The Synthesis and Antitumor Activity of Twelve Galloyl Glucosides

Twelve galloyl glucosides 1–12, showing diverse substitution patterns with two or three galloyl groups, were synthesized using commercially available, low-cost d-glucose and gallic acid as starting materials. Among them, three compounds, methyl 3,6-di-O-galloyl-α-d-glucopyranoside (9), ethyl 2,3-di-O-galloyl-α-d-glucopyranoside (11) and ethyl 2,3-di-O-galloyl-β-d-glucopyranoside (12), are new compounds and other six, 1,6-di-O-galloyl-β-d-glucopyranose (1), 1,4,6-tri-O-galloyl-β-d-glucopyranose (2), 1,2-di-O-galloyl-β-d-glucopyranose (3), 1,3-di-O-galloyl-β-d-glucopyranose (4), 1,2,3-tri-O-galloyl-α-d-glucopyranose (6) and methyl 3,4,6-tri-O-galloyl-α-d-glucopyranoside (10), were synthesized for the first time in the present study. In in vitro MTT assay, 1–12 inhibited human cancer K562, HL-60 and HeLa cells with inhibition rates ranging from 64.2% to 92.9% at 100 μg/mL, and their IC50 values were determined to be varied in 17.2–124.7 μM on the tested three human cancer cell lines. In addition, compounds 1–12 inhibited murine sarcoma S180 cells with inhibition rates ranging from 38.7% to 52.8% at 100 μg/mL in the in vitro MTT assay, and in vivo antitumor activity of 1 and 2 was also detected in murine sarcoma S180 tumor-bearing Kunming mice using taxol as positive control.


The in Vitro Antitumor Activity of 1-12
The in vitro antitumor activity of 1-12 was assayed by the MTT method on human cancer K562, HL-60 and HeLa cell lines, using 5-flurouracil (5-FU) and docetaxol (DOC) as positive controls at first. In the MTT assay, 1-12, 5-FU and DOC inhibited the tested three human cancer cell lines by the inhibition rates (IR%) at 100 μg/mL shown in Table 1, and the half inhibitory concentration (IC50) of 1-12 was determined as given in Table 2. We then further tested the inhibitory effect of 1-12 on the murine sarcoma S180 cell line by the same MTT assay using taxol as positive control. In the test, 1-12 and taxol inhibited the S180 cells with following IR% values at 100 μg/mL: 1, 40

The in Vivo Antitumor Activity of 1 and 2 in Mice
The in vivo antitumor activity of 1 and 2 was tested on the murine sarcoma S180 tumor-bearing Kunming mice, and taxol was used as positive control. Compounds 1-2 at 15 and 30 mg/kg dosages and taxol at 20 mg/kg dosage were administered by intravenous injection via the mouse tail vein every day for six continuous days for 1-2 and every other day for six days for taxol, respectively.
On the third day of the last administration, the mice were sacrificed, and the body and tumor weights were weighed. Then, the inhibition rate (IR%) of the test and taxol groups on the tumor growth was calculated using mean tumor-weight (TW) values by the following formula: In the present test, both 1 and 2 at the 30 mg/kg dosage could significantly inhibit the tumor growth, although their inhibitory effect was a little weaker than the effect of the positive control taxol (Table 3), while 1 and 2 at 15 mg/kg dosage did not show significant inhibitory effects on the tumor growth as shown in Table 3. Moreover, in contrast to the significantly reduced body weight of the mice in taxol group, 1 and 2 both at 15 and 30 mg/kg dosages did not significantly affect the body weight of tested mice (Table 3). Table 3. Inhibitory effect of 1 and 2 on the S180 tumor growth in mice (mean ± S.D., n = 8 or 11). Notes: * p < 0.01, ** p < 0.05, compared with model group; n = 8 for taxol group and n = 11 for other groups.
It is noteworthy that in the 4,6-O-benzylidende ring forming reaction we used triethyl orthoformate for the first time for the pTSA-catalyzed reaction of D-glucose and benzaldehyde, as shown in Scheme 2, to obtain the 4,6-O-benzylidende protected ethyl glycoside 17 by a one-pot reaction. Under the acidic reaction conditions, the triethyl orthoformate underwent hydrolysis to produce EtOH, which in turn further reacted with the reactive 1-OH in glucose, eventually affording 17. In contrast, the 1-OH free, 4,6-O-protected product 16 could be obtained by the same pTSA-catalyzed reaction of the D-glucose and benzaldehyde without triethyl orthoformate (Scheme 2). Generally, triethyl orthoformate was used for the purpose of dehydration in relevant reactions by its hydrolysis to produce ethyl formate or formic acid and EtOH. Triethyl orthoformate has also been used in the pTSA-catalyzed reaction of a β-D-glucoside and benzaldehyde for protecting the 4,6-OH groups in glucose and the target 4,6-Obenzylidende-protected product with an unchanged β-D-glycoside bond was obtained in [50]. Similarly, the pTSA-catalyzed reaction of methyl α-D-glucoside and benzaldehyde in the presence of triethyl orthoformate in the present study exclusively produced the unchanged α-glycoside bond 4,6-Obenzylidende protected product 14 (Scheme 2), indicating the stability of the methyl α-glycoside bond under the acidic reaction conditions. Under the acidic reaction conditions, triethyl orthoformate should undergo hydrolysis (by H2O from the reaction of methyl α-D-glucoside and benzaldehyde) to produce formic acid and EtOH. If the methyl α-D-glucoside were also hydrolyzed, the reactive 1-OH in the D-glucose produced should further react with EtOH to also produce an ethyl 4,6-O-benzylidene-Dglucoside as byproduct, as seen in the synthesis of 17.
The hydrolysis-resistant nature of the methyl α-D-glycoside bond was also indicated by the acid-catalyzed hydrolysis of 15. In the hydrolysis of 15 by 2 N HCl at 80 °C for 16 h, shown in Scheme 3, in addition to the hydrolyzed product 18, quite a large amount of the 1-OCH3 unhydrolyzed material methyl 2,3-di-O-benzyl-α-D-glucopyranoside was produced in every round of hydrolysis, and the 41.2% total yield of 18 was obtained by repeated hydrolyses of the 1-OCH3 unhydrolyzed materials. Incidentally, later by an additional hydrolysis of 15 using 30% trifluoroacetic acid (TFA) as catalyst in CH3CN-TFA-H2O (3:3:4) at 98 °C for 48 h according to the method in the literature [51], we could obtain 18 in quite a high 40% yield through a one-step hydrolysis.
In the in vitro MTT assay, compounds 1-12 inhibited the human cancer K562, HL-60 and HeLa cells with inhibition rates ranging from 64.2% to 92.9% at 100 μg/mL (Table 1), and their IC50 values were determined to vary from 17.2-124.7 μM on the tested cancer cell lines (Table 2). In the in vitro MTT assay, 1-12 also inhibited the murine sarcoma S180 cells with inhibition rates ranging from 38.7% to 52.8% at the 100 μg/mL (see Section 2.2.1). The MTT assay for the synthesized 1 and 2 on K562 cells (IC50: 77.9 μM for 1 and 68.2 μM for 2) reconfirmed our assay previously reported result for the natural products (IC50: 80.8 μM for natural 1 and 64.8 μM for natural 2) [44]. In addition, it seems from the IC50 values in Table 2 that the galloyl glucosides with three galloyl groups (2, 5, 6 and 10) possess stronger inhibitory activity than those with two galloyl groups (1, 3, 4, 7-9, 11 and 12) on the tested three human cancer cell lines, except for 6 on the K562 cell line, and further on the whole, 1-12 all showed stronger activity on the HL-60 cells than on the K562 and HeLa cells (Table 2). Moreover, two pairs of the α,β-anomers, 5/6 and 11/12, showed no difference in their inhibition of tested human cancer cell lines between the corresponding α,β-anomers (Table 2), and no relationship between inhibitory activity and substitution patterns in 1-12 could be found from the IC50 values given in Table 2. We then carried out antitumor tests in mice for 1 and 2, which were obtained in a larger amount for the test, in vivo antitumor activity of 1 and 2 was detected on the murine sarcoma S180 tumor-bearing Kunming mice using taxol as positive control as shown in Table 3. The same S180 tumor-bearing mice model were also used in [52,53] for an in vivo antitumor test by a similar procedure [52,53], including the use of taxol as positive control [52]. Although the in vivo antitumor activity of 1 and 2 in the present test was certainly weaker than that of taxol, it is noteworthy that 1 and 2 showed also very weak toxicity to the host, as indicated by body weight changes of the tested mice (Table 3), as documented for PGG in [18,20]. In view of the present results, it seems worthy to further evaluate the in vivo antitumor activity of these galloly glucosides in detail, including whether they would exert preventive effect on the tumor development in mice by pre-administration, and so on.

General
Melting points were measured on a Beijing Tiandiyu X-4 exact micro melting point apparatus (Tiandiyu Science and Technology Co., Ltd., Beijing, China) and the temperature was not corrected. Optical rotations were measured on an Optical Activity Limited polAAr 3005 spectropolarimeter (Optical Activity Limited, Ramsey, UK). ESIMS was recorded on an Applied Biosystems API 3000 LC-MS spectrometer (AB SCIEX, Framingham, MA, USA). 1   A mixed solution of 18 (3.0 g, 0.0083 mol) and 13 (10.3 g, 0.0225 mol) in 30 mL of anhydrous pyridine was stirred at 60 °C for 48 h and then evaporated under reduced pressure to obtain a reaction mixture. The reaction mixture was separated by silica gel column chromatography (bed, 3.5 × 15.0 cm; eluted by Pet.Et2O-acetone 20:1) to afford crude 19 (8.0 g, 0.0066 mol, 79.5% yield), which did not give pseudo-molecular ion peaks in both positive and negative ESI-MS measurements. Crude 19 (8.0 g, 0.0066 mol) was dissolved in 200 mL of THF-95% EtOH (15:5) solution, transferred into a 0.5 L high-pressure reactor, and then reduced by H2 under 10 atmospheric pressure at 40 °C for 12 h using 10% Pd-C (0.8 g) as catalyst. The reaction mixture was filtered to filter out insoluble materials and evaporated under reduced pressure to give a reaction product. This product was separated by repeated Sephadex LH-20 column chromatography eluted with H2O-MeOH (75:25→25:75) to afford crude 1. The crude 1 was recrystallized in 30% MeOH to obtain pure 1 (1. Notes: a Chemical shifts were recorded in δ C values using the solvent signal (CD 3 OD: δ C 49.00) as reference.
Signal assignments were based on the results of 1 H-1 H COSY and HMQC experiments.

1,4,6-Tri-O-galloyl-β-D-glucopyranose (2)
A mixed solution of 18 (4.5 g, 0.0125 mol) and 13 (42.0 g, 0.0917 mol) in anhydrous pyridine (170 mL) was stirred at 60 °C for 48 h and then evaporated under reduced pressure to obtain a reaction mixture. The mixture was subjected to silica gel column chromatography eluted with CH2Cl2-EtOAc (100:1) to afford crude 20 (15.5 g, 0.0095 mol, 76.0% yield), which did not give the corresponding pseudo-molecular ion peaks in positive and negative ESI-MS measurements. The crude 20 (15.5 g, 0.0095 mol) was dissolved in THF-95% EtOH (17:6, 230 mL), transferred into a 1 L high-pressure reactor, and reduced by H2 under 10 atmospheric pressure at 40 °C for 12 h using 10% Pd-C (1.0 g) as catalyst. The reaction mixture was filtered to filter out insoluble materials and evaporated to give a reaction product. This product was separated by repeated Sephadex LH-20 column chromatography eluted with H2O-EtOH (45:55) to afford crude 2, which was recrystallized in 40% MeOH to give pure 2 (2.3 g, 0.0036 mol, yield 37.9%). Needles (MeOH-H2O) Tables 4 and 5. (6) A mixed solution of 16 (2.0 g, 0.0075 mol) and 13 (8.5 g, 0.0186 mol) in anhydrous pyridine (40 mL) was stirred at 60 °C for 48 h and then evaporated under reduced pressure to obtain a reaction mixture that was separated by silica gel column chromatography (bed, 3.0 × 13.0 cm) using CH2Cl2 as eluting solvent to afford a mixture of three esterified products 21 (5.8 g). The mixture was dissolved in THF-95% EtOH (14:1, 150 mL), transferred into a 0.5 L high-pressure reactor, and then reduced by H2 under 10 atmospheric pressure at 40 °C for 12 h using 10% Pd-C (0.7 g) as catalyst. The reaction mixture was filtered to filter out the undissolved materials and evaporated under reduced pressure to obtain a reaction product. This reaction product was separated by Sephadex LH-20 column (bed, 1.8 × 60.  Tables 4 and 5.

Methyl 2,3-Di-O-galloyl-α-D-glucopyranoside (7)
A mixed solution of 14 (1.5 g, 0.0053 mol) and 13 (10.0 g, 0.0218 mol) in anhydrous pyridine (50 mL) was stirred at 60 °C for 48 h and evaporated under reduced pressure to give a reaction mixture. The reaction mixture was subjected to column chromatography on silica gel (bed: 5.0 × 10.0 cm), eluted by CH2Cl2, to afford 22 (5.0 g, 0.0044 mol, yield 83.0%). A portion of 22 (3.5 g, 0.0031 mol) was dissolved in 150 mL of THF-95% EtOH (13:2), transferred into a 0.5 L high-pressure reactor, and then reduced by H2 under 10 atmospheric pressure at 40 °C for 12 h using 10% Pd-C (1.5 g) as catalyst. The reaction mixture was filtered to filter out insoluble substances and evaporated to give a reaction product containing 23 ( A mixed solution of 24 (1.9 g, 0.0051 mol) and 13 (8.0 g, 0.0175 mol) in anhydrous pyridine (30 mL) was stirred at 60 °C for 48 h and then evaporated under reduced pressure to obtain a reaction mixture that was separated by silica gel column chromatography (bed: 3.0 × 15.0 cm, eluted with a gradient of Pet. Et2O-CH2Cl2 2:1→CH2Cl2) to afford a crude 25 (5.5 g, 0.0045 mol, 88.2% yield). The crude 25 (5.5 g, 0.0045 mol) was dissolved without further purification in THF-95% EtOH (13:2, 150 mL), transferred into a 0.5 L high-pressure reactor, and then reduced by H2 under 10 atmospheric pressure at 40 °C for 12 h using 10% Pd-C (2.0 g) as catalyst. The reaction mixture was filtered to filter out insoluble substances and evaporated under reduced pressure to give a reaction product. This product was separated by Sephadex LH-20 column (bed, 2.8 × 60.0 cm; eluted by 40% EtOH) to give  Tables 4 and 5.

Ethyl 2,3-Di-O-galloyl-α-D-glucopyranoside (11) and Ethyl 2,3-di-O-galloyl-β-D-glucopyranoside (12)
A mixed solution of 17 (2.0 g, 0.0068 mol) and 13 (7.2 g, 0.0157 mol) in anhydrous pyridine (40 mL) was stirred at 60 °C for 48 h and then evaporated under reduced pressure to give a reaction mixture that was separated by silica gel column chromatography (bed, 3.8 × 13.0 cm; eluted by a gradient of Pet. Et2O-CH2Cl2 2:1→CH2Cl2-MeOH 30:1) to afford crude 30 (a 1α,1β mixture, 7.3 g, 0.0064 mol, 94.1% yield). The crude 30 without further purification was dissolved in THF-95% EtOH (14:1, 150 mL), transferred into a 0.5 L high-pressure reactor, and reduced by H2 under 10 atmospheric pressure at 40 °C for 12 h using 10% Pd-C (1.5 g) as catalyst. The reaction mixture was filtered to filter out the insoluble substances and evaporated under reduced pressure to give a reduction product which was then separated by repeated Sephadex LH-20 column chromatography (bed, 2.8 × 60.  Tables 4 and 5. (13) To a solution of anhydrous K2CO3 (200 g) in DMSO (500 mL) was added dried gallic acid (60 g, 0.3529 mol) under stirring and heated to 140 °C. Then, benzyl chloride (300 mL) was added dropwise over 1.5 h under nitrogen atmosphere and then reacted at 140 °C for 8 h. After cooling down the reaction solution to the room temperature, the reaction mixture was diluted with cooled-distilled water (2000 mL) and extracted with CH2Cl2 (2000 mL × 4). The CH2Cl2 solution was combined and evaporated under reduced pressure to give a solid product. This product was suspended in MeOH (1800 mL), harvested by filtration, and washed with a suitable amount of methanol in the same filter to give a white powder. The powder was dried in vacuo and dissolved in CH2Cl2 (1000 mL). To the solution, MeOH (approximately 1000 mL) was added in portions, and when a lot of fine crystal species appeared, held at room temperature for two hours. Then, the crystals formed in the solution were filtered and then dried in vacuo to obtain benzyl tri-O-benzylgallate ( A mixture of methyl-α-D-glucopyranoside (20 g, 0.10 mol), freshly distilled benzaldehyde (60 mL), triethyl orthoformate (20 mL), p-toluenesulfonic acid monohydrate (1.0 g) and THF (200 mL) was refluxed at 85 °C for 16 h. After cooling the reaction mixture to room temperature, K2CO3 (1.0 g) was added and the mixture was stirred at room temperature for 30 min. The mixture was filtered, and the filtrate was suspended in distilled water (400 mL) and a suitable amount of 95% EtOH was added to dissolve the suspended materials in full. Then, the solution was evaporated under reduced pressure, until a fine crystalline species appeared, and kept at 4 °C for 12 h. The crystals formed in the solution were filtered, washed with water (100 mL ×2) and cyclohexane (100 mL ×3), and dried in vacuo to give pure methyl 4,6-O-benzylidene-α-D-glucopyranoside (14, 20 g, 0.071 mol, 71.0% yield). Needles

Methyl 2,3-di-O-Benzyl-4,6-O-benzylidene-α-D-glucopyranoside (15)
A mixed solution of 14 (17.5 g, 0.062 mol) and NaH (5 g) in DMF (300 mL) was stirred for 30 min, during which it became hot and was thus cooled down to room temperature. To the mixture benzyl bromide (45 mL) was added dropwise with stirring and reacted at room temperature for 4 h. To the reaction mixture was added distilled water (200 mL), and then it was extracted with CHCl3 (500 mL). The CHCl3 solution was washed with water (200 mL × 3), then dried with anhydrous MgSO4, filtered, and evaporated under reduced pressure, at a lower temperature at first to remove CHCl3 and further at 90 °C to remove the remaining DMF, to obtain a yellow-colored product. This product was separated by silica gel column chromatography (bed, 2.0 × 50.0 cm), eluted with Pet. Et2O-EtOAc (10:1), to obtain a fraction containing the target product 15. Recrystallization of the fraction in Pet. Et2O gave 15 (

MTT Assay
Compounds 1-12, 5-FU, DOC and taxol were dissolved in DMSO to prepare 10.0 mg/mL stock solutions, respectively, and serial dilutions were made for the MTT assay. 5-FU, DOC and taxol were used as positive control, and DMSO was used as blank control. The MTT assay was performed according to our previous procedure [30][31][32][33][34][35][36][37][38][39][40][41][42][43][44], and exponentially growing K562, HL-60 and HeLa cells in RPMI 1640 medium and S180 cells in DMEM medium were treated with samples at 37 °C for 48 h. The assay was run in triplicate, and the OD value was read at 570 nm. The IR% was calculated using OD mean values by the formula, IR% = (ODcontrol − ODsample)/ODcontrol × 100%, and the IC50 value for a sample was obtained from its IR% values at different concentrations.

The in Vivo Test in Mice for Antitumor Activity of 1 and 2
Compounds 1 and 2 (30 mg of each) and taxol (20 mg) was dissolved in DMSO (1 mL) and mixed with oxidized castor oil (1 mL) to prepare stock solutions of 1 and 2 at 15 mg/mL and taxol at 10 mg/mL, respectively. Stock solutions were repeatedly prepared according to the need. Before administration, each 0.4 mL of the stock solution was diluted with 3.6 mL of 5% glucose aqueous solution to obtain diluted solutions of 1 and 2 at 1.5 mg/mL and taxol at 1.0 mg/mL for each round of administrations. A blank solvent without test samples was also prepared in the same manner and used for the model group as blank control.
The in vivo antitumor activity of 1 and 2 was tested by a procedure similar to that described in [52,53]. Sixty three Kunming mice (18)(19)(20), which had received a hypodermic injection of murine sarcoma S180 cells (0.2 mL each of fresh cell suspensions at density of 1 × 10 7 cells/mL) into the armpit, were randomly divided into six groups: the model (blank solvent) group with 11 mice, positive control taxol (20 mg/kg) group with eight mice, and four test groups with 11 mice in each group for 1 and 2 at both 15 and 30 mg/kg, respectively. The drug administration was performed by intravenous injection via the mouse tail vein from the next day of tumor cell injection. The four test groups were injected every day continuously 6 days, while the taxol group was injected every other day for 6 days (for a total of three administrations). The model group was continuously injected the same volume of blank solvent as that of 30 mg/kg test sample solutions every day for 6 days. On the third day of the last administration, the mice were sacrificed, the body and tumor weights were weighed, and the inhibition rate of 1-2 and taxol on the tumor growth was calculated as described in the main text. Statistical analysis for the test and positive control taxol groups was performed by the Student's t test using body and tumor weights in mean ± S.D. values in comparison with the model group.

Conclusions
Twelve galloyl glucosides 1-12 with two or three galloyl groups were synthesized from D-glucose and gallic acid. Three of them, 9, 11 and 12, were new compounds and six others, 1-4, 6 and 10, were synthesized for the first time. In an in vitro MTT assay, 1-12 inhibited human cancer K562, HL-60 and HeLa cells with inhibition rates ranging from 64.2% to 92.9% at 100 μg/mL and their IC50 values were determined to vary between 17.2-124.7 μM. In addition, 1-12 also inhibited murine sarcoma S180 cells with inhibition rates ranging from 38.7% to 52.8% at 100 μg/mL in an in vitro MTT assay, and the in vivo antitumor activity of 1 and 2 was also detected on murine sarcoma S180 tumor-bearing Kunming mice using taxol as positive control.