Tannin-rich extracts from Lannea stuhlmannii and Lannea humilis (Anacardiaceae) exhibit hepatoprotective activities in vivo via enhancement of the anti-apoptotic protein Bcl-2

The potential hepatoprotective activities of two Lannea species were explored in vivo. Furthermore, the binding activities of their main polyphenols to the antiapoptotic protein Bcl-2 were investigated. Based on HPLC-MS/MS results, 22 secondary metabolites were characterized in L. stuhlmannii (mainly tannins), while 20 secondary metabolites (mainly sulphated tannins) were identified in L. humilis. Both extracts exhibited substantial antioxidant activities in vitro and counteracted D-galactosamine induced intoxication in rats in vivo and increased the total antioxidant capacity (TAC) of liver tissues. In addition to reducing the elevated levels of AST and total bilirubin, both extracts significantly attenuated the deleterious histopathologic changes in liver after D-galactosamine-intoxication. Also, both extracts protected hepatocytes from apoptotic cell death and increased the expression of the anti-apoptotic protein Bcl-2. The identified compounds from both extracts can bind to the Bcl-2: Bim (BH3) interface with an appreciable binding free energy. Hydrogen and ionic bonds and hydrophobic interactions with amino acid residues in the hydrophobic face of Bim (BH3) domain were discovered. To sum up, L. humilis and L. stuhlmanni exhibited promising hepatoprotective activities in vivo against D-GalN-induced liver injury and their hepatoprotection is due to the antioxidant and anti-apoptotic effects of tannins and proanthocyanidins.

products distributed over different phytochemical classes such as tannins, cyanidins, flavonoids, alkylphenols, and dihydroalkyhexenones 4,5 . For instance, a prenylated chromenflavanone was isolated from L. acida leaf extract 6 . Also, three dihydroalkylhexenones were discovered in L. edulis in a bio-guided fractionation 5 .
Several biological activities have been reported from the genus Lannea. In particular, a L. coromandelica leaf extract has shown anti-conceptive activities while its bark extract was found to have hypoglycemic activities, in addition to anti-inflammatory activities in rats 7,8 . Okoth et al. have reported antioxidant and antibacterial activities for L. alata root and the isolated compounds myricitrin, myricetin-3-O-a-arabinofuranoside, lupeol, and sitosterol 4 .
Lannea stuhlmannii and L. humilis are deciduous trees widely distributed in the tropics 3 . In Tanzania, Lannea species are being used traditionally for the treatment of cancer. Also, decoctions of L. stuhlmannii root are still in  use against fever, asthma, stomachache, and dysentery while decoctions of the stembark are considered beneficial to counteract headache and stomach pain. The leaf extract is used to relieve abdominal pain and to hasten childbirth. Moreover, stembark extract has shown antibacterial activity 9 . In Namibia, L. stuhlmannii roots are used as ethnomedicine for the treatment of HIV/AIDS infections. In East Africa, the plant is employed traditionally in treatment of anaemia 10 . The bark of L. humilis is used traditionally to treat fever, and diarrhea 11 . In Tanzania, a root decoction is drunk to treat anaemia, stomach pains, nausea and general body weakness 12 . L. humilis trees excrete a water-soluble gum rich in amino acids, galactose, and other polysaccharides 13,14 . The root of L. stuhlmannii and the bark from L. humilis demonstrated promising anti-trypanosomal and cytotoxic activities in a panel of 40 plants extracts 15 .
Phytochemical studies are missing for both species. However, a preliminary study by Chhabra et al. reported the presence of some alkaloids and flavonoids in L. stuhlmannii bark 16 . Herein we characterized the polyphenolic constituents of L. stuhlmannii and L. humilis bark extracts using HPLC-PDA-ESI-MS/MS. Moreover, in vitro antioxidant activities were studied and hepatoprotective activities were investigated in rats treated with the toxic D-galactosamine that causes liver damage. Molecular modelling was employed to find out if the main polyphenols of the plants bind to Bcl-2: Bim (BH3) interface that might be modulated by the drug.

Materials and Methods
Plant material. Lannea humilis (Oliv.) Engl. and Lannea stuhlmannii (Engl.) Engl. bark samples were collected from trees growing in Lupaga Site in Shinyanga, Tanzania and kept under accession number P7311 and P7317, respectively, at IPMB, Heidelberg University 15 . The bark samples were ground and extracted with 100% methanol at room temperature for three days (6 × 500 mL). The combined extracts were filtered and evaporated using a rotary evaporator at 40 °C. The residues were frozen at −70 °C, and then lyophilized to fine dried powder. The extraction yield was calculated based on initial dry weight and found to be 22% and 15% for L. humilis and L. stuhlmannii bark, respectively. HPLC-PDA-MS/MS. The extracts were analyzed utilizing a ThermoFinnigan LCQ-Duo ion trap mass spectrometer (ThermoElectron Corporation, Waltham, Ma, USA) with an ESI source (ThermoQuest Corporation, Austin, Tx, USA). A ThermoFinnigan HPLC system using a C18 reversed-phase column (Zorbax Eclipse XDB-C18, Rapid resolution, 4.6 × 150 mm, 3.5 µm, Agilent, Santa Clara, CA, USA) was used 17 . Water and acetonitrile (ACN) (Sigma-Aldrich GmbH, Germany) (0.1% formic acid each) were used as a mobile phase. At 0 min, ACN was 5% and then increased to 30% in 60 min at 1 mL/min with a 1:1 split before the ESI source. Autosampler surveyor ThermoQuest was utilized to inject the extracts and the system was controlled by Xcalibur software (Xcalibur TM 2.0.7, Thermo Fischer Scientific, Waltham, Ma, USA). The MS operated in the negative mode as reported before 18 . The ions were detected in a full scan mode and mass range of 50-2000 m/z.

Biological experiments. Antioxidant activities in vitro.
The Folin-Ciocalteu method quantified the total phenolic content at the extracts as described before 19 . The 2,2′-diphenyl-1-picryl-hydrazyl-hydrate (DPPH) radical scavenging and ferric reducing antioxidant power (FRAP) assays were utilized to evaluate the antioxidant properties of the extracts and were carried out as described earlier 19 .  group (total animals: n = 42). To induce acute liver failure, the rats were injected with D-GalN (800 mg/kg, i.p). The animals were treated with vehicle, extracts or silymarin for three consecutive days before the injection of D-GalN. Blood samples were drawn from the retro-orbital plexus of the eye 24 h after D-GalN injection under pentobarbital anesthesia (50 mg/kg, i.p.). The serum was separated by centrifugation at 3500 × g for 15 min at 4 °C. These serum samples were used for measurement of AST activity and total bilirubin level. The livers were excised and used for light microscopic observations, oxidative stress markers measurement and immunohistochemical staining of Bcl-2.
Liver proteins. Serum aspartate aminotransferase (AST) activity was measured using a commercial assay kit, (Diamond diagnostic, Egypt). Total bilirubin levels were measured using commercially available analytical kit (BioMed Diagnostics, Egypt). Determination of oxidative stress markers. The generation of reactive oxygen species following D-GalN injection was detected in liver tissues using commercially available kits provided by Biodiagnostic (Giza, Egypt) to measure the lipid peroxidation product content, malondialdehyde (MDA).Measurement of Total antioxidant capacity (TAC) was performed by reacting with the endogenous antioxidants in the serum sample with certain amount of hydrogen peroxide (H 2 O 2 ). The remaining H 2 O 2 was measured by an enzymatic reaction with 3,5-dichloro-2-hydroxy benzene sulphonate producing a colored compound. The intensity of the resulting color was measured by spectrophotometer (JENWAY 6105 UV/V Spectrophotometer) at wavelength of 505 nm.
Histology. Excised livers were processed for light microscopic examination, according to standard procedures. The livers were preserved in phosphate-buffered 10% formalin, after which the livers were cut into small pieces, embedded in paraffin wax, sectioned at 4 µm thickness, and stained with haematoxylin and eosin. Immunohistochemical determination of Bcl-2 (anti-apoptotic marker). Liver sections were prepared by soaking them in xylene to remove paraffin and graded alcohol was used (70%, 90%, 100%). The activity of endogenous peroxidase was blocked by addition of 5% H 2 O 2 in absolute methanol for 10 min at room

Results and Discussion
The methanol extracts obtained from the bark of L. stuhlmannii and L. humilis were analyzed by HPLC-PDA-ESI-MS/MS. A total of 22 and 20 secondary metabolites were tentatively identified in L. stuhlmannii and L. humilis bark, respectively. L. stuhlmannii extract is rich in condensed tannins, whereas sulphated condensed tannins and phenolic acids dominated in L. humllis (Table 1 (Fig. 3a,b).  (Fig. 4a,b).  Biological activities. Antioxidant and hepatoprotective activities. Catechins are renowned for their powerful potential to scavenge various free radicals such as hydroxyl, peroxyl, superoxide, and other radicals. Antioxidant activity of catechins is mediated through different mechanisms. They are able to transfer an electron to bind a reactive radical, while they change to the more stable and less reactive phenoxyl radical. They are also able to chelate Cu 2+ and Fe 3+ ions, thus limiting free radicals generation. Indirectly, catechins exert an antioxidant effect via increasing the level of endogenous antioxidants such as glutathione reductase, catalase, and superoxide dismutase. Moreover, catechins are reported to have an inhibitory effect on xanthine oxidase that catalyzes the metabolism of purines into uric acid and reactive oxygen species. A direct relationship between the number of the hydroxyl groups a catechin possess and the exerted antioxidant potential has been demonstrated 17,18 .
Here, the antioxidant potential of the methanol extract of bark from L. humilis and L. stuhlmannii was investigated using two assays, namely DPPH and FRAP. The total phenolic content (TPC) was determined using Folin Ciocalteu assay and expressed as mg gallic acid (GAE)/g extract. Both extracts showed substantial activities in all assays, data are shown in Table 2. Such promising scavenging activities were reported from other Lannea species such as L. alata 4 , other bark-tannins rich extracts from Albizia harveyi, Senna singueana, Turraea fischeri, and proanthocyanidin rich root extracts from Cassia abbreviata and Ximenia americana var. caffra 17,18,21,23,24 .
In another set of experiments, the in vivo antioxidant potential was examined in rats treated with toxic D-galactosamine. We determined the total antioxidant capacity (TAC) and lipid peroxidation product, MDA in liver tissues as markers for oxidative stress. Except for the low dose level of L. stuhlmannii, both extracts were able to counteract the D-galactosamine-intoxication and increased the TAC of liver tissues (Fig. 8A). However, the low dose levels of L. humilis and L. stuhlmannii (100 mg/kg b.w.) were able to attenuate the MDA increase (Fig. 8B).

Extract effects on markers of liver dysfunction and histopathological changes. To test whether
the two studied Lannea species have hepatoprotective activities in vivo, the biochemical and histopathological changes were evaluated after a liver cell injury had been induced by single dose injection of D-galactosamine (D-GalN, 800 mg/kg, i.p.). All treatments reduced AST and total bilirubin levels (p < 0.05), except the high dose of L. humilis which failed to decrease serum AST activity (p > 0.05), Fig. 9.
To confirm the in vivo antioxidant activities and the observed biochemical changes, a histopathological evaluation was performed. It is obvious from Fig. 10B that D-GalN has induced a mononuclear cellular infiltration indicating the development of inflammation in liver tissues (arrow head) in the portal area and congested blood vessel (BV). Together with the observed hyperplasia of bile ducts (arrow), these changes provide evidence for the     successful establishment of animal liver injury model. In general, these deleterious changes were significantly attenuated by L. humilis and L. stuhlmannii treatment. However, a dose of 100 mg/kg L. humilis has shown little improvement and some mononuclear cellular infiltration (arrow) was spotted in the portal tract area and additionally a congested blood vessel (BV) was detected (Fig. 10C). A higher dose of L. humilis (200 mg/kg) showed marked improvement indicated by little mononuclear cellular infiltration (arrow) and central vein (CV) with minimum congestion, (Fig. 10D). The low dose level (100 mg/kg) of L. stuhlmannii had little improvement as seen in Fig. 10E, where mononuclear cellular infiltration (arrow head) was clearly noticed in the portal area accompanied with bile duct hyperplasia (arrow). Figure 10F revealed that the high dose level (200 mg/kg) of L. stuhlmannii was able to exert some improvements where mononuclear cellular infiltration was observed only in the portal tract area (arrow head) and less congested blood vessel (BV) was also observed. Figure 10G showed that silymarin pretreatment (a lignin with known hepatoprotective properties) caused partial improvement.
Extract effects on anti-apoptotic marker Bcl-2. When cellular antioxidants are depleted and ROS accumulation reaches a critical threshold, mitochondrial damage occurs. This process leads to the release of cytochrome c (Cyt-c) from the mitochondria, which in turn activates the apoptotic pathway 25 . Several proteins can inhibit or decrease apoptosis, among them the anti-apoptotic members of Bcl-2 family proteins 26 . To clarify the mechanism that the extracts attenuated hepatotoxicity induced by D-GalN, we stained liver sections with avidin-biotin peroxidase stain with haematoxylin counter stain for Bcl-2, anti-apoptotic marker, and determined the percentage of positively stained hepatic cells. A significant increase in the number of Bcl-2 positive hepatocytes was observed in rats treated with all dose levels of the extracts (Fig. 11C-F), suggesting that both Lannea species protected hepatocyte from apoptotic cell death. The extracts produced similar effect to silymarin (Fig. 11G).

Molecular modelling study.
The family of B-cell lymphoma-2 (Bcl-2) proteins is known for regulating programmed cell death through mediating mitochondrial or intrinsic initiated apoptosis. The Bcl-2 proteins are either pro-survival (anti-apoptotic) such as Bcl-XL, Bcl-2, Bcl-B, and Mcl-1 proteins or anti-survival (pro-apoptotic) proteins including Bim, Bad, Bak, Bax, and others 27 .
The molecular structures of the majority of these proteins have been resolved. The anti-apoptotic members have been found to share one or even more of the crucial BH domains namely BH1, BH2, BH3, and BH4. All the pro-apoptotic proteins, however, contain the BH3 domain that is very essential for cell killing activity 28 .
In order to get an insight how the plant extracts were able to protect against hepatic cells death, we decided to dock some of the identified compounds to the crystal structure of the interaction surface of Bcl-2:Bim (BH3) complex, pdb code: 4b4s, using molecular operating environment (MOE), 2013.08; Chemical Computing Group Inc., Montreal, QC, Canada, H3A 2R7, 2016 according to our previously applied protocol 24 . The binding site in the Bcl-2: Bim (BH3) crystal structure as well as the amino acid residues crucial for the efficient binding between the two proteins are well established, thus docking the plant major chemical components to this site would be beneficial to understand how they are able to hinder the Bcl-2: Bim complex formation and thus blocking apoptosis. The docked compounds are shown in Fig. 12 and docking results are summarized in Table 3.
Despite the fact that the actual mechanism by which these proteins regulate apoptosis is still not clearly understood, it is well accepted now that this happens through heterodimerization of two members of pro-and anti-apoptotic agents into a protein complex that activates apoptotic pathways 29 . Resolving the loop residues in the Bcl-2:Bim (BH3) complex for instance, has shown that Bcl-2 surface has a hydrophobic cleft in which the hydrophobic face of the BH3 of Bim binds. In this regard, Bim (BH3) domain was shown to extend from Arg 53 to Ala 74 amino acid residues. Within this domain, the side chains of five hydrophobic residues are interacting with the hydrophobic cleft of Bcl-2 and conserved in all pro-apoptotic BH3 domains, namely Ile 58, Leu 62, Ile 65, Phe 69, and Tyr 73 30 . Additionally, some more amino acid residues were later reported to contribute favorably to the binding free energy of the complex such as Arg 64 and Glu 68. The residue Tyr 72 contributes to a lesser extent to the complex formation as its side chain is more solvent exposed 31 . Studies have shown that a single mutation in any of these residues such as Leu 62, for instance, could lead to a severe impairment in the binding of the two opposing proteins and hence blocking the apoptotic cell death 32,33 .
As shown in Table 3, the docked compounds were able to bind to the Bcl-2: Bim (BH3) interface with a scoring function of a range between −11.98 to −15.94 reflecting an appreciable binding free energy. Additionally, the compounds have afforded different polar and hydrophobic interactions to the amino acid residues in the hydrophobic face of Bim BH3 domain, (Table 3). Except for 6c, all other compounds with S-configuration of the stereo-center(s) were able to interact with the essential conserved Tyr 73 and/or Tyr 72 residues. However, compounds 27b and 40b having stereo-centers with R-configuration have also shown to interact with Tyr 73 residue. This interaction stabilized by different other polar and hydrophobic interactions could result in conformational changes in the protein structure leading to interference with dimerization of the two opposing proteins Bcl-2: Bim, thus hindering apoptosis. Figures 13 and 14 show the 3D poses and the 2D-interactions, respectively, of two selected compounds34a, 40a, and their enantiomers 34b, and 40b docked to the Bcl-2: Bim (BH3) interface.
In view of the immunohistochemical and molecular modeling studies, the anti-apoptotic effect of the extracts could be presumably attributed to their tannin content. This comes to agreement with other studies that have reported anti-apoptotic activity for proanthocyanidin rich extracts such as the bark extract of Senna sengueana 21 . Polyphenols are powerful biologically active natural products which can bind to several important proteins in cells by forming multiple hydrogen and ionic bonds that can modulate the 3D structure of proteins 34 . In summary, evidence is provided that L. humilis and L. stuhlmanni protect against D-GalN-induced liver injury. The hepatoprotection of the extracts is due to antioxidant and anti-apoptotic effects.

Conclusions
The chemical profiling of the methanol extracts of Lannea stuhlmannii and L. humilis bark was analyzed by HPLC-PDA-ESI-MS/MS. A total of 42 secondary metabolites were identified in the investigated extracts. Noteworthy, this work provides the first comprehensive phytochemical study of L. stuhlmannii and L. humilis bark and confirms their traditional use. The extracts could protect rat liver from intoxication with D galactosamine, apparently through antioxidant and increasing the antiapoptotic protein Bcl-2. Thus, Lannea is another candidate for the potential treatment of liver injury that needs further and deeper analyses.