Chemical composition, antioxidant and hepatoprotective activities of methanol extracts from leaves of Terminalia bellirica and Terminalia sericea (Combretaceae)

Plant material

Fresh mature plant leaves of Terminalia bellirica were collected from trees growing in Bangladesh Agricultural University Campus, Mymensingh, Bangladesh during the spring season. As for Terminalia sericea, the leaves were collected from the trees at Lupaga Site in Shinyanga, Tanzania. Voucher specimens of leaf samples are kept at IPMB, Heidelberg University under accession numbers P8654 and P7332 for Terminalia bellirica and Terminalia sericea, respectively.

The leaves of Terminalia bellirica (250 g) and Terminalia sericea (250 g) were air-dried, ground, and extracted with 100% methanol at room temperature for three days (6 × 500 mL). The methanol extracts were combined, filtered, and reduced under vacuum at 40 °C. After freezing at −70 °C, the extracts were freeze-dried (lyophilized) yielding fine dried powder with a yield of 10% and 12% for Terminalia bellirica and Terminalia sericea, respectively.

HPLC-PDA-MS/MS

ThermoFinnigan LCQ-Duo ion trap mass spectrometer (ThermoElectron Corporation, Waltham, MA, USA) with an ESI source (ThermoQuest Corporation, Austin, TX, USA) was used to investigate the phytochemical composition of the leaves extracts of both Terminalia species. A C18 reversed-phase column (Zorbax Eclipse XDB-C18, Rapid resolution, 4.6 × 150 mm, 3.5 µm, Agilent, Santa Clara, CA, USA) was employed with a ThermoFinnigan HPLC system. The mobile phase was composed of water, acetonitrile (ACN, Sigma-Aldrich GmbH, Steinheim, Germany) and 0.1% formic acid. Initially ACN was 5% then increased to 30% over 60 min. The flow rate was kept at one mL/min with a 1:1 split before the ESI source. The extracts were injected using ThermoQuest surveyor autosampler. Xcalibur was utilized as operating software (Xcalibur™ 2.0.7; Thermo Fischer Scientific, Waltham, MA, USA). The MS operated in the negative mode as reported before (Sobeh et al., 2017a). The ions were detected in a full scan mode and over a mass range of 50–2,000 m/z. The compounds were identified based on their retention times, molecular weights, and fragmentation patterns as well as comparison with reported data from the same plant, available literature, and authentic compounds.

Biological experiments

Evaluation of hepatoprotective effect

Induction of liver damage

Acute liver failure was induced by intraperitoneal injection of 800 mg/kg dose of D-galactosamine (D-GalN) dissolved in normal saline just before use (Mahmoud et al., 2014). After one week of acclimatization, 42 rats were randomly distributed into seven groups with six animals per group. Group 1 received vehicle (normal saline, 0.9% NaCl) for the whole duration of the experiment and served as the control group. Group 2 received vehicle for three days then, injected with 800 mg/kg D-GalN intraperitoneally. Groups 3 and 4 received Terminalia sericea in an oral dose of 100 and 200 mg/kg, respectively. Groups 5 and 6 received Terminalia bellirica in an oral dose of 100 and 200 mg/kg, respectively. Group 7 received the hepatoprotective standard drug, silymarin in an oral dose of 100 mg/kg and served as positive control. Groups 3–7 received the corresponding drugs or extracts once daily for three consecutive days before injecting D-GaIN (800 mg/kg) dissolved in normal saline.

Blood and tissue sampling

Blood samples were obtained from the retro-orbital plexus 24 h after D-GalN injection. The blood was centrifuged (3,000 g, 4 °C, 20 min) to separate serum. The obtained serum was used to determine liver injury markers and total bilirubin. Thereafter, animals were injected with ketamine for anesthesia and euthanized by decapitation. The animals were dissected; livers were isolated and washed with cold saline to remove clotted blood. Each liver was dissected into two parts. The first part was placed in buffered formalin for further histopathological examination and the second part was flash frozen in liquid nitrogen and stored at −80 °C for determination of oxidative stress markers.

Biochemical analysis

Markers of liver injury including serum aspartate aminotransferase (AST), alanine aminotransferase (ALT) activities, and total bilirubin level were measured colorimetrically (JENWAY 6105 UV/V Spectrophotometer) based on the method of Young (1995) and Balistreri & Shaw (1987), respectively and using a commercial assay kit (Diamond Diagnostic, Cairo, Egypt). Malondialdehyde (MDA) was measured according to the method of Ohkawa, Ohishi & Yagi (1979) using commercially available kits provided by Biodiagnostic (Giza, Egypt). Measurements of total antioxidant capacity (TAC) were performed as described by (Koracevic et al., 2001). In brief, TAC was measured after the reaction of the endogenous antioxidants in serum of rats with certain amount of hydrogen peroxide (H₂O₂). The residual H₂O₂ was measured by reaction with 3,5-dichloro-2-hydroxy benzene sulfonate yielding a colored product, which was detected by spectrophotometry (JENWAY 6105 UV/V Spectrophotometer) at a wavelength of 505 nm.

Histopathological study

The isolated liver tissues were sliced and placed in 10% buffered formalin, dehydrated in ascending concentrations of ethyl alcohol (70%, 90%, and 100%), transferred to xylene, and embedded in paraffin wax. A rotary microtome was used to prepare liver sections of five μm thickness that were stained with H&E stain and examined for histopathological changes under a light microscope (Olympus BX-50 Olympus Corporation, Tokyo, Japan) (Bancroft, Stevens & Turner, 1996).

Immunohistochemical study

The expression of the anti-apoptotic marker Bcl-2 was determined in liver tissues according to (Ozen et al., 2008). Briefly, the obtained liver tissues were deparaffinized in xylene and dehydrated using ethanol. The endogenous peroxidase activity was blocked by the addition of l.5% H₂O₂ in absolute methanol to the liver tissues which were then washed by phosphate buffered saline (PBS) before the addition of Bcl-2 primary antibodies. Streptavidin biotin peroxidase kit was used to measure protein expression. The Bcl-2 chromogen diaminobenzidine (DAB) was used to stain tissues and was then counterstained with hematoxylin for the protein detection (Zhou et al., 2005).

Morphometric analysis

Five randomly high power microscopic fields were selected to count the number of hepatocytes positive to Bcl-2. They were analyzed using computerized image system (a Leica Qwin 500 image analyzer connected to a Leica microscope). This number represents cell number per µm² (Salama et al., 2013).

Statistical analysis

All values of the hepatoprotective study are presented as mean ± standard error of mean. Statistical difference between various groups was determined using GraphPad Prism software (GraphPad, version 5; GraphPad Inc., La Jolla, CA, USA). Statistical significance of the results was performed using one-way analysis of variance test (ANOVA) followed by Tukey’s post hoc test to determine statistical difference between each two separate groups. A p value of <0.05 was considered as statistically significant.

HPLC-PDA-MS/MS identification of secondary metabolites

Structural analysis of polyphenolic compounds found in the leaf extracts of Terminalia bellirica and Terminalia sericea resulted in the separation and tentative identification of 85 compounds in total. Ellagitannins and proanthocyanidins represented the major chemical composition of Terminalia bellirica leaves extract while flavonoid glycosides and stilbenoids were the most abundant phytochemicals in Terminalia sericea leaves extract. The identified compounds along with their retention times, observed molecular and fragment ions are listed in Table 1. LC-MS profiles for Terminalia bellirica and Terminalia sericea leaves extracts in the negative ion mode are shown in Fig. 1.

Several ellagitannins were detected in Terminalia bellirica. For instance, the deprotonated molecular ion peak of compound 31 showed [M – H]⁻ at m/z 785. The MS² experiments revealed an ion at m/z 633 [M – H – 152], which was attributed to the loss of a galloyl group. Another ion was detected at m/z 483 indicating the removal of the hexahydroxydiphenoyl (HHDP) moiety. The ion at m/z 301 was formed due to the elimination of digalloylglucose [M – H – 483]. This fragmentation pattern suggested that the molecule is digalloyl-HHDP-hexoside (Fig. 2A) (Wyrepkowski et al., 2014).

Compound 33 was identified as corilagin. It displayed an [M – H]⁻ at m/z 633 with fragment ions at m/z 463 [M – H – 170] and m/z 301 [M – H – 170 – 162] (Fig. 2B) (Kinoshita et al., 2007). The identification was also confirmed using the available standard compound.

Compound 35 was identified as galloylpunicalagin (Fig. 2C) (Chen et al., 2009). It showed a molecular ion peak [M – H]⁻ at m/z 1235 and its MS² spectrum revealed a main fragment ion at m/z 1083 which corresponded to punicalagin. This was confirmed by two fragments ions at m/z 781 and 601 as previously described by (Fahmy et al., 2016). The ESI-MS analysis of compound 38 displayed a [M – H]⁻ at m/z 635. It showed product ions at m/z 483, 465, and 313 in MS² experiments and thus was identified as tri-galloyl-hexoside (Ben Mansour et al., 2017).

MS data analysis of Terminalia sericea leaves extract revealed the presence of myricetin, quercetin, and resveratrol derivatives as major constituents. Peaks numbered 70 and 71 were identified as quercetin galloyl-glucosides. They showed [M – H]⁻ at m/z 615 and fragment ions at m/z 463 [M – H – 152] and 301 [M – H – 152 – 162] as reported before (Sobeh et al. 2017a).

Several resveratrol derivatives were also detected. In particular, resveratrol glucoside was identified based on its [M – H]⁻ at m/z 389 and a daughter ion at m/z 227 [M – H – 162], (Fig. 2D). Other two peaks showed a deprotonated ion at m/z 535 and two daughter ions at m/z 389 and 227 were characterized as resveratrol coumaroyl-glucoside (Table 1).

Antioxidant activity and total phenolic content

Terminalia bellirica and Terminalia sericea leaves extracts exhibited substantial antioxidant activity in the DPPH and FRAP assays when compared to the well-known antioxidant epigallocatechin gallate (EGCG) from green tea. Folin-Ciocalteu assay revealed high total phenolic content in the studied extracts (Table 2).

Hepatoprotective potential

The effects of Terminalia sericea and Terminalia bellirica leaves extracts (100 and 200 mg/kg dose levels) and the positive control silymarin from Silybum marianum (100 mg/kg) on D-GalN-induced hepatic toxicity in rats were evaluated. A significant increase in the serum levels of AST (p < 0.0001), ALT (p < 0.001) and total bilirubin (p < 0.0001) was observed in the D-GalN treated rats compared to the control group (Fig. 3). Pre-treatment with Terminalia sericea (100 and 200 mg/kg b.w.), Terminalia bellirica (100 mg/kg), and silymarin (100 mg/kg) caused significant reduction in the levels of ALT compared to D-GalN group (p < 0.05). There was no significant difference between the extracts and silymarin (p > 0.05). On the other hand, only Terminalia sericea (in a dose of 200 mg/kg) was able to decrease serum AST level compared to D-GalN group (p < 0.05). The low dose level of Terminalia sericea, the two dose levels of Terminalia bellirica and silymarin did not significantly affect AST enzyme activity compared to D-GalN group (p > 0.05). Terminalia sericea in the high dose level, Terminalia bellirica in the two dose levels and silymarin decreased total bilirubin level when compared to D-GalN group.

D-galactosamine intoxication led to a decrease in the total antioxidant capacity as shown in Fig. 4A. Both extracts failed to restore the total antioxidant capacity (p > 0.05). Silymarin on the other hand was able to show an improvement when compared to D-GalN group (p < 0.0007).

Lipid peroxidation in hepatic tissues, manifested as elevated production of MDA, was increased significantly relative to the control group (p < 0.0018) upon administration of D-GalN (Fig. 4B). No change in the level of MDA was observed in all pre-treated groups relative to D-GalN group.

Structural changes of the liver

Histopathological examination of the liver sections from the control group revealed the usual liver architecture. The hepatic lobules and central veins were normal (Fig. 5A). Liver tissues isolated from D-GalN treated group showed mononuclear cell infiltration in the portal area with congested blood vessels and hyperplasia of bile ducts (Fig. 5B). Liver sections from Terminalia sericea (100 mg/kg) pre-treated rats showed mononuclear cellular infiltration and small areas of hemorrhage (Fig. 5C). Liver sections from Terminalia bellirica (100 mg/kg) pre-treated rats showed mononuclear cellular infiltration in the portal area and congested blood vessels (Fig. 5E). Meanwhile, liver sections from rats pre-treated with Terminalia sericea and Terminalia bellirica in the high dose level (200 mg/kg) showed marked improvement with little mononuclear cellular infiltration (Figs. 5D and 5F). Silymarin pre-treated rats showed partial improvement of the structural changes (Fig. 5G).

Effect on Bcl-2 expression

The anti-apoptotic Bcl-2 protein promotes survival and prevents apoptotic cell death. The current study revealed that D-GalN administration was able to induce free radical generation and decrease the endogenous antioxidants. This was in turn associated with reduction in the number of hepatocytes with positive Bcl-2 expression indicating increased apoptotic cell death (Fig. 6). All extract doses, except the low dose level of Terminalia sericea, were able to increase the positive Bcl-2 expression hepatocytes indicating protection against apoptosis. This effect did not significantly differ from that of silymarin on Bcl-2 expression (p < 0.0001).

Molecular modeling

To verify their ability to interfere with apoptosis and minimize hepatic cell death, the most abundant compounds identified in the investigated extracts were docked into Bcl-2: Bim (BH3) interaction surface, pdb code: 4b4s, with the help of molecular operating environment, 2013.08; Chemical Computing Group Inc., Montreal, QC, Canada, H3A 2R7, 2016 according to our previously applied protocol (Sobeh et al., 2017d). The selected compounds and their docking results are shown in Table 3.

It was reported that a heterodimerization of pro- and anti-apoptotic two protein members into a complex is crucial to activate the apoptotic pathways (Reed, 2000). The surface of the anti-apoptotic Bcl-2 has a hydrophobic cleft to which the BH3 domain of apro-apoptotic protein such as Bim binds. The BH3 domain of Bim extends from Arg 53 to Ala 74 residues. Out of these amino acids, five hydrophobic residues, namely, Ile 58, Leu 62, Ile 65, Phe 69, and Tyr 73 were reported to interact with the Bcl-2 hydrophobic cleft and are conserved in all the pro-apoptotic BH3 domains (Rautureau et al., 2012). Other residues such as Arg 64, Glu 68, and Tyr 72 were reported to contribute with different degrees to the complex binding free energy (Delgado-Soler et al., 2012). Introduced mutation to any of these residues was reported to block apoptosis as it impairs severely the binding of the two proteins (Sattler et al., 1997; Boersma et al., 2008).

Eight highly abundant compounds in Terminalia sericea and Terminalia bellirica leaves extracts were selected for docking into Bcl-2: Bim (BH3) interaction surface. As shown in Table 3, all compounds were able to bind to the binding site with an appreciable scoring function range of – 11.32 to – 24.44. The docked compounds have shown different polar and hydrophobic interactions with the amino acids in the binding site, most importantly Tyr 73 and/or Tyr 72. These interactions could result in changing the protein structural conformation leading to interference with the binding of Bim (BH3) to Bcl-2 and thus blocking apoptosis. Among the glycosides of Terminalia sericea, myricetin rutinoside and quercetin rutinoside have shown the best score of –17.34 and –16.45 respectively followed by the more rigid—by virtue of the alkene moiety in its side chain—resveratrol glucoside, which showed a scoring function of –13.98. As for Terminalia bellirica, the bulky compounds containing hexahydroxydiphenoyl and/or dehydrohexahydroxydiphenoyl units such as punicafolin, corilagin, granatin B, and the benzopyran tannin chebulagic acid have shown the best scores of –24.44, –15.6, –14.9, and –17.7 respectively while the small sized ellagic acid has shown the weakest binding among this group with –12.38 scoring function. Figures 7 and 8 show the 3D poses and 2D-interactions of compounds 42 and 69 as selected examples of the docked compounds.