Bioorganic Chemistry 92 (2019) 103258

Contents lists available at ScienceDirect

Bioorganic Chemistry

journal homepage: www.elsevier.com/locate/bioorg

Effects of quassioides and its active constituents on Alzheimer's T disease in vitro and in vivo Eryan Guoa, Yunwei Hub, Tao Dua, Huilin Zhua, Lei Chenb, Wei Qua,c, Jie Zhanga, Ning Xied, ⁎ ⁎ Wenyuan Liub, Feng Fenga,c,e, , Jian Xua, a Department of Natural Medicinal Chemistry, School of Traditional Chinese Pharmacy, Pharmaceutical University, Nanjing 210009, People’s Republic of China b Key Laboratory of Drug Quality Control and Pharmacovigilance (China Pharmaceutical University), Ministry of Education, China Pharmaceutical University, Nanjing 210009, People’s Republic of China c Key Laboratory of Biomedical Functional Materials, China Pharmaceutical University, Nanjing 211198, People’s Republic of China d State Key Laboratory of Innovative Natural Medicines and TCM Injections, Jiangxi Qingfeng Pharmaceutical Co., Ltd., Ganzhou 341000, Jiangxi, China e Jiangsu Food & Pharmaceutical Science College, Huaian 223003, People’s Republic of China

ARTICLE INFO ABSTRACT

Keywords: Alzheimer disease (AD), a prevalent neurodegenerative disorder, is one of the leading causes of dementia. However, there is no effective drug for this disease to date. Picrasma quassioides (D.Don) Benn, a Chinese tra- Alkaloids ditional medicine, was used mainly for the treatment of inflammation, fever, microbial infection and dysentery. Alzheimer’s disease In this paper, we reported that the EtOAc extract of Picrasma quassioides stems showed potential neuroprotective

activities in L-glutamate-stimulated PC12 and Aβ25-35-stimulated SH-SY5Y cell models, as well as improved memory and cognitive abilities in AD mice induced by amyloid-β peptide. Moreover, it was revealed that the

anti-AD mechanism was related to suppressing neuroinflammatory and reducing1-42 Aβ deposition using ELISA assay kits. To clarify the active components of the EtOAc extract of Picrasma quassioides stems, a systematic phytochemistry study led to isolate and identify six β-carboline alkaloids (1–6), seven canthin-6-one alkaloids (7–13), and five quassinoids (14–18). Among them, four β-carbolines (1–3, and 6) and six canthin-6-ones (7–11, and 13) exhibited potential neuroprotective activities in vitro. Based on these date, the structure-activity re- lationships of alkaloids were discussed. Furthermore, molecular docking experiments showed that compounds 2 and 3 have high affinity for both of dual-specificity tyrosine phosphorylation-regulated kinase 1A (DYPKIA) and butyrylcholinesterase (BuChE).

1. Introduction reducing neuronal cell death, and suppressing neuronal hyperexcit- ability [8]. Picrasma quassioides (D.Don) Benn, belonging to the family Alzheimer disease (AD), one of the prime leading causes of de- , is widely distributed in southern China, Korea, and mentia, is a prevalent neurodegenerative disorder with the character- Japan. This was used as anti-inflammation and anti-microbial ization of language disorder, irreversible memory loss, and cognitive infection agent in Chinese traditional medicine [1–3]. Its preparations, impairment [9,10]. The World Alzheimer Report 2015 showed that such as Kumu injection, Kumu mixture, and Xiaoyan Lidan tablets, have over 46 million people have been suffering from AD and this number is been widely used as anti-inflammatory agents. Previous studies re- expected to exceed 131.5 million in 2050 [11]. Currently, there are five vealed that the main chemical constituents of this plant are canthin-6- prescription drugs approved by the U.S. FDA for the symptomatic one alkaloids, β-carboline alkaloids, and quassinoids, and alkaloids are treatment of AD, including acetylcholinesterase inhibitors (donepezil, considered as the major anti-inflammatory active components [4,5]. galantamine, and rivastigmine), and N-methyl-D-aspartate (NMDA) re- Moreover, several alkaloids have been reported to possess neuropro- ceptor (antagonist, and memantine) [12]. Unfortunately, these drugs tective activity [6–8]. For example, picrasidine O exhibited a cerebral only provide symptomatic relief [13]. Hence, it is urgent to develop protective effect through boosting learning and memory performance, more effective and safer drugs and therapeutic strategy forAD.

⁎ Corresponding authors at: Department of Natural Medicinal Chemistry, School of Traditional Chinese Pharmacy, China Pharmaceutical University, Nanjing 210009, People’s Republic of China. E-mail addresses: [email protected] (F. Feng), [email protected] (J. Xu). https://doi.org/10.1016/j.bioorg.2019.103258 Received 29 April 2019; Received in revised form 30 July 2019; Accepted 4 September 2019 Available online 05 September 2019 0045-2068/ © 2019 Elsevier Inc. All rights reserved. E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258

The etiology of AD remains unclear due to complex and diverse 2.2.4. Cell proliferation assay causes, but several factors, such as neuroinflammation, Aβ deposits, tau EA, NB, and WT were dissolved in DMSO (100 mg/mL) and then hyperphosphorylation, DYRK1A overexpression, oxidative stress, low diluted in medium with 1% FBS to achieve final concentrations. levels of choline (acetylcholine, butyrylcholine), etc., have been greatly Neuroprotection of those extracts were assayed by MTT method on implicated in the AD pathogenesis [12,14]. The neuroinflammation is 10 mM L-glutamate-induced PC12 and 5 μM Aβ25-35 induced SH-SY5Y considered as one of the pathological hallmark of AD [15]. In the pa- cell models. Briefly, the cells (53 ×10 /well) were seeded into 96-well thological state of AD, microglia and astrocytes become activated and plates and incubated overnight. The treatment group were pretreatment release pro-inflammatory cytokines and neurotoxic materials, including with different concentrations of extracts (1, 5, and 10 μg·mL−1, re- tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), IL-6, nitric spectively) for 2 h. Then, the medium was removed, and the mixtures oxide (NO), and superoxide through activating nuclear factor kappa B with 10 mM L-glutamate (or 5 μM Aβ25-35) and different concentrations (NF-κB) [16,17]. And a series of studies showed that overexpression/ of extracts (1, 5, and 10 μg·mL−1) were added. After incubation for activation of dual-specificity tyrosine phosphorylation-regulated kinase 24 h, 20 μL of MTT solution (5 mg/mL in PBS) was added for each well 1A (DYRK1A) can led to increase Aβ deposits and tau hyperpho- and the cells were incubated in dark for another 4 h. After removal of sphorylation via activating multiple signaling pathways, and DYRK1A the media, 150 μL of DMSO was added. The cell proliferation rate was is considered to be an important target in AD therapy [14]. Besides, the measured, using a microplate reader (Thermo Fisher Scientific Co., BChE is markedly activated in the brains of AD patients, and BuChE Waltham, MA, USA) at 492 nm. inhibitors have become a therapeutic trend for AD [18,19]. However, Cell proliferation rate()( % = A A ) the one-compound-one-target therapeutic strategy for AD has failed, 490 treated 490 model which led to the concern is growing that multi-target therapeutic /A( 490 blankA 490 model) × 100%. strategy could be more promising [20]. Herbal medicines, characterized by multi-component and multi-target, are generally recognized to be effective and safe. Hence, many studies aiming to find a desirable 2.3. Anti-AD tests therapy for AD has focused on traditional medicines. Based on the above considerations, we investigated the neuropro- 2.3.1. Animals and reagents tective effects of different extracts of P. quassioides stems in vitro and in 48 ICR mice (weighing 20 ± 2 g; Qinglong Mountain Animal vivo. Then, we clarified the active constituents of EtOAc extract of Breeding Farm, Nanjing, China) were accommodated in standard an- Picrasma quassioides stems though systematic phytochemical and imal room at temperature 23 ± 2 °C, 60 ± 5% humidity and 12-h pharmacology studies. day/night cycle with free obtain to water and food. The IL-1β, TNF-α, and IL-6 Elisa assay kits were purchased from NeoBioscience

Technology Co., Ltd. The Aβ1-42 Elisa assay kit was purchased from 2. Material and methods Nanjing Jin Yibai Biological Technology Co., Ltd. EA was suspended in 0.5% sodium carboxymethyl cellulose (CMC-Na) to obtain different 2.1. Plant materials concentration of EA suspensions. Donepezil (Jiangsu Haosen Pharmaceutical Co., Ltd) was used as positive control. The stems of Picrasma quassioides (D.DON) BENN were collected in August 2016 at Ganzhou, Jiangxi Province, People’s Republic of China, 2.3.2. Building of AD animal model and drug treatment and identified by Prof. Feng Feng. A voucher specimen (No. Piqu- After a week of adaptive feeding, all mice were divided randomly 2016GEY-A) was deposited in the Department of Natural Medicinal into 6 groups of 8 each: control group (CT) and model group (Aβ) with Chemistry, China Pharmaceutical University, China. 0.5% CMC-Na, 1.3 mg/kg Donepezil group (Dpz), 25 mg/kg EA group (Low), 50 mg/kg EA group (Medium) and 100 mg/kg EA group (High). 2.2. Active fraction screening After mice were anesthetized fixed on a stereotactic frame (SR-5,

Narishige, Tokyo, Japan), 4 μL saline with or without Aβ25–35 (8 μg/ 2.2.1. Sample extraction and fractionation mouse) was infused bilaterally into the left and right hippocampus by a The dry Picrasma quassioides stems (20 kg) was extracted two times glass micropipette with a microinjection pump (Dakumar machinery, with 100 L 80% EtOH by refluxing for 2 h, and the aq. EtOHwas Sweden) at a site 2.0 mm caudal to the bregma, 1.5 mm from the combined and concentrated in vacuum at 60 °C to get the crude extract midline, and 2.0 mm below the dural surface [16]. Then, the heads (290 g). The extract was suspended in water, and then partitioned with were sterilized and stitched. Three days later, mice started to receive ethyl acetate and n-butanol to obtain ethyl acetate extract (64 g, EA), n- 0.5% CMC-Na, Donepezil (1.3 mg/Kg) and EA (25, 50, or 100 mg/kg) butanol extract (78 g, NB) and water extract (130 g, WT), respectively. by oral administration once daily, respectively. Brief experimental de- sign is demonstrated as Fig. 1. 2.2.2. Reagents and chemicals preparation Dulbecco’s modified Eagle’s medium (DMEM) and fatal bovine 2.3.3. Morris water maze (MWM) task serum (FBS) were purchased from KeyGEN Biotech Co., Ltd. (Nanjing, Spatial memory of mice was estimated using the MWM task [16]. China). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide The entire experiment comprised 2-days visible platform training, three (MTT) was obtained from Sigma-Aldrich Co., (St. Louis, MO, USA). days invisible platform training, and a day probe trial. Mice were Aβ25-35 (Aladdin, Shanghai, China) was dissolved to the concentration trained one by one in a 60 cm radius and 50 cm height rounded pool of 2 mg/mL using saline, and aggregated by incubating at 37 °C for with a depth of 30 cm with water. A platform (4.5 cm radius) was in- 7 days, and then stored at −20 °C. L-glutamate was purchased from stalled in the middle of one quadrant of the pool. Each mouse was Huixing Biochemical Reagent Co., Ltd. (Shanghai, China). trained in both visible-platform (1–2 day) and hidden-platform (3–5 day) versions, with the visible-platform training undertook to rule 2.2.3. Cell culture out differences in vision between individuals, and the hidden platform PC12 and SH-SY5Y cell lines were obtained from Chinese Academy version was utilized to evaluate spatial learning. Visible-platform in- of Sciences cell bank in Shanghai, China. Two cell lines were cultured in dicated by a 5 cm height flag placed 1 cm below the surface ofthe DMEM with 10% FBS, 100 U/·mL of penicillin and 0.1 mg/mL of water, but the invisible platform didn't have flag. The mice were sub- streptomycin in a suitable environment of 95% air with 5% CO2 at jected to four trials daily with a 1-h interval in different trials. Each trial 37 °C. lasted for 90 s unless the animal arrived in the platform early. The mice,

2 E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258

I.C.V. injection of distilled EA treatment by oral gavage water or Aβ25-35

0 3 17 23 Day Morris-water maze test

HE staining, IL-1β, TNF-a, IL-6, and Aβ1–42 assay

Fig. 1. The animal experimental schedule.

which didn’t find the platform within 90 s, were gently directed tothe by silica gel CC eluted with MeOH-H2O (10:90–100:0) to give four platform for 10 s. On day 6, the platform was removed away followed subfractions (Fr.B1–B5). Fr.B4 was further subdivided by semi-pre- by the probe trial to let the mouse find the platform within 90 s. The parative HPLC to give compound 10 (3.1 mg). Fr.C repeatedly purified number of target crossings and the time spent in the target quadrant by Sephadex LH-20 CC (eluted with CH2Cl2-MeOH = 1:1) to yield with the platform were recorded. All experimental data, including es- compounds 1 (30.5 mg) and 2 (2.5 mg). Fr.D was purified by silica gel cape latency, the number of target crossings and time spent in the target CC and recrystallized to afford compounds 3 (20.3 mg), 4 (6.2 mg), 6 quadrant, were obtained by the video tracking equipment and handled (20.5 mg), 7 (1600 mg), 8 (2600 mg), and 13 (200 mg). Fr.D was sub- with computer equipped with an analysis-management system (Viewer jected to silica gel CC, Sephadex LH-20 (CH2Cl2-MeOH = 1:1), ODS 2 Tracking Software, Ji Liang Instruments, China). (MeOH-H2O = 4:6-8:2) and semi-preparative HPLC to afford 5 (4.5 mg), 9 (6.8 mg), 11 (3.6 mg), 12 (5.2 mg), 14 (4.6 mg), 15 2.3.4. Brain tissues preparation (7.8 mg). Fr.E was subjected to silica gel CC, Sephadex LH-20 CC After Morris water maze test, all the mice were sacrificed and the (eluted with CH2Cl2-MeOH = 1:1) and ODS silica gel CC (eluted with brains were collected quickly. In each group, three complete brains H2O-MeOH) to yield compounds 16 (9.7 mg), and 17 (7.8 mg). were fixed in 4% Formaldehyde Solution for histological examination, and the rest brains were carefully dissected to obtain hippocampus for 2.5. Neuroprotective activity analysis of compounds 1–18 enzyme-linked immunosorbent assay (ELISA) and stored at 80 °C. The neuroprotective activities of compounds 1–18 (10 μM, 20 μM, 2.3.5. Histopathological examination and 40 μM, respectively) were evaluated using 10 mM L-glutamate sti-

After fixed, dehydrated, and decalcification, the brains were cutinto mulated PC12 and 5 μM Aβ25-35 stimulated SH-SY5Y cell models. transverse sections (5 µm) with a microtome and stained with hema- toxylin and eosin. Then, the prepared sections were examined using an optical microscope [21]. 2.6. Molecular modeling studies

The docking study was performed by CDOCKER module im- 2.3.6. ELISA assay kits plemented in Discovery Studio 3.0 [19] to clarify interaction com- The levels of IL-1β, TNF-α, IL-6, and Aβ1-42 in hippocampus were pounds 2 and 3 with DYPKIA and butyrylcholinesterase (BuChE). determined using ELISA assay kits following the operational instruc- tions. 2.7. Statistical analysis 2.4. Extraction and isolation All data were exhibited as mean ± SEMs from independent ex- EA (56 g) was subjected to isolate by silica gel CC eluted with periments in vitro and in vivo. Statistical analysis was implemented

CH2Cl2-MeOH (50:1–50:50) to obtain five fractions (Fr.A – E). Fr.A was using GraphPad Prism 6.0 software (GraphPad Software, Inc., San applied to silica gel CC, eluted with PE-EtOAc (50:1–1:1), the main Diego, CA). For statistical comparisons, the data were analyzed using subfraction was further purified using Sephadex LH-20 CC (eluted with one-way analysis of variance (ANOVA) followed by Dunnett's post-hoc

CH2Cl2-MeOH = 1:1) to yield compound 18 (5.4 mg). Fr.B was purified test.

Fig. 2. Cell proliferation rates of the different extracts in different cells models: (A) L-glutamate-stimulated PC12 cell models; (B) Aβ25-35-stimulated SH-SY5Y cell models. Values shown are expressed as mean ± SEMs; n = 3. ***p < 0.001 vs. blank group. #p < 0.05, ##p < 0.01, ###p < 0.001 vs. model group.

3 E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258

Fig. 3. The effect of EA on spatial memory impairment induced25–35 byAβ in mice was evaluated using MWM test. (A) Latencies to find a hidden platform in the water maze during visible platform two days training and invisible platform three days training. (B) representative swim paths during the spatial probe test; (C) the percentage of total time in the target quadrant during the probe trial test; (D) the number of target crossings during the probe trial test. Values shown are expressed as mean ± SEMs; n = 8. *p < 0.05, **p < 0.01 vs. CT group. #p < 0.05 vs. 0 mg/kg EA-treated Aβ group.

3. Results hippocampal region and cortex in different groups showed remarkable differences. The neurons of CT group were arranged in an organized 3.1. Neuroprotection of the different extracts in vitro pattern with clear boundaries (Fig. 4A). By contrast, nuclear pyknosis and karyolysis were observed, as well as the neurons arranged in an As present in Fig. 2, the cell proliferation rates both on PC12 and irregular pattern in the hippocampal region and cortex in Aβ group SH-SY5Y cells decreased significantly after treatment with L-glutamate (Fig. 4B). After treatment with EA and donepezil, the number of neu- and Aβ25-35. However, the phenomenon had been improved in different rons increased and the boundaries were clearer with an improved ar- extents after treatment with EA, NB, and WT. Especially, EA showed the rangement pattern as compared to untreated mice. And the EA treat- most potential treatment effect with the cell proliferation rates of ment groups showed similar characteristic to CT group, even better 58.11%-129.47% and 6.50%-78.63% on PC12 and SH-SY5Y cells, re- than the donepezil group (Fig. 4C–F). These suggested that EA could spectively. Thus, the EA was chosen for further investigation. reduce Aβ25–35-induced neuronal damage.

3.2. Anti-AD effects of EA in25-35 Aβ -induced mouse model 3.2.3. Effect of EA on the IL-1β, TNF-α, IL-6, andAβ1-42 levels in the hippocampus of Aβ25-35-injected mice 3.2.1. Morris water maze test The levels of IL-1β, TNF-α, IL-6, and Aβ1-42 in the hippocampus of MWM test was employed to assess the effect of EA on memory Aβ25-35-treated mice were significantly higher than those of theCT dysfunction in AD mice. As shown in Fig. 3A, there was no significant group (***p < 0.001, Fig. 5). After treatment with EA, the alterations difference between the six groups in the initial visible platform tests were significantly attenuated as shown in Fig. 5 (###p < 0.001). (p > 0.05). On the third day of the hidden platform test, Low group # found the hidden-platform faster than Aβ group ( p < 0.05; Fig. 3A). 3.3. The isolated compounds During the space exploration test on the sixth day, the mice in CT group and Low group spent more time in the target quadrant than Aβ group EA was fractionated by several chromatographic techniques to yield # (*p < 0.05, p < 0.05; Fig. 3B–C). The number of crossing the plat- 18 known compounds. By comparing their 1D NMR and MS data with # form of Low group was more than Aβ group ( p < 0.05; Fig. 3D). previous literatures, these compounds were identified as six β-carboline These results showed EA reversed the spatial memory deficits of the alkaloids, 1-hydroxymethyl-β-carboline (1) [22], 1,6-dihydroxy-β-car- Aβ25-35-induced AD mice. boline (2) [23], 1-formly-β-carboline (3) [24], β-carboline-1-carboxylic acid (4) [25], β-carboline-1-carboxamide (5) [26], 4-hydroxy-1-meth- 3.2.2. Results of histological examinations oxycarbonyl-β-carboline (6) [23], seven canthin-6-one alkaloids, 4,5- As demonstrated by HE staining, the neuronal structures of the dimethoxy-canthin-6-one (7) [27], 4-methoxy-5-hydroxyl-canthin-6-

4 E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258

Fig. 4. Effect of EA on the histopathological changes in hippocampus region and cortex: CT group (A); Aβ group (B); Low group (C); Medium group (D);Highgroup (E); Dpz group (F). one (8) [27], 9-hydroxy-canthin-6-one (9) [28], 4,5-dimethoxy-10-hy- using MTT method. As shown in Table 1, four β-carbolines (1–3, and 6) droxy-canthin-6-one (10) [29], 5-hydroxycanthin-6-one (11) [30], and six canthin-6-ones (7–11, and 13) exhibited neuroprotective ac- canthin-6-one (12) [31], 3-methyl canthin-5,6-dione (13) [27], five tivities. Especially, compounds 2 and 3 exhibited markedly neuropro- quassinoids, nigakilactone F (14) [32], nigakilactone B (15) [33], ni- tection with the cell proliferation rate of 33.74%–48.8% and gakilactone Q (16) [5], nigakilactone I (17) [34], desbenzoyl-picraja- 1.11%–82.1% at the dosage ranging from 10 μM to 40 μM in above cell vanin A (18) [35] (Fig. 6). models, respectively.

3.4. Neuroprotecition of compounds 1–18 in L-glutamate-stimulated PC12 3.5. Molecular docking of studies and Aβ25-35 stimulated SH-SY5Y cells To reveal the potential targets, compounds 2 and 3 molecular The effects of all isolates on proliferation rateof L-glutamate-sti- docking studies were performed on two key targets, DYRK1A and mulated PC12 and Aβ25-35 stimulated SH-SY5Y cells were determined BuChE, in AD. As shown in Fig. 7 and Fig. 8, compounds 2 and 3 could

5 E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258

Fig. 5. Effect of EA on the IL-1β (A), TNF-α (B), IL-6 1-42(C) andAβ (D) levels in the hippocampus of Aβ25-35-injected mice. Each value is expressed as mean ± SEMs (n = 4). ***p < 0.001 vs. CT group. ###p < 0.001 vs. 0 mg/kg EA-treated Aβ group.

Fig. 6. Structures of compounds 1–18.

6 E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258

Table 1 a The proliferation rate of L-glutamate-stimulated PC12 and Aβ25-35-stimulated SH-SY5Y cells after treatment with compounds 1-18.

Compounds L-glutamate-stimulated PC12 (%) Aβ25-35-stimulated SH-SY5Y (%)

10 μM 20 μM 40 μM 10 μM 20 μM 40 μM

Blank 100.00 ± 2.36 100.00 ± 5.87 Model 0 ± 3.24 0 ± 3.57 1 21.85 ± 5.13 30.29 ± 3.07 43.24 ± 5.26 11.55 ± 2.43 9.7 ± 1.77 8.53 ± 4.59 2 33.74 ± 7.38 36.64 ± 5.31 41.28 ± 2.34 35.46 ± 7.47 39.3 ± 9.95 48.8 ± 4.71 3 55.7 ± 11.64 82.1 ± 3.7 1.11 ± 6.7 47.82 ± 4.6 51.98 ± 2.66 5.15 ± 1.77 6 3.82 ± 4.99 7.34 ± 2.46 6.34 ± 1.07 19.89 ± 9.83 18.81 ± 1.52 19.49 ± 1.32 7 35.63 ± 0.65 39.94 ± 7.33 56.25 ± 3.23 12.72 ± 4.58 5.14 ± 0.33 1.06 ± 1.53 8 –b 6.3 ± 6.17 23.48 ± 2.3 6.66 ± 4.09 –b –b 9 8.74 ± 2.46 9.18 ± 0.82 25.27 ± 2.25 7.29 ± 5.62 –b –b 10 18.87 ± 1.27 19.93 ± 2.12 20.45 ± 2.63 12.67 ± 2.7 10.81 ± 5.62 24.85 ± 3.37 11 12.66 ± 1.43 26.71 ± 10.86 35.22 ± 7.46 8.18 ± 2.64 2.48 ± 4.59 –b 13 –b 3.24 ± 2.03 6.97 ± 3.82 15.69 ± 1.86 23.6 ± 4.24 21.21 ± 7.29

a Compounds 4, 5, 12, and 14–18 were inactive on two cell models. b Not active.

bind to the ligand bonding pockets for DYRK1A and BuChE, as well as neuroinflammatory, and reduced Aβ1-42 deposition in AD mice. These formation of Van der waals, and conventional hydrogen bond with suggested that EA could be a potential and valuable anti-AD agent. protein residues. Beside, pi – pi T – shaped and amide – pi staked were Despite the pathogenesis of AD is still ambiguous, but it is widely also observed between the molecules (2 and 3) and DYRK1A. The re- accepted that neurofibrillary tangles and Aβ deposition are the mainly sults showed that compounds 2 and 3 have high affinity for both of the neuropathological hallmarks in AD [36]. Intensive researches have re- above proteins. vealed that Aβ deposits and extracellular tangles can trigger neuroin- flammatory response [37], subsequently resulting in generation of pro- 4. Discussion inflammatory mediators include cytokines, such as IL-1β, IL-6, and TNF-α [38]. The excessive secretion of these pro-inflammatory med- AD, the most common neurodegenerative disease, has attracted iators may trigger signaling cascades in neurons, and causing neuron interest from pathologist and pharmacologist because of its high in- damage and enhancing AChE activity, ultimately leading to AD [39]. cidence and risk [11]. Unfortunately, there is no effective drug and Meanwhile, overproduction of inflammatory cytokines can activate β- treatment strategy for this disease. In the present study, we found that secretase-1 (BACE1) and γ-secretases to reduce Aβ clearance, thereby

EA significantly reduced L-glutamate and Aβ25-35 induced neuronal cells promoting Aβ production [40]. It is note that taking nonsteroidal anti- damage. Meanwhile, EA significantly ameliorated cognitive function, inflammatory drugs (NSAIDs) can reduce the risk ofAD [41]. P. quas- protected against neuronal loss responses, suppressed sioides is a good anti-inflammatory traditional Chinses medicine, but its

Fig. 7. 3D Binding mode prediction of compounds 2 (A) and 3 (B) with DYRK1A (PDB id: 3ANR) and compounds 2 (C) and 3 (D) with BuChE (PDB id: 4TPK).

7 E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258

Fig. 8. 2D Binding mode prediction of compounds 2 (A) and 3 (B) with DYRK1A (PDB id: 3ANR) and compounds 2 (C) and 3 (D) with BuChE (PDB id: 4TPK).

anti-AD activity has never been reported. Our results showed that EA neuroprotective activities roughly defined some structure-activity re- could significantly reduce the IL-1β, TNF-α, IL-6, andAβ1-42 levels in lationships. Firstly, both the β-carboline type and canthinone-6-one the hippocampus of AD mice, which suggested that EA could reverse type showed the potential neuroprotective activities, which suggested pro-inflammatory mediators increase by Aβ25-35 stimulating. Therefore, that the alkaloid skeleton might not be crucial for their structure-ac- we believe that suppressing neuroinflammatory and reduce Aβ1-42 de- tivity. Moreover, the types of functional groups in β-carboline alkaloids position could be involved in the mechanisms of anti-AD of EA. have the important influence on the activity. For example, the activities

Previous reports have verified that the main chemical constituents of those compounds with weak polarity groups (1 and 2) at R1 were of P. quassioides are alkaloids, and quassinoids, and alkaloids are the more significant than which with strong polarity groups (4 and 5) at R1 major anti-inflammatory components [4,5]. Moreover, some canthi- (Fig. 6). Base on this conclusion, comparison the active compound 2 none and β-carboline alkaloids have been reported to possess potential and the inactive compounds (4 and 5) revealed that hydroxyl group at neuroprotectvie activities in recent years [8,42,43]. Such as, picrasidine R2 might increase the activity. However, hydroxyl group at R3 might O, isolated from P. quassioides, possessed an extent cerebral protective decrease the activity when comparison the active compound (1 and 3) effect through boosting learning and memory performance, reducing and the inactive compound (6). neuronal cells death, and suppressing neuronal hyperexcitability [8]. To reveal the potential targets of the active compounds, molecular Harmine, a natural β-carboline alkaloid, have also been described as docking studies were performed. As mentioned above, DYRK1A and multifunctional agent for the treatment of Alzheimer's disease including BuChE are important targets for AD treatment, and several β-carboline interfering neuroinflammation, upregulating glutamate transporter ex- alkaloids were reported as the ligands of these two proteins [7,42,43]. pression, inhibiting DYRK1A, tau phosphorylation and AChE [42,43]. Therefore, DYRK1A and BuChE were employed in the molecular Harmaline can effectively ameliorated memory impairments in asco- docking study. Our results showed that compounds 2 and 3 have high polamine-induced mouse model via improvement of cholinergic system affinity for both of the above proteins, which suggested DYRK1Aand function, suppression of oxidative stress and inflammation damage, and BuChE could be the potential targets for β-carboline alkaloids and β- modulation of vital neurotransmitters in dementia mice [7]. But, the carboline alkaloids could be potential multi-target anti-AD agents. neuroprotective and anti-inflammatory activities of quassinoids were rarely reported. In this work, we isolated 13 alkaloids (1–13), and five 5. Conclusion quassinoids (14–18) from EA. Interestingly, only the alkaloids exhibited neuroprotective activities in vitro (Table 1), which suggested that al- In summary, EA is a potential multi-target anti-AD agent by sup- kaloids could be the main anti-AD constituents of P. quassioides. Be- pressing neuroinflammation and reducing1-42 Aβ deposition. sides, the structural diversity of these alkaloids and their Furthermore, 13 alkaloids (1–13) and five quassinoids (14–18) were

8 E. Guo, et al. Bioorganic Chemistry 92 (2019) 103258 isolated from EA. Among them, alkaloids were proved to be the mainly through the inhibition of TLR-4/NF-kappaB signaling in mice, J. Ethnopharmacol. anti-AD active components, and DYRK1A and BuChE could be the po- 191 (2016) 398–407. [17] P. Eikelenboom, C. Bate, W.A.V. Gool, J.J.M. Hoozemans, J.M. Rozemuller, tential targets for β-carbolines. Besides, the structure-activity relation- R. Veerhuis, A. Williams, Neuroinflammation in Alzheimer's disease and prion ships of alkaloids were discussed in this paper. disease, Glia 40 (2002) 232–239. [18] T. Zhao, K.-M. Ding, L. Zhang, X.-M. Cheng, C.-H. Wang, Z.-T. Wang, Acetylcholinesterase and Butyrylcholinesterase inhibitory activities of β-carboline Acknowledgement and quinoline alkaloids derivatives from the of GenusPeganum, J. Chem. 2013 (2013) 1–6. This work was supported by the National Natural Science [19] Y. Zhao, F. Ye, J. Xu, Q. Liao, L. Chen, W. Zhang, H. Sun, W. Liu, F. Feng, W. Qu, Foundation Youth Project of China [grant number 8170332], the China Design, synthesis and evaluation of novel bivalent beta-carboline derivatives as multifunctional agents for the treatment of Alzheimer's disease, Bioorg. Med. Chem. Postdoctoral Science Foundation (2017M621888). 26 (2018) 3812–3824. [20] R. León, A.G. Garcia, J. Marcocontelles, Recent advances in the multitarget-directed Declaration of Competing Interest ligands approach for the treatment of Alzheimer's disease, Med. Res. Rev. 33 (2013) 139–189. [21] A. Liu, X. Zhao, H. Li, Z. Liu, B. Liu, X. Mao, L. Guo, K. Bi, Y. Jia, 5- The authors declare no competing financial interest. Hydroxymethylfurfural, an antioxidant agent from Alpinia oxyphylla Miq. improves cognitive impairment in Aβ1–42 mouse model of Alzheimer's disease, Int. Immunopharmacol. 23 (2014) 719–725. Appendix A. Supplementary material [22] Y. Kondo, T. Takemoto, The structure of a new β-carboline alkaloid from Picrasma ailanthoides PLANCHON, Chem. Pharm. Bull. 21 (1973) 837–839. Supplementary data to this article can be found online at https:// [23] W.H. Jiao, H. Gao, C.Y. Li, G.X. Zhou, S. Kitanaka, A. Ohmura, X.S. Yao, β-carboline alkaloids from the stems of Picrasma quassioides, Magn. Reson. Chem. MRC 48 doi.org/10.1016/j.bioorg.2019.103258. (2010) 490–495. [24] Y. Yang, S. Luo, X. Shen, Y. Li, Chemical investigation of the alkaloids of Ku-Mu References [Picrasma quassioides (D. Don) Benn.] (author's transl), Yao Xue Xue Bao 14 (1979) 167–177. [25] Z.Q. Lai, W.H. Liu, S.P. Ip, H.J. Liao, Y.Y. Yi, Z. Qin, X.P. Lai, Z.R. Su, Z.X. Lin, Seven [1] J.W. Lee, J.W. Park, N.R. Shin, S.Y. Park, O.K. Kwon, H.A. Park, Y. Lim, H.W. Ryu, alkaloids from Picrasma quassioides and their cytotoxic activities, Chem. Nat. H.J. Yuk, J.H. Kim, Picrasma quassiodes (D. Don) Benn. attenuates lipopoly- Compd. 50 (2014) 884–888. saccharide (LPS)-induced acute lung injury, Int. J. Mol. Med. 38 (2016) 834. [26] P. Forgacs, J. Provost, A. Touché, Indole alkaloids of Odyendea gabonensis, Planta [2] W. Zhao, C. Sun, J. He, L. Chen, Y. Zhang, W. Sun, The possible mechanisms of Med. 46 (1982) 187–189. Picrasma quassiodes (D. Don) Benn. in the treatment of colitis induced by 2,4,6- [27] T. Ohmoto, K. Koike, Studies on the constituents of Picrasma quassioides BENNET. trinitrobenzene sulfonic acid in mice, J. Ethnopharmacol. 145 (2013) 424–430. III. The alkaloidal constituents, Chem. Pharm. Bull. 32 (1984) 3579–3583. [3] Y. Niiho, K. Mitsunaga, K. Koike, T. Ohmoto, Studies on the gastric antiulcer [28] Y. Sung, K. Koike, T. Nikaido, T. Ohmoto, U. Sankawa, Inhibitors of cyclic AMP components from the of Picrasma quassioides (Simaroubaceae), Nat. Med. 48 phosphodiesterase in Picrasma quassioides BENNET, and inhibitory activities of re- (1994) 116–121. lated β-carboline alkaloids, Chem. Pharm. Bull. 32 (1984) 1872–1877. [4] Q. Lin, J. Xu, F. Feng, Chemical constituents from the stems of Picrasma quassioides [29] M. Jiang, Y. Zhou, Canthin-6-one alkaloids from Picrasma quassioides and their Bennet, J. China Pharmaceut. Univ. 48 (2017) 675–679. cytotoxic activity, J. Asian Nat. Prod. Res. 10 (2008) 1009–1012. [5] J. Xu, D. Xiao, W.W. Song, L. Chen, W.Y. Liu, N. Xie, F. Feng, W. Qu, Quassinoids [30] E.V. Lassak, J. Polonsky, H. Jacquemin, 5-Hydroxycanthin-6-one from Simarouba from the stems of Picrasma quassioides and their cytotoxic and NO production-in- amara, Phytochemistry 16 (1977) 1126–1127. hibitory activities, Fitoterapia 110 (2016) 13–19. [31] T. Ohmoto, K. Koike, Studies on the constituents of Picrasma quassioides BENNET. II. [6] W. Horton, A. Sood, S. Peerannawar, N. Kugyela, A. Kulkarni, R. Tulsan, C.D. Tran, On the alkaloidal constituents, Chem. Pharm. Bull. 31 (1983) 3198–3204. J. Soule, H. LeVine 3rd, B. Torok, M. Torok, Synthesis and application of β-carbo- [32] T. Murae, T. Tsuyuki, T. Nishihama, S. Masuda, T. Takahashi, Bitter principles of lines as novel multi-functional anti-Alzheimer's disease agents, Bioorg. Med. Chem. picrasma ailanthoides planchon, Tetrahedron Lett. 10 (1969) 219–226. Lett. 27 (2017) 232–236. [33] K. Koike, M. Yokoh, M. Furukawa, S. Ishii, T. Ohmoto, Picrasane quassinoids from [7] S.P. Li, Y.W. Wang, S.L. Qi, Y.P. Zhang, G. Deng, W.Z. Ding, C. Ma, Q.Y. Lin, , Phytochemistry 40 (1995) 233–238. H.D. Guan, W. Liu, X.M. Cheng, C.H. Wang, Analogous β-Carboline alkaloids har- [34] T. Sadaaki, F. Narihiko, O. Masayoshi, K. Junko, K. Kazuo, T. Harukuni, A. Wataru, maline and harmine ameliorate scopolamine-induced cognition dysfunction by at- T. Junko, K. Masashi, N. Hoyoku, Three new quassinoids, ailantinol E, F, and G, tenuating acetylcholinesterase activity, oxidative stress, and inflammation in mice, from Ailanthus altissima, Cheminform 34 (2003) 385–389. Front. Pharmacol. 9 (2018) 346. [35] M. Yoshikawa, E. Harada, S. Aoki, J. Yamahara, N. Murakami, H. Shibuya, I. [8] T. Sasaki, W. Li, T. Ohmoto, K. Koike, Evaluation of canthinone alkaloids as cerebral Kitagawa, Indonesian Medicinal Plants. VI. On the chemical constituents of the bark protective agents, Bioorg. Med. Chem. Lett. 26 (2016) 4992–4995. of Picrasma javanica BL. (Simaroubaceae) from Flores Island. Absolute [9] Z. Xu, S. Chen, Y. Wang, S. Chen, W. Yao, X. Gao, Neuroprotective effects of silk Stereostructures of Picrajavanins A and B, Cheminform 41 (1993) 233-238. fibroin hydrolysate against Aβ25–35 induced cytotoxicity in SH-SY5Y cells and pri- [36] J.M. Rubio-Perez, J.M. Morillas-Ruiz, A review: inflammatory process in mary hippocampal neurons by regulating ROS inactivation of PP2A, J. Funct. Foods Alzheimer's disease, role of cytokines, Sci. World J. 2012 (2012) 756357. 45 (2018) 100–109. [37] Y. Shen, L. Lue, L. Yang, A. Roher, Y. Kuo, R. Strohmeyer, W.J. Goux, V. Lee, [10] J. Chen, L.-X. Tao, W. Xiao, S.-S. Ji, J.-R. Wang, X.-W. Li, H.-Y. Zhang, Y.-W. Guo, G.V. Johnson, S.D. Webster, Complement activation by neurofibrillary tangles in Design, synthesis and biological evaluation of novel chiral oxazino-indoles as po- Alzheimer's disease, Neurosci. Lett. 305 (2001) 165–168. tential and selective neuroprotective agents against Aβ25–35-induced neuronal da- [38] Z. Hui, W. Su-Lei, Q. Lai, J. Jia-Li, L. Hui, X. Yun, Z. Xiao-Lei, Diammonium gly- mage, Bioorg. Med. Chem. Lett. 26 (2016) 3765–3769. cyrrhizinate attenuates Aβ(1–42) -induced neuroinflammation and regulates MAPK [11] World Alzheimer Report, 2015. and NF-κB pathways in vitro and in vivo, CNS Neurosci. Ther. 19 (2013) 117–124. [12] C. Lu, Y. Guo, J. Yan, Z. Luo, H.-B. Luo, M. Yan, L. Huang, X. Li, Design, synthesis, [39] W.S.T. Griffin, L. Liu, Y. Li, R.E. Mrak, S.W. Barger, Interleukin-1 mediates and evaluation of multitarget-directed resveratrol derivatives for the treatment of Alzheimer and Lewy body pathologies, J. Neuroinflammation 3 (2006) 5. Alzheimer’s disease, J. Med. Chem. 56 (2013) 5843–5859. [40] S.E. Hickman, E.K. Allison, E.K. Joseph, Microglial dysfunction and defective beta- [13] B.C.-K. Tong, A.J. Wu, M. Li, K.-H. Cheung, Calcium signaling in Alzheimer's disease amyloid clearance pathways in aging Alzheimer's disease mice, J. Neurosci. 28 & therapies, Biochimica et Biophysica Acta (BBA) – molecular, Cell Res. 1865 (2008) 8354–8360. (2018) 1745–1760. [41] P.L. Mcgeer, J. Rogers, E.G. Mcgeer, Inflammation, anti-inflammatory agents and [14] B. Smith, F. Medda, V. Gokhale, T. Dunckley, C. Hulme, Recent advances in the Alzheimer disease: the last 12 years, J. Alzheimers Disease Jad 9 (2006) 271. design, synthesis, and biological evaluation of selective DYRK1A inhibitors: a new [42] N. Gockler, G. Jofre, C. Papadopoulos, U. Soppa, F.J. Tejedor, W. Becker, Harmine avenue for a disease modifying treatment of Alzheimer’s? ACS Chem. Neurosci. 3 specifically inhibits protein kinase DYRK1A and interferes with neurite formation, (2012) 857–872. FEBS J. 276 (2009) 6324–6337. [15] A. Nazem, R. Sankowski, M. Bacher, Y. Al-Abed, Rodent models of neuroin- [43] Y. Li, R. Sattler, E.J. Yang, A. Nunes, Y. Ayukawa, S. Akhtar, G. Ji, P.W. Zhang, flammation for Alzheimer’s disease, J. Neuroinflammation 12 (2015) 1–15. J.D. Rothstein, Harmine, a natural β-carboline alkaloid, upregulates astroglial [16] L. Chen, L. Hu, J. Zhao, H. Hong, F. Feng, W. Qu, W. Liu, Chotosan improves Aβ1-42- glutamate transporter expression, Neuropharmacology 60 (2011) 1168–1175. induced cognitive impairment and neuroinflammatory and apoptotic responses

9