Phytoconstituents from Adenanthera pavonina L. as antioxidants and inhibitors of inducible TNF-α production in BV2 cells

Abstract Adenanthera pavonina L. has been used traditionally to relieve inflammation. This study aimed to expand the phytochemical knowledge of A. pavonina and evaluate its constituents for their antioxidant and anti-inflammatory potentials as tumor necrosis factor alpha (TNF-α) inhibitors. The antioxidant activity was evaluated using the DPPH radical scavenging assay, and the inhibition of TNF-α was assessed through ELISA and qRT-PCR techniques. Interestingly, five previously undescribed metabolites, including a pentadienoic acid derivative, a triterpenoid glycoside, and three tamarixetin glycosides, were identified alongside seven known compounds. Most compounds evaluated had higher DPPH radical scavenging capabilities than the standard, trolox. Tamarixetin 3-O-(α-ʟ-rhamnopyranosyl)-(1→2)-β-ᴅ-galactopyranoside (11), a previously undescribed compound, was the most effective at suppressing TNF-α protein and m-RNA levels. Other flavonoid glycosides, quercetin 3-O-(α-ʟ-rhamnopyranosyl-(1→2)-β-ᴅ-xylopyranoside (4), isovitixin (5), and quercetin-3-O-[α-ʟ-rhamnopyranosyl-(1→2)]-[α-ʟ-rhamnopyranosyl-(1→6)]-β-ᴅ-galactopyranoside (9), also significantly lowered TNF-α production. Graphical Abstract


Introduction
Inflammation is a serious condition that occurs in conjunction with several diseases afflicting humans (Wang et al. 2021). Chronic inflammation causes damaging cellular and vascular changes, oxidative stress, and over-expression of various pro-inflammatory cytokines including TNF-a and Interleukins (ILs) (Medzhitov 2010). A vast array of pathological conditions are also associated with activation of TNF-a (Giollo et al. 2021). Thus, TNF-a inhibitors could avoid the serious untoward effects of the conventional anti-inflammatory NSAIDs and glucocorticoids (Wongrakpanich et al. 2018).
Adenanthera pavonina L. (Fabaceae) is a wild ethno-medicinal tree endemic to India, South-East China, and Malaysia. Earlier phytochemical research on the plant revealed a variety of bioactive phytochemicals, including sterols, fatty acids, pavonin lactone, flavonoids, and their glycosides (Gennaro et al. 1972;Ali et al. 2005;Mohammed et al. 2014). Several traditional applications of the plant have been reported in diarrhea, hemorrhagic conditions, rheumatism, lung ailments, and gout, reflecting its astringent and anti-inflammatory potentials (Khare 2007). Olajide et al. have demonstrated the analgesic and anti-inflammatory effects of its seed extract (Olajide et al. 2004). A decoction containing A. pavonina has displayed a cytotoxic effect against human laryngeal carcinoma (Lindamulage and Soysa 2016). These traditional uses and medicinal importance fueled our interest in exploring the incompletely studied phytochemical profile of the aerial parts.
Flavonoid glycosides were found to be the major components in this phytochemical investigation of A. pavonina. Several studies have proved the antioxidant and antiinflammatory activities of flavonoid derivatives through various mechanisms (Okawa et al. 2001;Choy et al. 2019;Habib et al. 2020;Owor et al. 2020). Based on the reported anti-inflammatory activities of A. pavonina and structurally related compounds to the plant's isolated constituents, selected isolates were assessed for their antioxidant and anti-inflammatory capacities to attenuate the LPS-induced inflammatory response in BV2 cells.

Phytochemical investigation
The chromatographic analysis resulted in isolation and identification of five previously undescribed compounds (1, 3, 8, 11, and 12) (Chart 1). Compound 1, a white amorphous powder, had a molecular formula of C 20 H 28 O 10 with seven degrees of unsaturation based on the detected molecular ion peak at m/z 427.1608 [M-H] À (calcd 427.1604) in High Resolution Electrospray Ionization Mass Spectrometry (HRESIMS) spectrum (negative mode). The 1 H NMR spectrum displayed two trans-olefinic protons at d H 7.90 and 6.35 (each d, J ¼ 15.9 Hz), a methyl singlet at d H 2.03, and an olefinic proton H-2 at d H 5.75. These signals are characteristic features for the substructure (3-methyl-2E,4E-pentadienoic acid) (Sasaki et al. 1991). HMBC correlations ( Figure S6, supplementary material) supported the position of methyl group and other functions in this substructure. For example, H-2 showed strong HMBC correlations with two carbons that resonating at d C 131.0 and 20.6. The configuration of the two double bonds was assigned as trans based on the coupling constants of H-4 and H-5 along with NOESY correlations between H-5 and 3-Me group. In addition, a substituted cyclohexyl moiety, attached to C-5, was deduced from the detailed analysis of NMR spectral data (Table  S1, Figures S1-S7, supplementary material). The 1 H NMR spectrum exhibited two singlets corresponding to additional two methyl substituents on the cyclohexyl moiety resonating at d H 1.24 and 0.95. The former methyl group showed significant HMBC correlations to carbons at d C 81.0, 87.9, and 41.3 (C-1 0 , 2 0 , and 3 0 ) confirming its angular location at C-2 0 . The other methyl group showed HMBC correlations with a quaternary carbon at d C 51.8 (C-6 0 ) and a carboxy group at d C 178.2 (6 0 -COOH) suggesting the presence of methyl and carboxylic functions attached to C-6 0 . An oxygenated methine (H-4) was detected at d H 3.63 and located at C-4 0 flanked by the two methylenes (C-3 0 and C-5 0 ) of the cyclohexyl ring based on 1 H-1 H COSY correlations. The b-configuration of the hydroxyl function at C-4 0 was judged from the observed multiplicities of H-4 (Rodr ıguez-Dele on et al. 2019). The DEPTQ-135 spectrum confirmed the deduced structural fragments and showed the presence of two additional carbonyl groups at d C 173.9 and 177.3 together with a methyl group at d C 17.3 (d, 7.15), a methylene at d C 38.9, and a methine at d C 36.1 suggesting the presence of (2-methylsuccinyl) fragment. The HMBC and 1 H-1 H COSY correlations ( Figure S4-6, supplementary material) clearly confirmed the succinyl structural assignment. The relative configurations at C-2 00 , C-2 0 , and C-6 0 were postulated through the NOESY experiment which showed no cross peaks between H-4 0 and any of the methyl signals 2 0 -Me, 6 0 -Me or 2 00 -Me. The configuration at C-1 0 was not identified because the small amount of 1 limited further analysis. Accordingly, the structure of 1 was identified as 5-[1 0 ,4 0dihydroxy-2 0 ,6 0 -dimethyl-6 0 -carboxy-2 0 -(2 00 -methyl-succinyloxy) cyclohexyl]-3-methyl-2E,4E-pentadienoic acid which is a previously undescribed natural metabolite. This class is reported herein for the first time from the genus.
The genus Adenanthera has been reported to have flavonoid glycosides; however, no previous studies have reported pentadienoic acid derivatives or hopane-type triterpene glycosides.

Effect of some isolated compounds on BV2 microglial cells viability
All investigated isolates did not exhibit cytotoxic effects on BV2 cells as indicated by their high IC 50 values. Calculated IC 50 values of the tested compounds are shown in (Table S4, supplementary material). These data indicate that the tested isolates have no influence on cell viability; hence, concentration of 100 mM was used in subsequent experiments.

Effect of some isolated compounds on LPS-induced TNF-a activation
Elevated concentrations of both TNF-a protein and m-RNA were successfully induced in BV2 cells through LPS exposure to serve as a positive control. TNF-a was chosen as it plays an important role in inflammations, and its dysregulation has been linked to many human diseases (Halaris et al. 2012). The effects of compounds (3-7, 9, 11, and 12) on the production of TNF-a were assessed; however, compounds 1, 2, 8, and 10 were excluded from biological evaluations because of insufficient yields. The highest suppression in TNF-a protein level was observed with 11, which was comparable to the normal secreted amount (the negative control). Compounds 4, 5, and 9 also displayed significant overall reduction of TNF-a release (Table S5, supplementary material). To determine whether the tested compounds inhibited TNF-a production at the transcriptional level or had post-transcriptional effects, the qRT-PCR analysis, a parallel experiment, was done. After 6 h, these compounds significantly reduced TNF-a mRNA expression, implying that they inhibited the gene expression of the tested cytokine (Table S5, supplementary material).
We found that quercetin derivatives 4, 9, and 11 reduce the pro-inflammatory mediator TNF-a secretion and downregulate the regulatory genes in LPS-stimulated BV2 microglia cells. These findings support a previous study in which quercetin has been shown to significantly reduce TNF-a production and gene expression in a dose-dependent manner, indicating its ability to regulate the immune response and potential anti-inflammatory properties (Nair et al. 2006).

Antioxidant activities
All tested compounds, except 5, effectively scavenged the DPPH radicals with IC 50 values less than that of the positive control, trolox (IC 50 47.85 mg/mL). Compounds 6 and 7 were the most potent antioxidants with IC 50 values of 18.24 and 16.42 mg/mL, respectively. The results are shown in (Table S6, supplementary material). These findings were consistent with previous structure activity studies that have linked antioxidant activity to the presence of 2, 3-unsaturation, 3 0 , 4 0 catechol moiety, 3-OH group, and 4-keto functional group (Aderogba et al. 2012b). Accordingly, the lack of a catechol group in compound 5 (isovitexin) resulted in its inactivity, whereas co-existence of all required structural aspects was behind the observed antioxidant activities of other quercetin glycosides investigated.

Plant material
The aerial parts of A. pavonina were collected in August 2017 from Aswan Botanical Garden, Aswan, Egypt. Dr. Hafeez Rofaeel has authenticated the plant. A voucher specimen (29,313) has been deposited at the herbarium of Flora and Phytotaxonomy Research, Horticultural Research Institute, Agricultural Research Center, Dokki (Cairo), Egypt.

Extraction and isolation
The dried pulverized aerial parts (4.5 kg) were repeatedly macerated using 70% methanol till exhaustion. A dried residue of 295 g was obtained after the concentration of the combined hydro-alcoholic extracts at 40 C under reduced pressure. The obtained residue was re-suspended in distilled water and partitioned successively using n-hexane, dichloromethane, and ethyl acetate to give three fractions (F 1 -F 3 ), respectively. The remaining aqueous solution from the partitioning process was dried to provide F 4 . Detailed information about the chromatographic separation of each compound is included in the supplementary material.

Tamarixetin
3.4.4. Determination of mRNA expression of TNF-a by quantitative real-time polymerase chain reaction qRT-PCR Total RNA was extracted from cell lysates using TRIzol reagent according to the manufacturer's instructions. The RNA was then reversed to create cDNA. For qRT-PCR, SYBR Green PCR Master Mix was used. The relative differences in expression between groups were expressed using Ct values. Data were normalized with b-actin. Relative differences between control and treatment groups were calculated. Primers sequences for TNF-a and b-actin were listed in (Table S7, supplementary material).

The antioxidant activities of the tested phytochemicals using DPPH antioxidant assay
The antioxidant activities of the tested phytochemicals and trolox were evaluated using the DPPH scavenging assay. Briefly, in a 96-well plate, the DPPH solution was first distributed to each well, and then different concentrations of each tested phytochemical (100-0.1 ng/ml) were added to each well. The plate was incubated for 30 min at room temperature in the dark, and finally, absorbance was recorded at 595 nm wavelength. To authenticate the process, trolox at different concentrations were used as a standard. The degree of scavenging was calculated by the following equation: Scavenging effect ð%Þ ¼ f Absorbance of control À Absorbance of sample ð Þ =Absorbance of controlg Â 100 3.4.6. Statistical analysis All values were analyzed by the GraphPad Prism 7.0 (GraphPad software, USA) and expressed as mean ± SEM. All the statistical analyses were performed using ANOVA test followed by turkey's post hoc comparison test applied across all groups. Differences with P < .05 were considered statistically significant.

Conclusions
Five new compounds were reported from A. pavonina L. aerial parts. Except for 5, all tested compounds showed higher antioxidant activities than the control, trolox. Compounds 4, 5, 9, and 11 showed anti-inflammatory potentials as indicated by TNFa suppression in LPS-stimulated BV2 cells, with 11 having the greatest inhibitory effect on TNF-a protein and m-RNA levels. These findings revealed the antioxidant and antiinflammatory attributes of A. pavonina components, which deserve further investigations.