Discovery of Potent Glucokinase and PPARγ Dual-Target Agonists through an Innovative Scheme for Regioselective Modification of Silybin.

Glucokinase (GK) and PPARγ are important targets for antidiabetic use. Silybin is one of the major active ingredients of Silybum marianum. The regioselective modification of the five hydroxyl groups in the silybin structure has always been a challenge. In this study, we found that silybin was an agonist of GK and PPARγ. A novel synthesis scheme of silybin derivatives was designed, and a series of novel silybin derivatives has been synthesized. The derivative 8d showed relatively strong activation activity for GK and PPARγ in enzyme activity and transactivation assays (GK activation fold: 1.86; PPARγ transactivation activation percentage: 90.32%). This research suggests that silybin and its derivatives could be used as novel GK and PPARγ dual-target agonists.

Introduction

Type 2 diabetes mellitus is a complex disease characterized by hyperglycemia, insulin resistance, and insulin secretion deficiency.[1] The World Health Organization predicted that the number of diabetic patients was expected to reach 360 million worldwide by 2030.[2] However, no single marketed oral hypoglycemic drugs can control blood glucose continuously without any side effects. Modulating multiple targets in the biological network simultaneously is recognized as being beneficial for treating a range of diseases, including cancer and diabetes.[3] At present, with the development of molecular biology, pathophysiology, and pharmacology, it has become feasible to develop multitarget therapies for diabetes to improve efficacy and to minimize side effects,[4,5] and recently several studies suggested that the glucokinase (GK) and peroxisome proliferator-activated receptor-γ (PPARγ) dual-target agonists could be a potent way in treating patients with diabetes.[6−8]

GK is mainly expressed in the pancreatic β-cells and the liver.[9] It constitutes a rate-limiting step in glucose metabolism in these tissues.[10] In addition, the GK level was found reduced in patients with type 2 diabetes.[11] Although the GK activators such as RO281675,[12] AZD165,[13] GKA22, and GKA50[14] were reported to decrease the blood glucose, they failed to improve insulin resistance.

PPARγ is a member of the nuclear receptor family, broadly expressed in skeletal muscle, adipose tissue, and liver.[15] It is a crucial regulator of glucose homeostasis and adipocyte differentiation.[16] As the target of rosiglitazone and other thiazolidinedione drugs, it promotes insulin secretion and improves insulin resistance.[17] However, the PPARγ agonists such as thiazolidinedione have serious side effects: rosiglitazone may cause cardiac failure, pioglitazone causes edema events, and troglitazone was banned because of its hepatotoxicity.[18−20] Thus, the clinical application has been greatly limited.

Silybin, a major bioactive component of silymarin, extracted from blessed Silybum marianum,[21] boasts various biological activities, including anti-inflammatory, antioxidant, and antitumor effects.[22−24] Silybin has been proved to protect pancreatic β-cells,[25] attenuate hepatic glucose production,[26] and increase glucose uptake in muscle in vitro and in vivo.[27] However, the underlying mechanism remains unclear. Until now, silybin derivatives for antidiabetic use have remained to be investigated.

Eva-Maria et al.[28] reported that silybin is a PPARγ agonist, and the results of target prediction showed that GK and PPARγ might be potential targets of silybin (GK docking score: −8.9 kcal/mol; PPARγ docking score: −9.9 kcal/mol). This result motivated our interest in exploring GK and PPARγ dual-target agonists from silybin derivatives. The binding pockets of PPARγ and GK are composed of hydrophobic amino acids, so the hydrophobic compounds may have a better affinity to the pockets.[29,30] During the initial process of molecular docking, as shown in Figure , around the hydroxy group at position C-7 of silybin, there were hydrophobic pockets (circled by light yellow) in PPARγ and GK, so we suggested that hydrophobic structural modification at this position could improve the affinity between the compounds and the pocket. Hidalgo-Figueroa et al.[31] reported that agonists of PPARγ should contain a central aromatic scaffold and an extra-lipophilic side chain. This is understandable because the fatty acid is an endogenous agonist of PPARγ.[15] On the other hand, the structures of GK agonists RO281675, AZD1656, GKA22, and GKA50 also contain a central aromatic scaffold (benzene ring) and a lipophilic side chain. Therefore, the above findings inspired us to design the dual-target agonist for GK and PPARγ by modifying the structure of silybin (Figure ).

Figure 1

Binding mode of silybin to PPARγ and GK.

Figure 2

Silybin structure modification scheme.

In this work, a series of C-7-OH hydrophobic derivatives of silybin has been designed, including hydroxyl and different acyl halides to form esters and halohydrocarbons to form ethers. Moreover, the other four hydroxyl groups are acetylated to enhance the fat solubility of silybin. The target compounds were determined by virtual screening.

The regioselective modification of the five hydroxyl groups in the silybin structure has always been a challenge.[32] We observed that the reaction of acetylated silybin with aliphatic amine can selectively aminolyse the ester group at position C-7. This finding provided a feasible scheme for synthetic target compounds. Then these target compounds were synthesized and evaluated as dual GK/PPARγ agonists.

The Glide software was used to simulate the molecular docking process. Discovery Studio 2019 software was used to simulate the surface of the protein with blue as hydrophilicity and brown as hydrophobicity. The hydrophobic pocket was circled with light yellow.

Results and Discussion

Chemistry

In the research, we discovered that the ester groups of 3,5,7,20,23-penta-O-acetyl-silybin (2) contained acylation ability, amine could react with these groups to generate amide and turn its ester into hydroxyl, and the reaction mechanism is shown in Figure . This process was not hydrolysis of ester under the condition of a base, because this reaction can be performed under anhydrous conditions; both the primary and secondary amines are able to complete this reaction while tertiary and quaternary amines failed. The acidity at position C-7-OH of silybin is strongest, so the acylation ability of the ester group at this position is also strongest.[33] Through controlling the amount of amine and the reaction time, the ester group at position C-7 of compound 2 is able to react completely with other ester groups unaffected. This discovery, never being reported before, has important implications for regioselective modification of silybin, flavonoids, and polyhydroxy compounds.

Figure 3

Mechanism of the ammonolysis of compound 2.

All target compounds were synthesized with silybin as the original material. The reactions were monitored by thin-layer chromatography (TLC), and the spots were visualized with UV light. The preparation route of target compounds 4a–m and 5a and b is outlined in Scheme . We used acetic anhydride to synthesize acetylated silybin (1) under DMAP catalysis rapidly, and the yield was close to 100%. For the synthesis of compound 3, we tried a variety of solvents (including methanol, THF, and acetone) and used TLC to monitor the reaction. Finally, it was determined that compound A was completely converted using the amine (2 equiv) in pyridine at 0 °C for 30 s. Extension of reaction time results in the appearance of byproducts. Compounds 4a–m were synthesized in THF using corresponding acid chloride and triethylamine. Compounds 5a and b were prepared by the reaction of compound 3 with NaHCO3 under acetone. In Scheme 2,3-dehydrosilybin (DHS 7) was generated in an ethanol solution of sodium hydroxide. The silybin and DHS derivatives 6a–c and 8a–d were synthesized in THF using corresponding acid chlorides.

Scheme 1

Preparation Route of Target Compounds 4a-m and 5a-b

Reagents and Conditions: (a) acetic anhydride, DMAP, pyridine, r.t.; (b) CH3CH2CH2NH2, pyridine, 0 °C; (c) acid chloride, triethylamine, THF, r.t.; acetone, halohydrocarbons, NaHCO3, reflux, 24 h

Scheme 2

Preparation Route of Target Compounds 6a-c and 8a-d

Reagents and conditions: (a) NaOH, EtOH, reflux; (b) acid chloride, triethylamine, THF, r.t.

In Vitro Biological Activity Evaluation

Through virtual screening, we chose those compounds with better docking scores for GK and PPARγ, shown in areas E and F in Figure (GK and PPARγ docking score < −8 kcal/mol). Compounds that are difficult to synthesize are removed. With the help of the virtual screening results, a series of novel silybin derivatives were first synthesized (Scheme ). In vitro biological assay of PPARγ (Figure A) elucidated that the PPARγ transactivation activation of compounds 4c, 4d, 4e, 4i, 4m, and 5a was higher than that of silybin. Compounds with aliphatic side chain substitution have higher PPARγ transactivation activation than aromatic substitution, showing that the binding pocket of PPARγ is more affinitive to the flexible lipophilic group. Among them, the compounds 4c, 4e, 4i, and 4m substituted with propionyl, acryloyl, isobutyryl, and acetyl, respectively, showed the best activities.

Figure 4

Virtual screening results of GK and PPARγ.

Figure 5

(A) In vitro PPARγ transactivation assay of the compounds 4a–m and 5a and b. Each value is average (±SEM) of triplicate samples conducted at 10 μM (rosiglitazone activity was set as 100%). ***p < 0.001, **p < 0.01, and *p < 0.05, compared with silybin (A). (B) In vitro GK activity assay of the compounds 4a–m and 5a and b. Each value is average (±SD) of triplicate samples conducted at 10 μM (GK activation by DMSO only was set as 100%) ***p < 0.001, **p < 0.01, and *p < 0.05, compared with DMSO. (C) In vitro PPARγ transactivation assay of the compounds. Each value is average (±SEM) of triplicate samples conducted at 10 μM (rosiglitazone activity was set as 100%). ***p < 0.001, **p < 0.01, and *p < 0.05. (D) In vitro GK activity assay of the compounds. Each value is average (±SD) of triplicate samples conducted at 10 μM (GK activation by DMSO only was set as 100%). ***p < 0.001, **p < 0.01, and *p < 0.05.

Next, in vitro GK activity assay was applied to all compounds. As shown in Figure B, compounds 4c, 4e, 4h, 4i, 4m, and 5a with less steric hindrance led to moderate efficiency for GK activation. When the substituted R was replaced with bulky hydrophobic groups (tert-butyl group or cyclopropyl) like compounds 4f and 4g, they exhibited slightly decreased efficiency for GK activation, while compounds 4a, 4b, 4k, 4l, and 5b with aromatic substitution exhibited low or no GK activation. This discovery is consistent with the Li team,[34] suggesting that the binding pocket of GK has a strict steric size limit. The binding pocket of GK, similar to PPARγ, rejects groups with large steric hindrances and prefers flexible lipophilic groups with less steric hindrance.

Silybin showed potent activity for GK and PPARγ, which was in accordance with the results of target prediction; this may be one of the antidiabetic mechanisms of silybin. Through the research studies above, we found that when the substituted R was replaced with propionyl, acryloyl, isobutyryl, and acetyl, these compounds acquired good activities on both GK and PPARγ (GK activation fold >1.35, PPARγ transcription activation percentage >50%). Thus, we synthesized compounds 6a–c and 8a–d (Scheme ) to further verify whether these groups were more effective for GK, and PPARγ, silybin, and 2,3-dehydrosilybin (DHS 7) were used as the central aromatic scaffold.

As shown in Figure C,D, we found that propionyl, acryloyl, and acetyl led to better efficiency for GK and PPARγ, except for compound 6a, whose activity on GK was slightly reduced (compared to compound 4a). Isobutyryl (compound 6c) would harm the affinity of the silybin scaffold to both GK and PPARγ. However, when the scaffold was changed into DHS, the affinity to GK was not affected and even increased to PPARγ, suggesting that the GK binding pocket is stricter than PPARγ, and it cannot tolerate the silybin scaffold with more than three carbon side chains. Therefore, the DHS is a more effective scaffold because it could tolerate all groups without losing activity for GK and PPARγ. This result indicates that the aromatic scaffold with a larger conjugated system is essential to improve the activity of the compound for GK and PPARγ. Within the scope of the experiment, compound 8d showed the best GK and PPARγ activity, which means that acetyl may be the most effective lipophilic side chain.

The abscissa represents the docking score of the designed compounds to PPARγ. The ordinate represents the docking score of the designed compounds to GK. The quadrants are marked with A–R.

Molecular Docking Studies

The binding mode of compound 8d to the pocket was different from what we simulated in Discovery Studio 2019 software using silybin as the ligand. The changed binding pose of 8d may result from the addition of multiple side chains and the change of the aromatic scaffold. As shown in Figure , compound 8d was chosen for molecular docking to investigate its interactions with GK and PPARγ. In the binding pocket of GK, the acetyl group at position C-7 of compound 8d formed the hydrogen bond with Gln 219. Moreover, the 1,4-dioxane in the aromatic scaffold formed a hydrogen bonding network with Gly 97 and Gln 98 and formed the Pi–Pi stacked with Trp 99. On the other hand, compound 8d perfectly matched the lipophilic binding pocket of PPARγ and interacted with Ser 289, Tyr 327, Ser 342 (hydrogen bonds), and Phe 264 (Pi–Pi stacked).

Figure 6

Possible binding mode of compound 8d in the corresponding targets. (A) Compound 8d with GK LBD (PDB ID 4ISE). (B) Compound 8d with PPARγ LBD (PDB ID 6DGL). Key amino acid side chains are shown in gray stick format and are labeled. Hydrogen bonds are shown in yellow dotted lines and Pi–Pi stacks are shown in blue dotted lines.

In Vitro Cytotoxicity Evaluation

We also used normal human hepatocytes (HL7702 cell) to perform cytotoxicity experiments on all compounds. Silybin is a marketed drug for chronic hepatitis with good safety.[35] Therefore, it was selected as the positive control drug. As shown in Table , all of the compounds exhibited good safety in vitro, except compound 4j. This increased toxicity of compound 4j may be due to the generation of carbocations by halogen ionization. When the chlorine of compound 4j was replaced with hydrogen, compound 4m showed safety equivalents to silybin. The toxicity of compounds 4b and 4k is reduced because of the inability of halogen to ionize (conjugation of halogen to the benzene ring), Figure .

Table 1

In Vitro Cytotoxicity (IC50, μM) of Compounds to Normal Hepatocytes

compound	IC50 (μM)a	compound	IC50 (μM)a	compound	IC50 (μM)a
Silybin 1	92.19 ± 11.59	4h	84.98 ± 6.69	6b	83.43 ± 722
3	80.77 ± 9.39	4i	87.55 ± 4.81	6c	87.56 ± 6.43
4a	>100	4j	52.27 ± 5.34**	DHS 7	77.54 ± 4.45
4b	>100	4k	>100	8a	80.77 ± 3.28
4c	82.09 ± 3.51	4l	>100	8b	75.32 ± 15.86
4d	>100	4m	>100	8c	>100
4e	77.98 ± 5.82	5a	84.72 ± 8.55	8d	93.91 ± 5.74
4f	>100	5b	78.56 ± 13.36
4g	64.21 ± 3.18	6a	65.92 ± 3.42

The IC50 value is the compound concentration that inhibited HL7702 cell growth by 50%. The data are mean of three measurements ±SD. ***p < 0.001, **p < 0.01, and *p < 0.05, compared with silybin.

Figure 7

Cytotoxic structure–activity relationship of compounds. The structural differences are presented, and the rest are omitted.

Conclusions

Herein, a new method with efficiency and high yield for regioselective modification of silybin was proposed. We discovered acylation ability of the ester groups in compound 2 and the acylation ability of ester groups was different because of the different chemical environment; with this mechanism, regioselective ammonolysis of 3,5,7,20,23-penta-O-acetyl-silybin (2) can be achieved. This discovery provided a new idea for regioselective modification of silybin, flavonoids, and polyhydroxy compounds.

In this work, we proved that silybin was a natural agonist for GK and PPARγ, and it exhibited good agonistic effects on GK and PPARγ (GK activation fold = 1.32, PPARγ transcription activation percentage = 37.25%). In addition, a series of silybin derivatives with safety as potent dual-acting agonists of GK and PPARγ was designed, screened, and synthesized. Silybin and the DHS as the aromatic scaffold, and adding flexible hydrophobic side chains led to discovering four compounds with high enzyme activating and transcriptional stimulating activities (6b, 8a, 8b, and 8d, GK activation fold >1.5, PPARγ transcription activation percentage >60%), and the preliminary structure–activity relationships were determined and summarized. The discovery of these compounds may be used as lead compounds for GK and PPARγ dual-target agonists.

Experimental Section

Virtual Screening

All designed hydrophobic derivatives of silybin were screened by molecular docking using the Glide program from Schrodinger (maestro version 11.5.011, MMshare version 4.1.011, released 2018-1). The protein structures of GK and PPARγ (PDB ID: 4ISE and 6DGL) were downloaded from http://www.rcsb.org/pdb. The Protein Preparation Wizard module was applied for protein structure preparation, such as adding hydrogen, filling in missing side chains, removing water molecules in the binding pocket, and so on. The Receptor Gride Generation module in Glide was used to generate a receptor grid file that determined the position of the binding pockets using the unique ligand in the downloaded protein structure to define the binding site and simulate the binding mode of candidate molecular ligands to GK and PPARγ. Finally, screening was carried out using the Virtual Screening Workflow module in Glide software with extra precision mode, and the ligands were treated as flexible. All the docked compounds were scored by the Glide scoring function, and the compounds with higher docking scores for GK and PPARγ were selected as target compounds.

General

The raw material of silybin was kindly provided by Liaoning Wode Pharmaceutical Co., Ltd. The raw material of silybin under normal conditions was an isometric mixture of two diastereoisomers: Silybin A and Silybin B. Individual pure isomers were difficult to separate (only prep-HPLC can separate the isoforms), and readymade isomers were too expensive to achieve commercialized production, so the research studies worldwide about silybin mostly used the mixture.[36,37] Therefore, after much thought, the mixture was adopted as a raw material to proceed the reaction and biology assay. All reagents and solvents were purchased from commercial suppliers and used as they were provided. Melting points were determined on a BUCHI B-540 melting point apparatus and are uncorrected. The reaction was monitored by TLC on a Shanghai TLC silica gel 60 GF254 with UV detection. 1H NMR spectra were recorded on a Bruker Fourier 300 spectrometer (Billerica, MA, USA), and the solvent is DMSO-d6, using trimethylsilane as an internal standard. The chemical shifts were expressed in values (ppm). The coupling constant (J) is given in hertz (Hz). The abbreviations used are as follows: s, singlet; d, doublet; and m, multiplet. ESI-MS was measured on a Thermo-Finnigan LCQ equipment from Thermo-Finnigan (San Francisco, CA, USA).

3,5,7,20,23-Penta-O-acetyl-silybin (1)

Silybin (0.964 g, 2 mmol) was dissolved in pyridine (10 mL). The mixture was stirred at room temperature, whereupon acetic anhydride (2 mL, 35 mmol) was added followed by DMAP (5 mg). The reaction mixture was stirred at room temperature for 5 min, and then 100 mL cold water was poured into the mixture. The precipitate was filtered out, washed with water, and then dried to obtain compound 2. The yielded pure product was in the form of powder with white color (1.365 g, 99%). m.p. 111–117 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.28–7.24 (m, 2H), 7.17–7.03 (m, 4H), 6.89 (d, J = 3.0 Hz, 1H), 6.78 (d, J = 2.2 Hz, 1H), 5.98 (t, J = 12.3 Hz, 1H), 5.68 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.67–4.59 (m, 1H), 4.14 (dt, J = 12.2, 2.7 Hz, 1H), 3.97 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.30 (s, 3H), 2.28 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.97 (s, 3H); ESI-MS m/z [M + H]+: 715.17.

3,5,20,23-Tetra-O-acetyl-silybin (3)

To a stirred solution of compound 2 (1.384 g, 2 mmol) in pyridine (10 mL) was added propylamine (330 μL, 4 mmol, 2.0 equiv), and the reaction mixture was stirred for 30 s at 0 °C. The reaction was quenched by adding glacial acetic acid (230 μL, 4 mmol) immediately. The reaction solution mixture was poured into 200 mL of ice water. The precipitate was filtered out, washed with water, and then dried to obtain compound 3. White solid, yield 1.27 g (98%). m.p. 132–135. 1H NMR (600 MHz, DMSO-d6): δ 11.12 (s, 1H), 7.35–6.89 (m, 6H), 6.31 (d, J = 2.3 Hz, 1H), 6.26 (d, J = 2.2 Hz, 1H), 5.76 (t, J = 12.1 Hz, 1H), 5.52 (d, J = 12.1 Hz, 1H), 5.11 (t, 1H), 4.66–4.59 (m, 1H), 4.14 (d, J = 11.6 Hz, 1H), 3.98 (dd, J = 12.4, 5.1 Hz, 1H), 3.79 (s, 3H), 2.27 (s, 3H), 2.26 (s, 3H), 2.01 (s, 3H), 1.96 (s, 3H); ESI-MS m/z [M + Na]+: 673.17.

General Procedure for the Synthesis of (4a–m)

To a solution of compound B (650 mg, 1 mmol) in THF (10 mL), corresponding acid chloride (1.1 mmol, 1.1 equiv) followed by triethylamine (280 μL, 2 equiv) was added. The reaction mixture was stirred at room temperature for 1 h. The precipitate (triethylamine hydrochloride) was removed by filtration under reduced pressure. The solvent (THF) was evaporated under reduced pressure. The residue is recrystallized in methanol.

7-O-Benzoyl-3,5,20,23-tetra-O-acetyl-silybin (4a)

White solid, yield 556 mg (77.8%), m.p. 130–134 °C. 1H NMR (600 MHz, DMSO-d6): δ 8.12 (d, J = 6.8 Hz, 2H), 7.77 (t, J = 7.5, 1.2 Hz, 1H), 7.62 (t, J = 7.4 Hz, 2H), 7.28 (dd, J = 9.0, 2.0 Hz, 2H), 7.17–6.97 (m, 6H), 6.02 (t, 1H), 5.72 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.66–4.61 (m, 1H), 4.14 (dt, J = 12.2, 2.6 Hz, 1H), 3.97 (dd, J = 12.4, 5.0 Hz, 1H), 3.79 (s, 3H), 2.31 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.99 (s, 3H); ESI-MS m/z [M + H]+: 715.22.

7-O-P-Fluorobenzoyl-3,5,20,23-tetra-O-acetyl-silybin (4b)

White solid, yield 622 mg (80.5%), m.p. 132–136 °C. 1H NMR (600 MHz, DMSO-d6): δ 8.20 (dd, J = 8.7, 5.6 Hz, 2H), 7.45 (t, J = 8.4 Hz, 2H), 7.28 (dd, J = 8.1, 1.8 Hz, 2H), 7.18–7.03 (m, 5H), 6.98 (d, J = 2.1 Hz, 1H), 6.02 (t, J = 12.0 Hz, 1H), 5.72 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.68–4.60 (m, 1H), 4.14 (d, J = 12.0 Hz, 1H), 3.97 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.31 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.99 (s, 3H); ESI-MS m/z [M + H]+: 772.26.

7-O-Propionyl-3,5,20,23-tetra-O-acetyl-silybin (4c)

White solid, yield 497 mg (70.4%), m.p. 117–121 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.29–6.75 (m, 8H), 5.98 (t, J = 12.3 Hz, 1H), 5.69 (d, J = 12.4 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.67–4.59 (m, 1H), 4.14 (dt, J = 12.4, 2.8 Hz, 1H), 3.97 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.61 (q, J = 7.4 Hz, 2H), 2.30 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.97 (s, 3H), 1.12 (t, J = 7.0 Hz, 3H); ESI-MS m/z [M + Na]+: 729.21.

7-O-3,3-Dimethacryloyl-3,5,20,23-tetra-O-acetyl-silybin (4d)

White solid, yield 601 mg (82.1%), m.p. 101–106 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.31–6.74 (m, 8H), 5.97 (t, 1H), 5.95 (s, 1H), 5.68 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 8.1 Hz, 1H), 4.68–4.59 (m, 1H), 4.14 (dt, J = 12.2, 2.8 Hz, 1H), 3.97 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.30 (s, 3H), 2.27 (s, 3H), 2.16 (s, 3H), 2.01 (s, 3H), 1.98 (s, 3H), 1.98 (s, 3H); ESI-MS m/z [M + H]+: 733.98.

7-O-Acrylic-3,5,20,23-tetra-O-acetyl-silybin (4e)

White solid, yield 602 mg (85.6%), m.p. 121–126 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.29–6.83 (m, 8H), 6.56 (d, J = 17.3 Hz, 1H), 6.40 (ddd, J = 17.2, 10.5, 1.0 Hz, 1H), 6.20 (d, J = 10.4 Hz, 1H), 6.00 (t, J = 11.9 Hz, 1H), 5.70 (d, J = 12.4 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.67–4.59 (m, 1H), 4.14 (d, J = 11.9 Hz, 1H), 3.97 (dd, J = 12.4, 5.0 Hz, 1H), 3.79 (s, 3H), 2.30 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.98 (s, 3H); ESI-MS m/z [M + Na]+: 727.21.

7-O-Pivaloyl-3,5,20,23-tetra-O-acetyl-silybin (4f)

Yellow solid, yield 563 mg (76.4%), m.p. 115–119 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.32–6.73 (m, 8H), 5.97 (t, J = 12.1 Hz, 1H), 5.69 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 8.3 Hz, 1H), 4.69–4.56 (m, 1H), 4.14 (d, J = 12.4 Hz, 1H), 3.97 (dd, J = 12.4, 5.0 Hz, 1H), 3.79 (s, 3H), 2.29 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.98 (s, 3H), 1.29 (s, 9H); ESI-MS m/z [M + H]+: 737.42.

7-O-Cyclopropionyl-3,5,20,23-tetra-O-acetyl-silybin (4g)

White solid, yield 591 mg (82.4%), m.p. 121–126 °C.1H NMR (600 MHz, DMSO-d6): δ 7.28–6.75 (m, 8H), 5.97 (t, 1H), 5.68 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.68–4.59 (m, 1H), 4.14 (d, J = 12.0 Hz, 1H), 3.97 (dd, J = 12.4, 5.0 Hz, 1H), 3.79 (s, 3H), 2.29 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.97 (s, 3H), 1.93–1.86 (m, 1H), 1.11–1.06 (m, 2H), 1.06–1.01 (m, 2H); ESI-MS m/z [M + Na]+: 741.19.

7-O-N-Butyryl-3,5,20,23-tetra-O-acetyl-silybin (4h)

White solid, yield 615 mg (85.9%), m.p. 122–126 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.34–6.74 (m, 8H), 5.98 (t, J = 12.2 Hz, 1H), 5.69 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 8.1 Hz, 1H), 4.68–4.57 (m, 1H), 4.14 (d, J = 12.2 Hz, 1H), 3.97 (dd, J = 12.4, 5.0 Hz, 1H), 3.79 (s, 3H), 2.57 (t, J = 7.2 Hz, 2H), 2.30 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.98 (s, 3H), 1.64 (h, J = 7.2 Hz, 2H), 0.95 (t, J = 7.4 Hz, 3H); ESI-MS m/z [M + H]+: 716.17.

7-O-Isobutyryl-3,5,20,23-tetra-O-acetyl-silybin (4i)

Yellow solid, yield 587 mg (82.1%), m.p. 99–103 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.39–6.58 (m, 8H), 5.98 (t, J = 12.2 Hz, 1H), 5.69 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 8.1 Hz, 1H), 4.68–4.57 (m, 1H), 4.14 (d, J = 12.4 Hz, 1H), 3.97 (dd, J = 12.4, 5.0 Hz, 1H), 3.79 (s, 3H), 2.82 (hept, J = 7.0 Hz, 1H), 2.30 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.98 (s, 3H), 1.22 (s, 3H), 1.21 (s, 3H); ESI-MS m/z [M + H]+: 716.21.

7-O-Chloroacetyl-3,5,20,23-tetra-O-acetyl-silybin (4j)

Yellow solid, yield 484 mg (66.8%), m.p. 120–125 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.35–6.79 (m, 8H), 6.00 (t, J = 12.5 Hz, 1H), 5.71 (d, J = 12.4 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.70 (s, 2H), 4.67–4.59 (m, 1H), 4.14 (d, J = 12.2 Hz, 1H), 3.97 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.31 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.98 (s, 3H); ESI-MS m/z [M + Na]+: 749.08.

7-O-P-Chlorobenzoyl-3,5,20,23-tetra-O-acetyl-silybin (4k)

White solid, yield 602 mg (77.6%), m.p. 133–139 °C. 1H NMR (600 MHz, DMSO-d6): δ 8.12 (d, J = 8.6 Hz, 2H), 7.69 (d, J = 7.4 Hz, 2H), 7.34–6.94 (m, 8H), 6.02 (t, J = 12.0 Hz, 1H), 5.72 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 7.8 Hz, 1H), 4.69–4.55 (m, 1H), 4.14 (d, J = 12.0 Hz, 1H), 3.97 (dd, J = 12.4, 5.0 Hz, 1H), 3.79 (s, 3H), 2.31 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.99 (s, 3H); ESI-MS m/z [M + H]+: 777.15.

7-O-Nicotinyl-3,5,20,23-tetra-O-acetyl-silybin (4l)

White solid, yield 608 mg (80.5%), m.p. 135–138 °C. 1H NMR (600 MHz, DMSO-d6): δ 9.25 (d, J = 2.1 Hz, 1H), 8.91 (d, J = 4.4 Hz, 1H), 8.46 (d, J = 8.0 Hz, 1H), 7.66 (dd, J = 8.0, 4.9 Hz, 1H), 7.36–6.93 (m, 8H), 6.02 (t, J = 12.2 Hz, 1H), 5.73 (d, J = 12.4 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.68–4.58 (m, 1H), 4.14 (d, J = 9.7 Hz, 1H), 3.97 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.32 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.99 (s, 3H); ESI-MS m/z [M + H]+: 756.20.

7-O-Acetyl-3,5,20,23-tetra-O-acetyl-silybin (4m)

Yellow solid, yield 561 mg (81.2%), m.p. 111–117 °C.1H NMR (600 MHz, DMSO-d6): δ 7.28–7.24 (m, 2H), 7.17–7.03 (m, 4H), 6.89 (d, J = 3.0 Hz, 1H), 6.78 (d, J = 2.2 Hz, 1H), 5.98 (t, J = 12.3 Hz, 1H), 5.68 (d, J = 12.3 Hz, 1H), 5.12 (t, J = 7.9 Hz, 1H), 4.67–4.59 (m, 1H), 4.14 (d, J = 12.2, 2.7 Hz, 1H), 3.97 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.30 (s, 3H), 2.28 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.97 (s, 3H); ESI-MS m/z [M + Na]+: 715.17.

General Procedure for the Synthesis of (5a–b)

To a solution of compound B (650 mg, 1 mmol) in acetone (10 mL), NaHCO3 (170 mg, 2 equiv) was added followed by iodomethane/benzyl bromide (3 mmol, 3 equiv) portion wise. The reaction mixture was refluxed for 24 h. The excess acid binding agent (NaHCO3) was removed by filtration under reduced pressure. The solvent (acetone) was evaporated under reduced pressure. The residue is recrystallized in methanol.

7-O-Methyl-3,5,20,23-tetra-O-acetyl-silybin (5a)

Yellow solid, yield 523 mg (78.9%), m.p. 103–107 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.32–6.97 (m, 6H), 6.59 (d, J = 2.5 Hz, 1H), 6.50 (d, J = 2.5 Hz, 1H), 5.84 (t, 1H), 5.58 (d, J = 12.1 Hz, 1H), 5.15–5.08 (m, 1H), 4.67–4.59 (m, 1H), 4.14 (d, J = 12.3 Hz, 1H), 3.98 (dd, J = 12.4, 5.1 Hz, 1H), 3.83 (s, 3H), 3.79 (s, 3H), 2.27 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.97 (s, 3H); ESI-MS m/z [M + Na]+: 687.17.

7-O-Benzyl-3,5,20,23-tetra-O-acetyl-silybin (5b)

Yellow solid, yield 514 mg (71.9%), m.p. 115–119 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.44 (d, J = 6.8 Hz, 2H), 7.40 (t, J = 7.4 Hz, 2H), 7.35 (d, J = 7.1 Hz, 1H), 7.26 (s, 1H), 7.24 (s, 1H), 7.15 (dd, J = 8.1, 3.0 Hz, 1H), 7.09–7.02 (m, 3H), 6.66 (d, J = 2.4 Hz, 1H), 6.57 (d, J = 2.4 Hz, 1H), 5.85 (t, 1H), 5.58 (d, J = 12.1 Hz, 1H), 5.21 (s, 2H), 5.12 (t, J = 8.0 Hz, 1H), 4.67–4.59 (m, 1H), 4.14 (d, J = 9.8 Hz, 1H), 3.98 (dd, J = 12.5, 5.0 Hz, 1H), 3.79 (s, 3H), 2.27 (s, 3H), 2.27 (s, 3H), 2.01 (s, 3H), 1.97 (s, 3H); ESI-MS m/z [M + H]+: 716.13.

General Procedure for the Synthesis of (6a–c)

To a solution of silybin (241 mg, 0.5 mmol) in THF (40 mL), corresponding acid chloride (3 mmol) followed by triethylamine (700 μL) was added. The reaction mixture was stirred at room temperature for 1 h. The precipitate (triethylamine hydrochloride) was removed by filtration under reduced pressure. The solvent (THF) was evaporated under reduced pressure. The residue is recrystallized in methanol.

3,5,7,20,23-Penta-O-propionyl-silybin (6a)

White solid, yield 301 mg (76.5%), m.p. 123–128 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.39–6.57 (m, 8H), 5.96 (t, 1H), 5.68 (d, J = 12.3 Hz, 1H), 5.11 (dd, J = 11.3, 7.7 Hz, 1H), 4.66–4.58 (m, 1H), 4.16 (dd, J = 12.4, 3.1 Hz, 1H), 4.00 (dd, J = 12.4, 4.9 Hz, 1H), 3.78 (s, 3H), 2.67–2.56 (m, 6H), 2.39–2.15 (m, 4H), 1.17–1.09 (m, 9H), 1.00 (t, J = 7.5 Hz, 3H), 0.91 (td, J = 7.5, 1.4 Hz, 3H); ESI-MS m/z [M + Na]+: 811.19.

3,5,7,20,23-Penta-O-acryloyl-silybin (6b)

Yellow solid, yield 181 mg (48.9%), m.p. 119–123 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.33–6.92 (m, 8H), 6.57–5.96 (m, 15H), 5.97 (t, 1H), 5.77 (d, J = 12.4 Hz, 1H), 5.19 (t, 1H), 4.74–4.66 (m, 1H), 4.27 (d, J = 12.6 Hz, 1H), 4.05 (dd, J = 17.4, 3.4 Hz, 1H), 3.77 (s, 3H); ESI-MS m/z [M+ Na]+: 763.21.

3,5,7,20,23-Penta-O-isobutyryl-silybin (6c)

Yellow solid, yield 293 mg (70.5%), m.p. 114–119 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.27–6.75 (m, 8H), 5.93 (dd, J = 12.6, 3.6 Hz, 1H), 5.69 (d, J = 12.3 Hz, 1H), 5.11 (dd, J = 14.3, 7.4 Hz, 1H), 4.68–4.53 (m, 1H), 4.19 (d, J = 11.4 Hz, 1H), 3.99 (dd, J = 12.4, 4.9 Hz, 1H), 3.78 (s, 3H), 2.89–2.65 (m, 5H), 1.27 (s, 3H), 1.26 (s, 3H), 1.24 (s, 3H), 1.22 (d, J = 6.5 Hz, 3H), 1.14 (d, J = 7.0 Hz, 3H), 1.09 (s, 3H), 1.06 (s, 3H), 1.05 (s, 3H), 0.99 (d, J = 5.0 Hz, 3H), 0.87 (d, J = 3.1 Hz, 3H); ESI-MS m/z [M+ Na]+: 855.9.

2,3-Dehydrosilybin (DHS 7)

Silybin (4.82 g, 10 mmol) and NaOH (80 mg, 2 mmol) were dissolved in 100 mL ethanol, and the mixture was heated under reflux for 4–5 h. The mixture was then left to cool to room temperature, and water containing HCl was poured into the mixture to adjust it to neutral. The precipitate formed was filtered off, washed with ethanol and water, and then dried. The residue is recrystallized in methanol. The yielded pure product was in the form of powder with light yellow color (4.27 g, 89%). m.p. 271–275 °C. 1H NMR (600 MHz, DMSO-d6): δ 12.42 (s, 1H), 10.81 (s, 1H), 9.56 (s, 1H), 9.16 (s, 1H), 7.77 (t, J = 2.1 Hz, 1H), 7.75 (d, J = 2.2 Hz, 0H), 7.12 (d, J = 8.5 Hz, 1H), 7.05 (d, J = 2.0 Hz, 1H), 6.89 (dd, J = 8.2, 2.0 Hz, 1H), 6.82 (d, J = 8.0 Hz, 1H), 6.46 (d, J = 2.0 Hz, 1H), 6.20 (d, J = 2.0 Hz, 1H), 4.99 (t, J = 5.6 Hz, 1H), 4.97 (d, J = 7.9 Hz, 1H), 4.28 (ddd, J = 7.5, 4.6, 2.5 Hz, 1H), 3.80 (s, 3H), 3.58 (ddd, J = 12.3, 5.0, 2.5 Hz, 1H), 3.40–3.35 (m, 1H); ESI-MS m/z [M + H]+: 481.12.

General Procedure for the Synthesis of (8a–d)

To a solution of DHS (240 mg, 0.5 mmol) in THF (40 mL), corresponding acid chloride (3 mmol) followed by triethylamine (700 μL) was added. The reaction mixture was stirred at room temperature for 1 h. The precipitate (triethylamine hydrochloride) was removed by filtration under reduced pressure. The solvent (THF) was evaporated under reduced pressure. The residue is recrystallized in methanol.

3,5,7,20,23-Penta-O-propionyl-2,3-dehydrosilybin (8a)

White solid, yield 330 mg (86.5%), m.p. 156–162 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.67–7.03 (m, 8H), 5.19 (d, J = 7.7 Hz, 1H), 4.80–4.68 (m, 1H), 4.21 (dd, J = 12.6, 3.1 Hz, 1H), 4.03 (dd, J = 12.6, 4.9 Hz, 1H), 3.79 (s, 3H), 2.70–2.57 (m, 8H), 2.38–2.25 (m, 2H), 1.18–1.10 (m, 12H), 1.00 (t, J = 7.5 Hz, 3H); ESI-MS m/z [M + Na]+: 785.26.

3,5,7,20,23-Penta-O-acryloyl-2,3-dehydrosilybin (8b)

Yellow solid, yield 301 mg (81.6%), m.p. 161–164 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.37–6.86 (m, 8H), 6.57–5.74 (m, 15H), 5.19 (dd, J = 13.1, 7.7 Hz, 1H), 4.74–4.66 (m, 1H), 4.27 (d, J = 4.8 Hz, 1H), 4.05 (dd, 1H), 3.77 (s, 3H); ESI-MS m/z [M + Na]+: 761.22.

3,5,7,20,23-Penta-O-isobutyryl-2,3-dehydrosilybin (8c)

Yellow solid, yield 330 mg (79.6%), m.p. 159–163 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.66 (d, J = 2.3 Hz, 1H), 7.59 (d, J = 2.2 Hz, 1H), 7.53–7.46 (m, 1H), 7.28 (d, J = 1.8 Hz, 1H), 7.21 (d, J = 8.5 Hz, 1H), 7.18–7.13 (m, 2H), 7.08 (d, J = 1.9 Hz, 1H), 5.20 (d, J = 7.7 Hz, 1H), 4.78–4.70 (m, 1H), 4.23 (dd, J = 12.6, 3.1 Hz, 1H), 4.02 (dd, J = 12.5, 4.7 Hz, 1H), 3.78 (s, 3H), 2.92–2.79 (m, 5H), 1.29 (s, 3H), 1.27 (s, 3H), 1.27 (s, 3H), 1.26 (s, 3H), 1.24 (s, 3H), 1.23 (s, 3H), 1.14 (d, J = 7.0 Hz, 3H), 1.10 (d, J = 7.0 Hz, 3H), 1.07 (d, J = 3.8 Hz, 3H), 1.06 (d, J = 3.8 Hz, 3H); ESI-MS m/z [M + Na]+: 853.39.

3,5,7,20,23-Penta-O-acetyl-2,3-dehydrosilybin (8d)

Yellow solid, yield 237 mg (68.9%), m.p. 144–148 °C. 1H NMR (600 MHz, DMSO-d6): δ 7.65 (d, J = 2.2 Hz, 1H), 7.60 (d, J = 2.2 Hz, 1H), 7.51 (dd, J = 8.6, 2.2 Hz, 1H), 7.31 (d, J = 1.9 Hz, 1H), 7.22 (d, J = 8.6 Hz, 1H), 7.18 (d, J = 8.0 Hz, 1H), 7.14 (d, J = 2.2 Hz, 1H), 7.10 (dd, J = 8.2, 1.9 Hz, 1H), 5.20 (d, J = 7.8 Hz, 1H), 4.75 (ddd, J = 7.9, 4.9, 3.0 Hz, 1H), 4.19 (dd, J = 12.5, 3.0 Hz, 1H), 4.02 (dd, J = 12.6, 5.0 Hz, 1H), 3.81 (s, 3H), 2.33 (s, 6H), 2.33 (s, 3H), 2.28 (s, 3H), 2.03 (s, 3H); ESI-MS m/z [M + Na]+: 713.16.

Biological Studies

PPARγ Transcription Activity Assays

Human embryonic kidney (HEK) 293 T cells were cultured in Dulbecco’s Modified Eagle’s Medium (Gibco-Life Technologies) containing 25 mmol L–1 of glucose and 10% heat-inactivated fetal bovine serum in 5% CO2 incubator at 37 °C. Then, 293T cells were seeded in six-well plates for 24 h to give a confluence of 70–80% at transfection and transferred to a 96-well plate (containing 60,000 cells/well). Cells were transfected with 6.6 μL PPARγ, 2.5 μL of PPRE-luciferase, 1.0 μL of Renilla, and 20 μL of lipofectamine for 5 h. After transfection, 293T cells were treated with target compounds (10 μM) or positive control rosiglitazone and pioglitazone for 24 h and then collected with cell culture lysis buffer. Luciferase activity was monitored on a luminometer (PerkinElmer, USA) using the luciferase assay kit (Promega) according to the manufacturer’s instructions. Rosiglitazone activity was set as 100%.

GK Activity Assays

The GK activities of the synthesized compounds were measured using the glucose-6-phosphate dehydrate (G6PDH) enzyme coupling method. The assay was performed in a final volume of 200 μL containing Tris–HCl (25 mM, pH 7.4), glucose (10 mM), potassium chloride (25 mM), magnesium chloride (1 mM), dithiothreitol (1 mM), ATP (1 mM), NAD (1 mM), G-6-PDH (2.5 U/mL), GK (0.5 μg), and compounds under investigation (10 μM). Absorbance was measured at 340 nm after 3 min incubation period and GK activation fold by the target compounds, and GK fold activation was calculated compared with control (GK activation by DMSO only was considered as 100%).

In Vitro Cytotoxicity Experiments

To evaluate the cytotoxicity, we used the HL7702 cell line as the normal human hepatic cells. The cytotoxicity of the compounds was evaluated in vitro using the MTT method.[38] The experiments were carried out in 96-well culture dishes (6000 cells/well), with silybin as the positive control, and the negative control contained cells, culture medium, MTT, and DMSO. HL7702 cells were cultured in supplemented with 10% FBS and 1% penicillin–streptomycin. All cells were cultured in a 37 °C incubator filled with 5% CO2 gas for 24 h. Then, cells were exposed to different concentrations of compounds (0.1, 1.0, 10, 100, and 200 μM) for 48 h. Then MTT solution (100 μL per well) was added and incubated at 37 °C for 4 h. The MTT formazan formed by metabolically viable cells was dissolved in 150 μL DMSO each well. The absorbance of each well was measured at a wavelength of 450 nm using a Robonik P2000 EIA reader. The cell viability was expressed as a percentage of control, and the concentration that induces 50% of maximum inhibition of cell proliferation (IC50) was determined using the Karber method. Each experiment was repeated at least three times, and the results were averaged.

Molecular Docking Study

The molecular docking study of synthesized compounds into the binding pocket of GK and PPARγ used the grid file in the virtual screening process. It was carried out to explore the possible binding patterns.

Statistical Analysis

Data analysis and statistical graphs were completed using Graph-Pad Prism 7.0 (San Diego, California, USA). All experimental data are shown as mean ± SEM or mean ± SD. One-way analysis of variance was performed to compare significant differences between multiple groups. p < 0.05 indicates statistical significance.