The four herbs in KXS (RS, FL, SCP, and YZ) were obtained from the First Affiliated Hospital of Guangzhou University of Chinese Medicine (Guangzhou, China) and were mixed at a ratio of 3:3:2:2 referenced the previous studies (Cao et al., 2012; Xu et al., 2019). The extraction process and quality control were implemented according to our previous study (Xu et al., 2019). Briefly, SCP was added to six parts of water as a solvent for 2 h, followed by heat reflux extraction for 8 h. Volatile oil and SCP dregs were saved. RS was extracted twice by 60% ethanol as solvent for each time with 1 hour. The extracts were combined and filtrated, and the RS dregs were saved. The dregs of SCP, RS, FL, and YZ were added to 10 parts of water and extracted twice, for 1 h each. All extracts were combined and evaporated on a rotary evaporator. Finally, volatile oil of SCP and original liquid after concentrated were mixed and refrigerated at −20°C for usage.

Kun-Ming mice (8 weeks old) weighing 30–35 g were obtained from Sibeifu Biotechnology (Beijing, China) and housed at 22 ± 2°C, with a relative humidity of 55% ± 5%, a 12 h light/dark cycle and ad libitum access to food and water. All animal procedures were performed in accordance with the principles and guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Guangzhou University of Chinese Medicine Animal Ethics Committee.

AD model was induced by intraperitoneally injected SCOP (3 mg/kg) for a consecutively week. In this week, Morris water maze and new objection recognition test were performed daily after 30 min of SCOP injection to evaluate animal model. In this study: mice were randomly allotted to five groups: Control group (n=10; 0.9% saline p.o.+ 0.9% saline, i.p.), SCOP group (n=10; 0.9% saline p.o.+ SCOP 3 mg/kg/d i.p.), low-dose KXS group (n=10; KXS 1.4 g/kg/d p.o.+ SCOP 3 mg/kg/d i.p.), high-dose KXS group (n=10; KXS 2.8 g/kg/d p.o.+ SCOP 3 mg/kg/d i.p.), and Don group (n=10; donepezil 3 mg/kg/d p.o.+ SCOP 3 mg/kg/d i.p.). The mice were orally administered 0.9% saline, KXS, or donepezil mice for 14 continuous days. Memory impairment was induced by SCOP treatment (3 mg/kg body weight) for 7 days.

The hippocampus and cortex were homogenized in SDS lysis buffer containing a protease inhibitor and phosphatase inhibitor mixture (Sigma-Aldrich). After sonication, lysates were centrifuged at 3,000×g and 4°C for 15 min. Supernatants were collected, and protein concentrations were determined by the Bradford assay (Bio-Rad). The same amounts of proteins were resolved on SDS-PAGE gels, transferred to PVDF membranes (Millipore), and probed with primary antibodies overnight at 4°C. Rabbit anti-TNFRSF1A (1:1,000, Abcam), anti-CHRNB2 (1:1,000, Abcam), and anti-β-actin (1:10,000, Abcam) were used as the primary antibodies. Immunoblots were visualized by the ECL western blot detection kit (Millipore) and quantified by densitometry and ImageJ software (National Institutes of Health) (Lee et al., 2017).

Mice were transcardially perfused with 0.9% saline. The mouse brains were fixed in 4% paraformaldehyde (PFA) for 72 h, followed by dehydration with different concentrations of ethanol and paraffin embedding. A paraffin microtome was used to cut 5-μm-thick coronal brain sections.

For immunofluorescence, sections were deparaffinized by ethanol and incubated with 3% H₂O₂. Antigen retrieval was performed by heating the sections in sodium citrate buffer (10 mM trisodium citrate, 0.5% Tween-20 in H₂O, pH 6.0) at 70°C for 30 min. Sections were then permeabilized and incubated with blocking solution containing the Iba-1 antibody (1:100, Abcam) overnight at 4°C. The sections were then stained with a fluorophore-conjugated secondary antibody (1:1,000, CST) for 1 h, followed by 4′6-diamino- 2-phenylindole (DAPI, Sigma) staining for another 1 h. Images were acquired by a fluorescence microscope Model DMi8 (Leica, Germany) and quantified using ImageJ software (National Institutes of Health) (Muhammad et al., 2019).

Mice cortical and hippocampal tissues were sequentially homogenized in ice-cold PBS, and the supernatants were centrifuged at 21,000×g for 20 min at 4°C. The levels of IL-1β, IL-6, and TNF-α were measured using an enzyme-linked immunosorbent assay (ELISA, Biological Technology, Jiangsu) (Karthivashan et al., 2018) according to the manufacturer's instructions.

All data were expressed as means ± SEM of at least three independent experiments. Statistical analyses were performed using SPSS (version 20.0, IBM, Armonk, NY). Statistical tests between multiple datasets were carried using a one-way analysis of variance (ANOVA) followed by Dunnett's post hoc test to determine statistical significance, as appropriate. A P value < 0.05 was considered statistically significant.

It is well known that abnormal Aβ aggregation and accumulation plays a critical role in the pathogenesis of AD. Amyloid precursor protein (APP) is cleaved by α-secretase, β-secretase, and γ-secretase via non-amyloidogenic and amyloidogenic pathways. In the non-amyloidogenic pathway, proteolysis of APP by α- and γ-secretases yields nonpathogenic fragments, including sAPPα and C-terminal fragments. However, the amyloidogenic pathway involves cleavage by β- and γ-secretases, leading to the production of sAPPβ, C-terminal fragments, and Aβs (Barage and Sonawane, 2015; Selkoe and Hardy, 2016). Accumulated Aβ can further directly or indirectly trigger other pathological processes, such as mitochondrial dysfunction and inflammation, inducing neuronal death in AD (Mao et al., 2012). Thus, targeting β- and γ-secretases is regarded as a primary therapeutic approach for reducing Aβ production in AD patients. As shown in Figure 5, KXS targets key proteins implicated in the Aβ-associated signaling pathway, including BACE1 (β-secretase 1), GAPDH (glycolytic glyceraldehyde-3-phosphate dehydrogenase), and SD17B10 (hydroxysteroid 17-beta dehydrogenase 10). GAPDH (glycolytic glyceraldehyde-3-phosphate dehydrogenase), an abundantly expressed oxidoreductase for glucose metabolism, has been reported to possess diverse non-glycolytic functions, including interaction with Aβ (El Kadmiri et al., 2014; Sen et al., 2018). Target prediction suggested that M693 (citric acid) could bind with GAPDH, which might alleviate Aβ accumulation by disturbing the cross-talk between Aβ and GAPDH (Gerszon and Rodacka, 2018). In addition, HSD17B10 (hydroxysteroid 17-beta dehydrogenase 10), a hydroxysteroid dehydrogenase localized in the mitochondria (He et al., 2018), was the potential target of the compounds in KXS, including M150, M388, and M500. This target has been recognized as a putative intracellular mediator of Aβ neurotoxicity via its dozen proteins that bind to Aβ (Reddy et al., 2012). Thus, KXS could disturb the combination of Aβ and HSD17B10 to alleviate mitochondria dysfunction as it activates β-secretase and γ-secretase and facilitates Aβ generation (Yang et al., 2014). Taken together, KXS might prevent and treat AD by regulating Aβ generation and aggregation.

NFTs (neurofibrillary tangles), aggregating by abnormal hyperphosphorylation of microtubule-associated protein tau (MAPT) in the brain, correlate well with the severity of cognitive deficits in AD patients (Simic et al., 2016; Jouanne et al., 2017). Tau protein can be hyperphosphorylated by various kinases including CDK5 (cyclin-dependent kinase 5) (Zhao et al., 2019) and GSK3B (glycogen synthase kinase 3beta) (Chu et al., 2017). Therefore, a reduction in tau protein levels (Wang and Mandelkow, 2015) as well as inhibition of tau protein hyperphosphorylation and aggregation are potential therapeutic avenues for AD treatment. As shown in Figure 5, KXS was predicted to target three crucial proteins (i.e., CDK5, GSK3B, and MAPT) involved in the regulation of NFTs. For example, the compounds M621, M622, and M909 were predicted to interact with CDK5 and GSK3B. Emerging evidence has revealed that inhibiting these kinases can prevent tau protein hyperphosphorylation and NFT formation (Congdon and Sigurdsson, 2018). Additionally, compounds such as M84, M257, and M686 were identified to bind to MAPT, suggesting that KXS might decrease tau protein levels to alleviate NFT formation and ameliorate Aβ-dependent neurotoxicity (Peters et al., 2019).

The loss of cholinergic neurons in AD patients leads to learning and memory deficits (Ferreira-Vieira et al., 2016). Cholinesterase inhibitors can prevent the breakdown of acetylcholine and preserve its activity at cholinergic synapses (Hampel et al., 2018). Figure 5 shows that the KXS formula regulates the cholinergic system signaling pathway by targeting ACHE (acetylcholinesterase) and CHRN2B (cholinergic receptor nicotinic beta 2 subunit). Four compounds, M21, M389, M597, and M944, interact with ACHE and CHRN2B, which might improve the level of acetylcholine and sustain its activity. Our previous study demonstrated that KXS could ameliorate cognitive function in SCOP-induced mice by regulating the cholinergic system through sustaining ACh levels, increasing ChAT activity, and decreasing AChE activity (Xu et al., 2019), which is in accordance with our systems pharmacology-based analysis.

Neuroinflammation refers to the inflammatory response in the central nervous system secondary to neuronal insult (Calsolaro and Edison, 2016). Activating microglia and astrocytes in AD leads to the production and release of inflammatory cytokines, including interleukin-6 (IL-6), interleukin-1β (IL-1β), and TNF-α (Heneka et al., 2015; Kinney et al., 2018; Ozben and Ozben, 2019). Figure 5 suggests that KXS, a more specifically the compounds M230, M885, and M606, could modulate neuroinflammation in AD by regulating TNF-α and its receptor TNF-receptor superfamily 1A (TNFRSF1A). TNF-α can exacerbate Aβ burden by increasing γ-secretase activity and β-secretase production (Cheng et al., 2014; Chang et al., 2017) or bind to TNFRSF1A. This effect is partly mediated through MAPK1 signaling, which leads to the expression of molecules that participate in inflammation and amyloid genesis (Montgomery and Bowers, 2012; Decourt et al., 2017). Furthermore, IL-1β and IL-6 are able to enhance the amyloidogenic process of APP (Tachida et al., 2008) and the hyperphosphorylation of tau epitopes (Kitazawa et al., 2011). Suppression of either IL-1β or IL-6 can ameliorate neuroinflammation and neurodegeneration (Basu et al., 2004; Rothaug et al., 2016). In summary, KXS might prevent and treat AD by regulating neuroinflammation.

The systems pharmacology analysis indicated that the therapeutic mechanism of KXS for AD involves several potential mechanisms, including the regulation of cholinergic system dysfunction and neuroinflammation.

Chronic neuroinflammation (sustained microglial activation and overproduction of pro-inflammatory cytokines) has been implicated in the pathophysiology of AD. To validate the mechanism of KXS on neuroinflammation in vivo, we measured proinflammatory cytokine levels in SCOP-induced mice treated with KXS. As shown in Figure 6, SCOP administration significantly increased the levels of TNF-α, IL-1β, and IL-6 compared to the non-treated control group (P < 0.05 or P < 0.01). However, KXS and Don significantly (P < 0.05) attenuated the upregulation of IL-1β, TNF-α, and IL-6 compared with the SCOP alone group both in the hippocampus and cortex.

TNFRSF1A is the main cell surface receptor for TNF. Next, we explored the expression of TNFRSF1A by western blotting. As shown in Figure 7, the protein expression of TNFRSF1A increased in the SCOP group in both the hippocampus and cortex (P < 0.01, vs. the control group). Both the low-dose and high-dose KXS groups and the DON wed group decreased TNFRSF1A expression (P < 0.05 or P < 0.01, vs. the SCOP group).

Microglia secrete a number of inflammatory cytokines including TNF-α, IL-1β, and IL-6 in response to inflammatory stimuli (Ozben and Ozben, 2019). To further probe the effect of KXS on neuroinflammation, we detected the expression of Iba-1, a microglia activation marker, in the hippocampus and cortex using immunofluorescence. As displayed in Figure 8, elevated microglia activation was observed in the SCOP group in the hippocampus (CA1) and cortex, whereas KXS and donepezil markedly attenuated microglia activation. These results demonstrate that KXS inhibits neuroinflammation in SCOP-induced mice

In our previous study, we demonstrated that KXS could alleviate cholinergic system dysfunction in SCOP-induced mice by increasing ChAT activity and the Ach content, and decreasing AChE activity (Xu et al., 2019). However, the detailed mechanism remains unclear. In this study, a systems pharmacology-based analysis suggested that KXS could regulate the expression of the target protein CHRNB2 (a nicotinic acetylcholine receptor). CHRNB2 is lost in the brains of AD patients and knockout of the CHRNB2 gene impairs neuronal survival in aging (Cook et al., 2004). Therefore, we further investigated the effect of KXS on cholinergic dysfunction via CHRNB2 expression in SCOP-induced mice. As shown in Figure 9, SCOP significantly downregulated the expression of CHRNB2 compared to the control group in both the hippocampus and cortex (P < 0.05, P < 0.01, respectively). High-dose KXS and Donepezil remarkably increased the expression of CHRNB2 both in the hippocampus and cortex (P < 0.05, vs. the SCOP group). These data suggest that CHRNB2 may play an important role in the effects of KXS ameliorating cholinergic system dysfunction of SCOP-induced mice.