYP1A2, bupropion 4-hydroxylation for CYP2B6, taxol 6-hydroxylation for CYP2C8, diclofenac 4′-hydroxylation for CYP2C9, S-Mephenytoin 4′-hydroxylation for CYP2C19, dextromethorphan O-demethylation for CYP2D6. The incubation conditions were consistent with the previous studies (He et al., 2015; Li et al., 2018).

2.4 IC₅₀ determination

Varying concentrations of 47 TT (25 nM-150 µM) were incubated with MDZ (5 µM) for 7.5 min. The inhibitory dose–response curves were obtained to fit the data with the program Prism (Version 8.0, GraphPad, San Diego, CA). Other conditions were consistent with the described above.

2.5 Inhibition kinetic characterization

Varying concentrations of 23-acetyl alisol C (AAC), HSA, CBE (0.1–25 µM) were incubated for 7.5 min with varying concentrations of MDZ (1–25 µM). The Lineweaver–Burk and Dixon plots were obtained to fit the data with the program Prism (Version 8.0, GraphPad, San Diego, CA). The inhibition constant (K_i) values were determined using the following equations:

When α is very large (a greater than 1), the mixed-model becomes identical to competitive inhibition.

2.6 LC–MS assay

A standard ExionLC AD HPLC system was equipped with a Luna Omega PS C18 (2.1 × 100 mm, 3 μm) analytical column. The mobile phase consisted of 0.1% formic acid aqueous solution (A) and acetonitrile (B), were conducted with the 0.30 mL/min flow rate as follows: 0–1.5 min (10–10% B), 1.5–3.0 min (10–45% B), and 3.0–5.0 min (45–90% B). An Applied Biosystems SCIEX QTRAP 5500 with electrospray ionization (ESI) source was used to analyze target metabolites. The analyses were performed in positive mode, and used the following values: 1′-hydroxylated MDZ (342.0 → 324.0), deethylated phenacetin (152.1 → 110.0), 4-hydroxylated bupropion (256.0 → 238.0), 6-hydroxylated paclitaxel (892.4 → 607.2), 4-hydroxylated diclofenac (312.1 → 231.0), 4′-hydroxylated S-mephenytoin (235.0 → 150.0), demethylated dextromethorphan (258.1 → 157.0), defluorinated gefitinib (445.0 → 128.0), 2- and 4-hydroxylated atorvastatin (575.0 → 440.0), S-oxidated quetiapine (400.1 → 177.9). Related parameters for the quantification assay setting refer to Table S1 in the supplementary materials.

2.7 3D-QSAR model studies

The comparative molecular field method (CoMFA) has been widely used in the discovery and optimization of lead compounds (Melo-Filho et al., 2014; Yang et al., 2011). In this study, the CoMFA method was used to construct a three-dimensional quantitative structure–activity relationship (3D-QSAR) model based on pIC₅₀.

During the modeling process, all molecules were prepared in SYBYL X1.1 by minimizing in the standard force field (Tripos), adding Gasteiger-Huckel charge and searching all molecules to produce the lowest energy conformation of small molecules. Then, the lowest energy conformation was used as the initial structure of the CoMFA model. In order to build the CoMFA model and verify its predictive ability, we divide all molecules into a training set and a test set. Considering the distribution of biological activity data and the diversity of chemical structures, 4, 6, 15, 24, 30, 37 and 44 were selected from 47 compounds as the test set, and the remaining 40 compounds as the training set. In order to modeling the biologically active conformations, compounds in the training set were superimposed based on a common scaffold, and 49 (HSA) with the largest pIC₅₀ value was set as template molecule of superimposition. During the construction of CoMFA, the steric field and electrostatic field are calculated. Then, Partial Least Squares (PLS) in SYBYL-X 1.1 was applied to perform CoMFA modeling. The pIC₅₀ value of the training set in the CoMFA model is set as Y, and the CoMFA parameter values as X. The optimal number of components (ONC) was obtained by leave-one-out (LOO) cross-validation analysis. After the model is built, internal and external validation are used to evaluate the robustness (R² and Q²) and predictive ability. The statistical parameters of the CoMFA model are listed in Table S2 in the supplementary materials.

2.8 Interactions between tetracyclic triterpenoid derivatives and clinical drugs

In the present study, quetiapine, atorvastatin and gefitinib that mainly undergoing CYP3A4 mediated metabolism were used as potential interacting clinical drugs with TT derivatives. The inhibitory effects of five representative TT derivatives (0.1–100 µM) toward the oxidation metabolism of the above-mentioned clinical drugs were evaluated. The concentration of clinical drugs in incubation system was seated according to the corresponding reported K_m values of probe substrates. The incubation conditions were consistent with the previous studies (Li et al., 2018).

3.1 Inhibitory effects of tetracyclic triterpenes derivatives toward CYP3A4

The inhibitory effects of 47 TT toward CYP3A4 were first evaluated (Scheme 1). It was found that, at the concentration of 10 μM, alismanin D (2), (23S)-11β,23-dihydroxy-8α,9β,14β-dammar-13(17)-ene-3,24-dione (3), 23-acetyl alisol C (AAC, 9), alisol B 23-acetate (10), alisol B (AB, 11), kadpolysperin I (24), kadcoccinone N (26), panaxatriol (PT, 32), (20S)-protopanaxdiol (S-PPT, 34), hemslecin A (HSA, 44) and cucurbitacin E (CBE, 47) exhibited a strong inhibitory effects toward CYP3A4 mediated MDZ 1′-hydroxylation, and the percentage remaining activity were all less than 25% (Fig. 1). The compound 47 (cucurbitacin E, CBE), a major component of Curcurbitaceae plants, was a potential inhibitor toward CYP3A4 with the percentage remaining activity determined as 5.3%. In contrast, the activity of CYP3A4 was slightly modulated by compounds including 7-oxo-ganoderlactone D (12), Ganolactone B (14), ganoderic acid H (15), banoderic acid B (16), 12β-hydroxyganoderenic F (17), ganoderic AM1 (18), ganoderic acid C (19), ganoderic acid F (20), ganoderic acid C6 (21), schisandronic acid (27), ginsenoside Rh1 (39), toosendanin (42) and pachymic acid (46) with the percentage remaining activity more than 75% at the same screening concentration. It can be found that ganoderic acid all exhibited weak inhibitory effects toward CYP3A4.

Close

Scheme 1

The structures of tetracyclic triterpenoids (1–47) are used.

Export to PPT

Close

Fig. 1

(A) A brief profile of inhibitory effects of tetracyclic triterpenoids (1–47) toward major human CYP isoforms in the human liver. (B) The representative tetracyclic triterpenoids exhibited strong inhibitory effects toward CYP3A4.

Export to PPT

The dose-dependent inhibition behaviors of representative TT toward CYP3A4 were evaluated (Fig. 2). The inhibitory activities of compounds including 9 (AAC), 11 (AB), 32 (PT), 33 (S-PPT), 44 (HSA), and 47 (CBE) expressed as IC₅₀ values and determined as 1.42, 1.27, 1.80, 4.94, 0.65 and 0.42 µM, respectively.

Close

Fig. 2

Dose-dependent inhibition behavior of tetracyclic triterpenoids toward CYP3A4.

Export to PPT

3.2 Inhibitory effects of tetracyclic triterpenoids toward other CYP subtypes

To explore the inhibitory selectivity of 47 TT, the inhibitory effects of these compounds toward other CYP subtypes that are mainly involved in drug metabolism were evaluated. Compound 47, which was the potential inhibitor of CYP3A4, exhibited very weak inhibitory effects toward other CYP subtypes at the concentration of 10 μM (Fig. 1). And the extensive strong inhibitory effects toward CYP3A4 of a majority of TT were not observed in other CYP subtypes. It was found that a few TT inhibited CYP2B6 with strong inhibition activity, including (24E)-coccinic acid (23), heteroclitalactone B (28), and panaxadiol (31). However, the TT all exhibited relatively weak inhibitory effects toward CYP2C8 and CYP2D6, with a percentage remaining activity were all more than 75%. Additionally, it was found that compound 23 strongly inhibited CYP1A2; compounds 30 and 42 strongly inhibited CYP2C9; and compound 36 strongly inhibited CYP2C19. Overall, TT derivatives exhibited a strong inhibitory effect toward CYP3A4 in comparison with other CYP subtypes.

3.3 3D-QSAR of tetracyclic triterpenoids by CoMFA

The strong inhibitory effects of TT prompted us to explore the interaction relationship between TT and CYP3A4. Above all, the IC₅₀ values were obtained for the 47 ligands to screen for their inhibitory ability toward CYP3A4. All the data were listed in Table S2, and it was found that the variation of the IC₅₀ values for these TT was very large.

Subsequently, the pIC₅₀ is set as the dependent variable for the construction of the CoMFA model. Considering the distribution of biological activity data and the diversity of chemical structures, compounds 4, 6, 15, 24, 30, 37, and 44 were selected from 47 compounds as the test set, and the remaining 40 compounds as the training set. The results of partial least square (PLS) analysis are listed in Table S3, and plots of the predicted values against experimental values are shown in Figure S1. The CoMFA model for the training set yields statistically significant results with the cross-validated q² of 0.522 and non-cross-validated r² of 0.936. The raw experimental and predicted pIC₅₀ values was shown in Table S2, and the correlation is shown in Fig. 3. All these parameters indicate that the CoMFA model is more reliable and the consistency of internal and external data sets is good.

Close

Fig. 3

The 3D quantitative structure–activity relationship of the tetracyclic triterpenoids based on the ligands. (A) The conformational overlap of 47 tetracyclic triterpenoids with compound 47 as a template molecule. (B) Equipotential diagram of CoMFA model based on template molecule 47. (C) Stereo field equipotential map of CoMFA model. The favorable area of the stereo field is shown in green, and the unfavorable area is shown in yellow. (D) The electrostatic field equipotential diagram of the CoMFA model. The positively charged area is shown in blue, and the negatively charged area is shown in red.

Export to PPT

The equipotential map generated by the PLS model can explain the relationship between the three-dimensional structures and inhibition activities of these compounds. Based on the conformational superposition of 47 TT inhibitors (Fig. 3A), the 3D-QSAR model of the steroid inhibitors was obtained by constructing a COMFA equipotential map (Fig. 3B). It can be seen that the inhibition activities of these TT were simultaneously affected by the stereo field and electrostatic fields, and the electrostatic contribution is nearly two times the contribution of the steric field. The steric field is mainly distributed on two sides of the TT scaffold (the A ring and D ring regions, see Fig. 3C), indicating that the modification of the substituents in these regions has a greater impact on the inhibitory activity of the TT inhibitors. The green contour indicates that the substitution of a larger group can increase the inhibition activity, while the yellow contour indicates that the substitution of a smaller group helps to increase the inhibition activity. As shown in Fig. 3D, the distribution of the electrostatic field is relatively dispersed. Among them, there is a favorable area of negative charge in the area near the A ring, indicating that the appropriate negative group substitution can enhance the activity of the inhibitor. The positively charged and negatively charged regions exist simultaneously around the D ring change, which indicates that increasing the positively charged and negatively charged groups in the corresponding regions can enhance the activity of the inhibitor.

3.4 Inhibition kinetics of representative tetracyclic triterpenes toward CYP3A4

The inhibition kinetics of CYP3A4 by CBE, AAC, and HSA were performed to explore the possible inhibition mechanism of TT. As shown in Fig. 4, the Dixon and Lineweaver–Burk plots of three TT indicated that CYP3A4 inhibition was all fitted in a mixed inhibition manner. Notably, the fitting parameters α were all greater than 1, indicating that the mixed model for CBE, AAC, and HSA was inclined to competitive inhibition (Table 1). The value of K_i for CBE, AAC, and HSA was determined as 0.162, 2.12, and 0.193 µM, respectively.

Close

Fig. 4

The Lineweaver-Burk and Dixon plots for CYP3A4 inhibition in the presence of CBE (A and B), AAC (C and D), and HSA (E and F), respectively.

Export to PPT

3.5 Interactions between representative tetracyclic triterpenes and clinical drugs

Some clinical drugs that mainly undergo CYP3A4-mediated metabolism including gefitinib, atorvastatin, and quetiapine were used to evaluate the potential interaction between TT and clinical drugs. The IC₅₀ values of CBE, AAC, HSA, S-PPT, and PT were determined and listed in Table 2. The involving TT exhibited significant inhibitory effects toward the CYP3A4-mediated oxidative metabolism of gefitinib, atorvastatin, and quetiapine. Therein, CBE strongly inhibited the metabolism of gefitinib, atorvastatin, and quetiapine, with the IC₅₀ lying between 0.94 and 3.75 µM.

CBE and HSA are major components of curcurbitaceae plants including Hemsleya amabilis Diels, Trichosanthes kirilowii Maxim, and Cucurbita moschata Duch that displayed anti-inflammatory, antipyretic, and antibacterial activity (Kaushik et al., 2015). Recently, this type of TT has been proven to have a strong anticancer activity the antiproliferative effect of CBE on PC 3, HL-60, and Bel-7402 lay between 10 and 43 nM, respectively (Chen et al., 2012). Cucurbitacin tablets and other drugs that contain these active ingredients have served as adjutant drugs in tumor therapy (Wang et al., 2017).

In the present study, the potential interaction between CBE and gefitinib was observed, indicating that more attention should be paid to the monitoring of gefitinib blood concentration and side effects when using CBE-containing drugs as adjutant agents. Additionally, S-PPT and PT are important metabolites of ginsenosides that are enriched in commonly used herbs such as Panax notoginseng (Burkill) F. H. Chen ex C. H. and Panax ginseng C. A. Meyer (Liu et al., 2006). The high frequency in the daily life of these herbal medicines may induce herbal-drug interaction due to the strong inhibitory effects of metabolites of ginsenosides toward CYP3A4. Overall, the results indicated that the interactions between TT and clinical drugs may be widespread, and the potential drug interaction induced by TT and they are containing herbal medicine should be taken attention.