A network pharmacological approach to investigate the pharmacological effects of CZ2HF decoction on Alzheimer's disease

Introduction

Alzheimer's disease (AD) is a progressive neuro-degenerative disorder, which was characterized as learning and memory impairment and disturbance of behavior in clinical due to amyloid beta (Aβ) plaques accumulation and neurofibrillary tangles formation (Voglein J, et al., 2019). Currently, approximately 35 million people are suffering with AD that brings heavy burden to sufferer and society (Molinuevo JL, et al., 2018). Unfortunately, up to now, there are no ideal tactics available and multiple candidate agents fail in late-stage clinical trials owing to the complex aetiology and pathophysiology of AD (Fang J, et al., 2017). Therefore, it is a dire clinical demand to explore valid drug for prophylaxis and treatment of AD.

Traditional Chinese medicine (TCM) has finds substantial application in treating multiple diseases in Asia (Liu L, et al., 2019). TCM formula are mainly composed of various natural herbs, which provide broad prospects for prophylaxis and treatment for diseases such as AD in (Sun YW, et al., 2019; Wang T, et al., 2018). However, the TCM formula are characterized by their multi-elements, multi-targets, multi-signaling pathways and holistic synergistic effects that complete particular preventive and therapeutic effects, thus a great challenge lying ahead of TCM (Chen X, et al., 2013; Yu G, et al., 2018). Therefore, it is of great significance to explore systematically and elucidate the mechanism of TCM formula via novel methods and strategies (Casas AI, et al., 2019). Network pharmacology is exactly a novel approach to investigate active constituents of TCM and their underlying targets (Ding M, et al., 2019). It is able to provide novel probabilities for illuminating multi-scale molecular mechanisms of TCM to treat AD.

Compound Cu-zhi-2-hao-fang (CZ2HF) is a Chinese herb preparation based on TCM prescription, which is of our research team’s patent (No. CN107441468A). This compound preparation is composed of Herba Epimedium, Rhizoma curculiginis, Morinda officinalis, Acorus calamus, Lycium barbarum, Scrophularia ningpoensis, Cinnamomum cassia Presl, Rhizoma zingiberis. Its validity in AD treatment has been verified in clinic. Moreover, in our previous study, CZ2HF also showed an obvious beneficial effects on Aβ25-35-induced learning and memory impairment and hippocampal neuronal injury in rats (Zeng L, et al., 2019); however, its exact molecular mechanism remains still unclear. Network pharmacology approach provides a novel insight into the mechanism of anti-AD formula and its active ingredients as the approach has superiority in the analysis of multi-compounds, multi-targets, and multi-effects. Thus, in this study, network pharmacology was applied to decipher the anti-AD ingredients in CZ2HF and their potential targets. Furthermore, their potential “ingredient-target-pathway” was also explored. The study will not only offer pharmacological basis, but also enhance the progress of TCM as candidate drugs for treating AD.

Materials and methods

Roadmap of the systems pharmacology approach

The roadmap was shown in Figure 1. In brief, the active ingredients of CZ2HF herbs were confirmed through an ADME-screening model, which together with the modules of oral bio-availability (OB) and drug-likeness (DL) screening. Thereafter, the latent targets of the CZ2HF formula were forecasted by target fishing. Furthermore, target-pathway networks were performed for network pharmacology analysis.

Active ingredients screening

To explore the active ingredients of CZ2HF that play an important role in anti-AD, the predicted OB and predicted DL values of them were predicted.

OB, the most important parameter in pharmacokinetic, measures absorption of medicine into blood and pharmacological action. (Zhao M, et al., 2018). In the present study, OBioavail 1.1, a potent internal model, was used to evaluate the OB values according to previous study (Xu X, et al., 2012). The molecules with OB ≥ 30% were selected as active compounds for further analysis.

DL is a comprehensive reflection of pharmacodynamics properties. It estimates molecules with “drug-like” characteristics in order to modulate corresponding targets. In current study, DL value was calculated with Tanimoto coefficient (Lv WJ, et al., 2019).

A represents the molecular descriptors of herbal components. B represents the average molecular properties of all compounds in Drug Bank database (http://www.drugbank.ca/). In the present study, compounds without information of ADME were excluded from the list and potential active chemicals were listed only if OB ≥ 30% and DL ≥ 0.18 according to previous study (Lee AY, et al., 2018).

Results

Selected compounds and its target genes

STITCH 5.0 database provided the information about predicted interactions between chemicals and genes. An interaction with a high “combine score” (≥ 700) was considered a strong interaction (Shen X, et al., 2017). Only 22 molecules of 106 compounds possessed potential target genes. Herbs-compounds network was established by Cytoscape 3.6.0 (Figure 2); moreover, 127 target genes of those molecules met the criteria of strong interactions (combine score ≥ 700) (Table 3). Since 29 target genes were duplications of the 127 genes, the compounds-genes network was composed of 98 genes and 22 compounds, which contained 120 nodes and 123 edges (Figure 3). The network of target compounds with target genes were screened based on ADME criteria of eight herbal medicines, which revealed 22 compounds and 98 target genes with 128 nodes and 153 edges (Figure 4).

Table 3. Interactions bewteen 127 target genes and molecules

NO.	Targets name	Molecule name	NO.	Targets name	Molecule name
1	SREBF2	beta-sitosterol	64	NSDHL	lophenol
2	ABCG8	beta-sitosterol	65	EBP	lophenol
3	ABCG5	beta-sitosterol	66	MSMO1	lophenol
4	APOE	beta-sitosterol	67	CYP51A1	obtusifoliol
5	DHCR24	beta-sitosterol	68	MSMO1	Sitosterol alpha1
6	CASP3	beta-sitosterol	69	C5orf4	Sitosterol alpha1
7	SREBF1	beta-sitosterol	70	ABCA1	stigmasterol
8	ABCB11	beta-sitosterol	71	ABCG8	stigmasterol
9	ICAM1	beta-sitosterol	72	ABCG5	stigmasterol
10	CYP7A1	beta-sitosterol	73	TNF	stigmasterol
11	SREBF2	sitosterol	74	IL8	stigmasterol
12	ABCG8	sitosterol	75	SLCO1B1	stigmasterol
13	ABCG5	sitosterol	76	IL10	stigmasterol
14	APOE	sitosterol	77	APOB	taxifolin
15	DHCR24	sitosterol	78	TNFSF11	taxifolin
16	CASP3	sitosterol	79	NQO1	taxifolin
17	SREBF1	sitosterol	80	ABCC1	taxifolin
18	ABCB11	sitosterol	81	MTTP	taxifolin
19	ICAM1	sitosterol	82	ASP3	peroxyergosterol
20	CYP7A1	sitosterol	83	SP7	icaritin
21	FDFT1	squalene	84	EIF2AK3	icaritin
22	LSS	squalene	85	BAK1	icaritin
23	DHCR24	squalene	86	CCND1	icaritin
24	CCL4	squalene	87	ESR1	icaritin
25	CCL5	squalene	88	ESR2	icaritin
26	CCL2	squalene	89	ABCC1	chrysoeriol
27	CCL3	squalene	90	NOTCH2	liquiritigenin
28	IL1RN	squalene	91	ESR2	liquiritigenin
29	IL8	squalene	92	PDE5A	icariin
30	NR1I2	kaempferol	93	ESR1	icariin
31	CDK1	kaempferol	94	NOS3	icariin
32	UGT3A1	kaempferol	95	PNPLA2	icariin
33	CYP1B1	kaempferol	96	SCARB1	icariin
34	RPS6KA3	kaempferol	97	JUN	icariin
35	AHR	kaempferol	98	SLCO2B1	icariin
36	UGT1A9	kaempferol	99	SLCO1B3	icariin
37	UGT1A3	kaempferol	100	RELA	icariin
38	UGT1A8	kaempferol	101	MAPK8	luteolin
39	UGT1A7	kaempferol	102	MMP9	luteolin
40	CHRM1	atropine	103	CASP3	luteolin
41	CHRM2	atropine	104	JUN	luteolin
42	CHRM4	atropine	105	FOS	luteolin
43	CHRM3	atropine	106	CDK2	luteolin
44	CHRM5	atropine	107	EGFR	luteolin
45	BCHE	atropine	108	SMAD2	luteolin
46	ACHE	atropine	109	CCNA2	luteolin
47	TAC1	atropine	110	AKT1	luteolin
48	GAST	atropine	111	MCL1	quercetin
49	DHCR24	campesterol	112	CYP1B1	quercetin
50	HSD3B2	campesterol	113	PIM1	quercetin
51	ABCG8	campesterol	114	HCK	quercetin
52	MSMO1	cycloeucalenol	115	SLC2A2	quercetin
53	C5orf4	cycloeucalenol	116	CYP2C8	quercetin
54	AKR1B1	fucosterol	117	CYP1A1	quercetin
55	PTPN1	fucosterol	118	ATP5B	quercetin
56	MMP13	glycitein	119	HIBCH	quercetin
57	MAPK1	glycitein	120	STK17B	quercetin
58	MAPK3	glycitein	121	STAT3	icariside II
59	JUN	glycitein	122	EGF	icariside II
60	SP1	glycitein	123	PARP1	icariside II
61	CDK4	glycitein	124	UGT1A10	icariside II
62	HSD17B7	lophenol	125	UGT1A1	icariside II
63	C5orf4	lophenol	126	UGT1A7	icariside II

Potential target genes and network analysis related to AD

The PPI network of the 98 target genes was created by Cytoscape, which included 98 nodes and 829 edges. Moreover, the size of each node is proportional to its degree of being involved in this network. Notably, AKT1 and TNF owned the first two numbers of degrees. The AD-associated genes were collected from the TTD. 28 repeating genes were selected through matching above-mentioned 98 genes with AD-related genes, which were labeled purple (Figure 5). Moreover, the network of target compounds with AD-related target genes, which were screened with ADME criteria of seven herbal medicines, revealed 9 compounds and 28 target genes with 44 nodes and 43 edges (Figure 6).

Pathway analysis

To identify the function and related signaling pathways of these target genes, we applied KEGG pathways and GO term to conduct functional enrichment analysis. We listed top 20 signaling pathways which related to aforementioned 98 genes (Figure 7). The results showed that the top 10 pathways were related to the process of BP, MF, CC (Figure 7). Top 20 signaling pathways related to AD-associated genes were also identified (Figure 7). Ultimately, herbs-compounds-genes network related to AD was created by Cytoscape (Figure 8).

Molecular docking

The results further revealed that the binding energy between TNF and stigmasterol, ICAM1 and sitosterol, ICAM1 and beta-sitosterol, MMP9 and luteolin, IL-10 and stigmasterol were -9.14 kcal/mol, -6.4 kcal/mol, -6.59 kcal/mol, -9.99 kcal/mol, -8.7 kcal/mol, respectively (Figure 9).

Discussion

Emerging evidence shows that TCM exerts excellent neuroprotective effects on AD due to its multiple active ingredients and targets. Although our previous study has revealed the therapeutic effects of CZ2HF formula on Aβ25-35-induced AD in rats (Zeng L et al., 2019), its therapeutic mechanisms remains still a mystery. Therefore, we applied a network pharmacology with large-scale text-mining, ADME screening, target predicting to decipher the active compounds, potential targets and possible mechanisms of CZ2HF formula for treatment of AD.

In this study, the results demonstrated that 22 compounds of eight herbs from CZ2HF were screened by ADME and 98 potential targets were identified among them as evidenced by STITCH DB database. Notably, sitosterol was detected in five different herbs of CZ2HF, as well as beta-sitosterol. Accordingly, it is reported that beta-sitosterol was considered as a potential anti-AD drug due to its beneficial effects on memory deficits (Adebiyi OE, et al., 2019). Furthermore, 98 targets were selected from the 22 compounds, of which there were 28 AD-related genes. Beta-sitosterol and sitosterol commonly aim 10 targets of 98 genes respectively; among them, APOE, ICAM1, DHCR24 and CASP3 were related to AD. Moreover, recent report indicated that luteolin, a natural flavonoids compound, exerted neuroprotective effects on AD by suppressing neuroinflammation (Contarini G, et al., 2019). Interestingly, our results showed that 9 potential targets of luteolin were confirmed by network pharmacology analysis. Among them, MAPK8, AKT1, EGFR and MMP9 were related to AD. Icariin and icariside II are the main pharmacological compounds of Herba Epimedium. More importantly, our previous studies have demonstrated that both icariin and icariside II could halt the disease progress of AD by targeting Aβ formation and neuroinflammation (Yan L, et al., 2017; Zong N, et al., 2016). In this study, network analysis successfully identified 6 AD-associated targets that belonged to icariin and icariside II. Among them, PDE5A was considered as a significant target because after inhibition of PDE5 by icariin or icariside II, the effects of anti-AD could be achieved (Jin F, et al., 2014; Yin C, et al., 2016). Evidences have demonstrated that PPI network could be useful in discovering the multiple interactions of numerous proteins in some complicated diseases including AD, and identifying potential therapeutic targets (Zhang JY, et al., 2019). In this study, AKT1, TNF, ERS1, EGFR, JUN, CASP3, EGF, STAT3, MAPK3 and FOS were considered as key target genes as evidence of its high degree.

The findings also revealed that 98-gene-related 20 top pathways were annotated by KEGG pathway and GO term. Among them, the molecular biology of specific pathways including lipopolysaccharide-mediated and TNF signaling pathways, which were pivotal signaling pathways in the network and the potential therapeutic targets in AD. Furthermore, the results demonstrated that the major target genes in the network including TNF, ICAM1, MMP9 and IL-10, which were involved in the inflammatory-related pathways. TNF is a pivotal pro-inflammatory cytokine in inflammatory responses, which plays a crucial role in normal brain function at physiological conditions (Clark IA, et al., 2018); however, excessive TNF contributes to the pathological process of neuroinflammatory disease like AD. That is link to neuroinflammation characterized by expression of TNF (Wu Y, et al., 2019). ICAM-1 is low at physiological level. However, it is quickly induced by pro-inflammatory cytokines. ICAM-1 is closely associated with neuroinflammation in AD (Walker DG, et al., 2017). Moreover, MMP9 is an crucial inflammatory component. Its activation has demonstrated to be involved in AD pathogenesis and is considered as a feature of AD due to its neurotoxic effect (Caldeira C, et al., 2017). Notably, IL-10 is an important anti-inflammatory regulator, which exerts against both central and peripheral inflammatory diseases (Guillot-Sestier MV, et al., 2015). Intriguingly, our results suggested that stigmasterol directly were bonded to TNF and IL-10, luteolin directly bonded to MMP9, and sitosterol directly bonded to ICAM1 with evidence collected from molecular docking method. Our findings demonstrated that these pathways were involved in the inhibitory effect on AD of CZ2HF. However, further research was dire to systematically decipher the other biological activities linked with CZ2HF, and elucidate its pivotal mechanisms. Among the target ingredients associated with the network pharmacology analysis, stigmasterol, sitosterol, luteolin were identified both as major active anti-inflammatory and anti-AD compounds of CZ2HF with evidence collected from compound-target-pathway networks. Stigmasterol is detected both in Lycium barbarum and Rhizoma curculiginis, which has been proved to be able to markedly attenuate cognitive impairment in vanadium-induced neurotoxicity(Adebiyi OE, et al., 2018); while, sitosterol is detected both in Morinda officinalis and Scrophularia ningpoensis, which has been regarded as a potential agent for the treatment of memory deficit disorders such as AD (Ayaz M, et al., 2017). Furthermore, luteolin, an active compound derives form Herba Epimedium also exerts therapeutic effects on Aβ1–40-induced injury in rats through mediation of NF-κB signaling pathways (Zhang JX, et al., 2017). Interestingly, further evidence showed that stigmasterol directly were bonded to TNF and IL-10, sitosterol and beta-sitosterol directly bonded to ICAM1, luteolin directly bonded to MMP9, which were consistent with the results of previous study as well as network analysis. Taken together, these results indicated that the major anti-inflammatory ingredients of CZ2HF were effective for the treatment of AD.

Conclusion

In summary, network pharmacological analysis of CZ2HF predicts that 22 compounds and 98 target genes are identified in the eight herbal medicines of CZ2HF with evidence collected from Network pharmacological analysis. In particular, 28 AD-related genes are further identified. The major compounds are sitosterol, beta-sitosterol, stigmasterol, and luteolin, which are associated with the core genes including TNF, ICAM1, MMP9 and IL-10. Our findings offer a further insight into the underlying mechanism of CZ2HF on treating AD from a multi-scale perspective.

Ethical statement

Not applicable.

Funding

This work was supported by the Natural Science Foundation of China (Grant No. 81760727), Science and Technology Support Plan of Guizhou Province (Grant No. Supporting Science and Technology Cooperation of Guizhou Province [2020]1Y010), Innovative Research Team of comprehensive utilization with Lithocarpus polystachyus Rehd, sweet tea in Zunyi City (Grant No. Technology Talents [2021]4), National key R & D plan for Research on modernization of Traditional Chinese Medicine (Grant No. 2017YFC1702005), Post subsidy project of State key R & D plan in social development field (Grant No. SQ2017YFC170204-05), Program for Changjiang Scholars and Innovative Research Team in University, China (Grant No. IRT-17R113), Program for Outstanding Youth of Zunyi Medical University (Grant No. 15zy-002).

Transparency statement

All the authors affirm that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Authors' contribution

Yu Wei, Qi-Hai Gong and Chang-Yin Yu contributed the central idea, analysed most of the data, made figures & tables and wrote the initial draft of the paper. Jian-Mei Gao, Fan Xu and Jing-Shan Shi contributed to refining the ideas, carrying out additional analyses and finalizing this paper.