Novel Combination of HuanglianJiedu Decoction in Behavioral and Psychological Symptoms of Dementia in Alzheimer's Disease

Alzheimer's disease (AD) is characterized by progressive cognitive decline. Besides cognitive decit, AD is also characterized by behavioral and psychological symptoms in dementia (BPSD). However, therapeutic management of BPSD remains challenging. HuanglianJiedu decoction (HLJDD), a traditional Chinese prescription, consisting of four herbs, is applied to treat AD, especially AD with BPSD. Though HLJDD, has the traditional combination with the principal herb Coptidis rhizoma (Huang-lian), it might, however, not be suitable for treating BPSD. Elucidating the mechanism underlying each herb is critical to the disease-matched combination of HLJDD. In this study, network pharmacology was used to determine the targets and biological processes regulated by HLJDD in the treatment of BPSD. Moreover, molecular docking was utilized to evaluate the binding activity between the herbs' main active ingredients and neurotransmitter receptors. The results showed that Scutellariae radix (Huang-qin) and Phellodendri chinrnsis cortex (Huang-bai) exhibited better anti-BPSD effects when compared to Coptidis rhizoma and Gardeniae fructus (Zhi-zi). Scutellariae radix exhibited superior anti-neuroinammation functions, with better blood vessel regulation effects. Phellodendri chinrnsis cortex showed a higher binding anity to the dopamine D2 receptor (DRD2) and 5-hydroxytryptamine receptor 2A (HTR2A). Coptidis rhizoma and Gardeniae fructus were better in neuronal signaling. In conclusion, for treating BPSD, Scutellariae radix and Phellodendri chinrnsis cortex are the principal herbs while Coptidis rhizoma and Gardeniae fructus are the ancillary herbs. testing compounds: poriferast-5-en-3beta-ol (beta-sitosterol), Stigmasterol, chelerythrine, and campesterol, which have a lower anity for DRD2 and HTR2A. Scutellariae radix and Phellodendri chinrnsis cortex were found to have the primary compounds for treating BPSD. Scutellariae radix plays an anti-inammatory role while Phellodendri chinrnsis cortex regulates apoptotic processes and response to nicotine. They both regulate blood vessels. Coptidis rhizoma and Gardeniae fructus play the assistant role in BPSD, and are involved in neuronal signaling.


Introduction
In traditional Chinese medicine (TCM), the combination of herbs into a prescription to enhance their overall therapeutic effects or eliminate the adverse effects [1], is referred to as TCM combination. Although prescriptions have conventional combinations, it is essential to consider novel combinations for different diseases. In this study, we propose an innovative method that is based on molecular mechanism for determining herbs combination. We investigated the novel combination of HLJDD in BPSD.
The prevalence of BPSD, including anxiety, agitation, aggression, irritation, depression, apathy, disinhibition, delusions, or hallucinations in AD patients is more than 90% [2]. Molecules involved in the pathogenesis of BPSD have been widely studied. Neuroin ammation, a response that involves neurons and microglia, has been reported to characterize several neurodegenerative diseases and neuropsychiatric conditions, resulting in the elevated production of pro-in ammatory cytokines, like IL-6, TNF-α, IL-8, and IL-4. Notably, microglia activation is the rst sign of neuroin ammation. Activated microglia can release various oxidants such as reactive oxygen species and activate several genes and proteins, such as inducible nitric oxide synthase [3]. Neurotransmitters and their receptors are also considered to play a crucial role in BPSD [4]. Evaluation of genetic risk factors provide a powerful approach for elucidating the mechanisms underlying BPSD. APOE epsilon4, the most common genetic risk factor for the late-onset of AD, elevates the risk of BPSD [5]. However, the e cacy of antipsychotics for the treatment of BPSD is scanty. Antipsychotic drugs used in the clinical management of BPSD have extrapyramidal severe side effects (EPS). Memantine, an NMDA receptor antagonist for treating moderate-to-severe AD, inhibits antipsychotic-induced EPS [6], but is controversial for treating BPSD. The current primary therapeutic options for AD, acetylcholinesterase inhibitors, are not effective for treating BPSD [7]. Several researchers turn their attention to natural products as an alternative or complementary method to BPSD for their clinical e cacy with minimal side effects. BPSD has different TCM syndrome differentiation. "Toxin damaging brain collateral" an essential BPSD differentiation, was proposed by Yongyan Wang. For this syndrome differentiation, brain collateral is injured by re toxins. For the treatment, purging re for removing toxins is commonly used. HLJDD, which is composed of Scutellariae radix, Coptidis rhizoma, Phellodendri chinrnsis cortex, and Gardeniae Fructus, is a classic prescription for heat-clearance and detoxi cation. Studies have documented that HLJDD is frequently used in AD management [8], especially with BPSD [9]. While the TCM physicians seldom treat BPSD with whole herbs in HLJDD or with HLJDD only. They prefer to choose two or more principal herbs in HLJDD with other syndrome-match herbs. According to HLJDD's traditional combination, the TCM physicians now prefer to choose Coptidis rhizoma as the primary herb to treat BPSD. However, in this research, Scutellariae radix and Phellodendri chinrnsis cortex are the remedying herbs in BPSD. So we put forward that the traditional combination of prescription need refers to the novel combination with modern technology. To elucidate HLJDD's combination in treating BPSD, we used network pharmacology and molecular docking to provide useful drugs for BPSD therapeutics by emphasizing their molecular activities ( Figure 1).

Material And Methods
2.1. Determination of active ingredients and their matched potential target proteins. A natural plant contains various chemical compounds. We used the TCM systems pharmacology database and analysis platform (TCMSP, https://tcmspw.com/index.php) to determine the main active ingredients in the herbs. Compounds were ltered by oral bioavailability (OB) ≥ 30%, drug-likeness (DL) ≥ 0.18, and drug half-life (HL) ≥ 4 h. The potential molecular targets for the active compounds were predicted using the search tool for interactions of chemicals database (STITCH, http://stitch.embl.de/), with the species limited to "Homo sapiens." 2.2. Potential targets prediction for BPSD. BPSD-associated protein targets were identi ed using the GeneCards database (http://www.genecards.org/) [10]. These target proteins exhibited a higher rank score, and a higher correlation with BPSD. The keywords used in the search were "behavioral and psychological symptoms of dementia in Alzheimer's disease." The top 50% of the predicted targets were selected as potential BPSD targets.
2.3. Network construction and analysis. We constructed the "compounds-targets-BPSD" networks by Cytoscape 3.7.2 software [11]. The .csv format les whose data combined target compounds and BPSD targets were imported into Cytoscape. The node size was based on the target proteins score values as provided by the GeneCards database. Overlapping targets between active compounds and BPSD were the herbs' putative targets, and set with a red rectangle node. Memantine was used to identify the potential neurotransmitter receptors.
2.4. Enrichment analysis of target proteins. We performed Gene Ontology (GO) functional enrichment analysis using the ClueGO plugin [12], with the species limited to "Homo sapiens" and p < 0.05.
2.5. Molecular docking for active compounds with DRD2 and HTR2A. Memantine was selected as the active control medicine. The molecular mechanism of memantine on BPSD was determined by network pharmacology, same as the HLJDD herbs. The results showed that DRD2 and HTR2A were the targets of neurotransmitter receptors. Then molecular docking studies were conducted using AutoDock software [13] to evaluate the a nity of the active compounds bind with DRD2 and HTR2A neurotransmitter receptors. Crystal structures of the proteins were downloaded from the RCSB Protein Data Bank (http://www.rcsb.org/). PDB ID "6CM4" and "6A93", whose ligand is Risperdal, were selected as DRD2 and HTR2A's crystal structure. Memantine, Risperdal, and HLJDD's active compounds were then docked with DRD2 and HTR2A.

Results
3.1. HLJDD's main active compounds and potential target proteins. After searching the TCMSP database, we identi ed 80 compounds in HLJDD. Scutellariae radix, Coptidis rhizoma, Phellodendri chinrnsis cortex, and Gardeniae fructus have 32, 10, 27, and 11 ingredients, respectively. Then, these compounds were imported into the STITCH database to achieve the potential target proteins (Table 1). 3.2. "Compounds-targets-BPSD" network and GO biological process analysis. In this network, the red rectangular nodes represent the key protein targets of the herbs. The targets for Scutellariae radix were the highest (24), followed by Phellodendri chinrnsis cortex (15). We used the GO biological process analysis to further study the herbs' key protein targets. Figure 2 b and Table 2 show the biological processes of the key target proteins of Scutellariae radix, including positive regulation of smooth muscle cell proliferation, lipopolysaccharide-mediated signaling pathway, and regulation of neuroin ammatory responses, etc. Figure 3 b, c, d, and Table 3 show the biological processes of the key target proteins of Coptidis rhizoma. Its biological processes include regulation of cellular responses to cadmium ions, responses to nicotine, and negative regulation of macroautophagy. Figure  4 b, c, d, and Table 4 show the biological processes of the key target proteins for Phellodendri chinrnsis cortex. These processes include regulating responses to nicotine, glial cell apoptotic process, plasma lipoprotein particle, etc. Figure 5 b, c, d, and Table 5 show the biological processes of the key target proteins of Gardeniae fructus, which are mainly associated with the regulation of amyloid-beta formation, regulation of membrane protein ectodomain proteolysis and regulation of nitric oxide biosynthetic process.    3.3. "Compounds-targets-BPSD" network of memantine. Patients with moderate to severe AD exhibit relatively severe cognitive and psychological symptoms. N-Methyl-D-aspartic acid (NMDA) is one of the main therapeutic options. Memantine is the most common therapeutic choice for NMDA [14]. In our study, memantine was used as the reference drug. We established the "compounds-targets-BPSD" network of memantine to determine its potential mechanism in BPSD, especially for the neurotransmitter receptor. It was found that DRD2 and HTR2A were the neurotransmitter receptors associated with BPSD.
3.4. Results of molecular docking. Serotonin and its receptors, particularly the HTR2A, play roles in cognitive behaviors and psychiatric conditions such as depression, schizophrenia, and AD [4]. A multitarget-directed ligand, acting on HTR2A and DRD2, has been shown to exert an anti-aggressive and antipsychotic activity and is, therefore, a promising therapeutic option for BPSD [15]. In our study, herbs active ingredients were docked with DRD2 and HTR2A. The 3D structures of the active ingredients were downloaded from the TCMSP database in .mol2 format. They were later converted to .pdb format using the Open Babel GUI software [16]. The local search parameters and rigid lenames of the macromolecule models were used when docking. We found that beta-sitosterol exhibited the lowest docking energy with DRD2 (-9.58 Kcal/mol) and HTR2A (-8.2 Kcal/mol) when compared to the other compounds, including memantine and Risperdal. Stigmasterol, chelerythrine, and campesterol also exhibited a lower docking energy than memantine and Risperdal. Phellodendri chinrnsis cortex was shown to contain most of these compounds. (Table 6 and Figure 6).

Discussion
TCM plays an important role in medical diagnosis and treatments. Based on the TCM theory, Chinese formulas contain a mixture of herbs, combined based on the following combination principle; "monarch (Jun), minister (Chen), assistant (Zuo), and guide (Shi)" meaning that herbs play primary, secondary, auxiliary, or harmonic roles, respectively [17]. Primary herbs are substances that provide the main therapeutic thrust. Secondary herbs enhance or assist the therapeutic actions of the primary. The rest have the following functions: treat accompanying symptoms, moderate the harshness or toxicity of the primary ones, guide the medicine to the proper organs, or exert a harmonizing effect [18]. In HLJDD, Coptidis rhizoma plays the monarch role, Scutellariae radix plays the minister role while Phellodendri chinrnsis cortex and Gardeniae fructus play the assistant and guide roles, respectively. However, this combination is not suitable for treating BPSD. We studied drugs from the molecular perspective, and proposed a novel method for testing the combination of formulas. We found that Scutellariae radix and Phellodendri chinrnsis cortex are the principal herbs, while Coptidis rhizoma and Gardeniae fructus are the assistant herbs.
In AD, the abnormal accumulation of amyloid-β released from amyloid precursor protein and neuroin ammation are its partially pathologic hallmarks. Amyloid-β accumulation also causes indirect injuries to neurons by inducing neuroin ammation [19]. Microglia, the resident innate immune cells in the brain, are important in AD immune responses. They act as sentinel and protective cells, but may become inappropriately reactive in AD to drive neuropathology [20]. With advances in age, microglia exhibit enhanced sensitivity to in ammatory stimuli, similar to that observed in brains with ongoing neurodegeneration [21]. The lipopolysaccharide (LPS), a gram-negative bacterial endotoxin released from the cell wall, mediates inflammation in the body, involving in regulating the expression of potential inflammatory factors [22]. Studies have linked schizophrenia with neuroin ammatory conditions and microglia, which have been correlated to the pathogenesis of schizophrenia. Neuroin ammatory changes observed in schizophrenia involve abnormal astrocyte functions [23]. In our study, Scutellaria baicalensis was shown to play anti-in ammatory roles, including regulating neuroin ammatory responses, mediating lipopolysaccharide-mediated signaling pathways, cellular response to interleukin-6, and regulation of interferon -α production during BPSD treatment. Gardeniae fructus was shown to regulate amyloid-beta formation in BPSD therapy.
Smooth muscle cell proliferation, especially vascular smooth muscle cells, are essential during cell growth or injury [24]. Blood vessels with vascular smooth muscle cells play an important role in normal brain functions. Apart from supplying adequate blood, they help maintain its structural integrity and function. Shortages in cerebral blood ow and blood-brain barrier dysfunction are early ndings in neurodegenerative disorders. Cerebral blood ow shortage, impaired cerebrovascular reactivity, and impaired hemodynamic responses are increasingly prevalent in the early stages of AD [25]. Platelets, critical blood ow factors, are anucleate blood cells whose principal function is to stop bleeding by forming aggregates for hemostatic reactions. Platelet aggregates are also involved in pathological thrombosis and play an essential role in in ammation [26]. Nitric oxide (NO) is a small free radical molecule with an endothelium-derived relaxing factor. Adequate levels of NO in the vascular endothelium are critical for regulating blood flow and vasodilation. Moreover, NO plays a vital neuronal signaling role [27]. Cadmium, a metal that resembles zinc and calcium, is also crucial for neuronal signaling. Cadmium exposure is associated with neurodegenerative diseases such as AD. It can alter neurotransmitters' release, cause oxidative stress, damage the mitochondria, and induce apoptosis [28]. In our study, Scutellariae radix was shown to positively regulate smooth muscle cell proliferation, endothelial cell proliferation, and blood vessel endothelial cell migration; Coptidis rhizoma was shown to mediate cellular responses to cadmium ion; Phellodendri chinrnsis cortex was shown to regulate platelet activation while Gardeniae fructus regulated the nitric oxide biosynthetic process.
Macroautophagy is an evolutionarily conserved dynamic pathway that functions primarily in a degradative manner. Various diseases are associated with macroautophagic dysregulation. Macroautophagy plays a critical role in cellular homeostasis. Insu cient or excessive macroautophagy can seriously compromise cell physiology [29]. Coptidis rhizoma was shown to negatively regulate macroautophagy while Phellodendri chinrnsis cortex regulated glial cell apoptosis.
Studies have documented that Nicotine might be involved in the pathophysiology of psychosis. Smoking is correlated with depression. In animal models, Nicotine exhibited anxiolytic properties. Depressed people are more likely to smoke and more likely to develop severe depressive episodes upon smoking cessation. Nicotine has also been observed to exhibit similar cognitive improvements in AD patients [30]. However, the relationship between smoking and AD has not been clearly elucidated [31]. In the study, Coptidis rhizoma and Phellodendri chinrnsis cortex can respond to nicotine.
It has been documented that serotoninergic, dopaminergic, and cholinergic systems are involved in BPSD pathogenesis, and the roles of HTR2A and DRD2 as therapeutic targets are evident [15]. We used molecular docking to determine the potential active ingredients that exhibited good binding activities to DRD2 and HTR2A. The best-docked compound was beta-sitosterol. The free binding energy of beta-sitosterol with DRD2 and HTR2A was -9.58 kcal/mol, -8.2 kcal/mol, respectively. Stigmasterol, chelerythrine, and campesterol also exhibited good binding activities. The active ingredients of Phellodendri chinrnsis cortex were found to be beta-sitosterol, Stigmasterol, chelerythrine, and campesterol. It is the only herb containing these ve ingredients.
The TCM theory veri es our results. Triple energizers in TCM theory mean upper, middle, and lower energizers. They are the birth and channel to run for Qi, blood, thin, thick uids, and essence. Moreover, they also contact ve Zang-organs and six Fu-organs. Scutellariae radix affects the upper energizer, which consists of the brain and heart; Phellodendri chinrnsis cortex affects the lower energizer, which is the kidney and liver while Coptidis rhizoma in uences the middle energizer that consists of the spleen and stomach. Kidney essence de ciency is a primary syndrome differentiation of AD in TCM theory. Due to the imbalance between yin and yang of liver functions, "Liver re" is the largest contributor to BPSD [32 , 33]. Phellodendri chinrnsis cortex acts on the lower energizer to purge the liver re. "Su Wen" puts forward that "the mind is the monarch's o cial, and the gods come out of it." In the compendium of Materia Medica, Shizhen Li of the Ming Dynasty proposed that "the brain is the house of primordial God." Scutellariae radix works on the heart and brain, belonging to the upper energizer. In conclusion, based on the TCM theory, Scutellariae radix and Phellodendri chinrnsis cortex are HLJDD's primary herbs.

Conclusions
The therapeutic value of natural products in the management of BPSD has increased due to their clinical e cacy and insigni cant side effects. Different types of compounds have been reviewed for their biological activities. Our results showed that Scutellariae radix has more molecular targets and biological processes involved in BPSD while Phellodendri chinrnsis cortex exhibited more well-docked compounds: poriferast-5-en-3beta-ol (betasitosterol), Stigmasterol, chelerythrine, and campesterol, which have a lower a nity for DRD2 and HTR2A. Scutellariae radix and Phellodendri chinrnsis cortex were found to have the primary compounds for treating BPSD. Scutellariae radix plays an anti-in ammatory role while Phellodendri chinrnsis cortex regulates apoptotic processes and response to nicotine. They both regulate blood vessels. Coptidis rhizoma and Gardeniae fructus play the assistant role in BPSD, and are involved in neuronal signaling. The data used to support the ndings of this study are available from the rst author upon request.

Competing interests
The authors declare no competing nancial interest. Suya Ma wrote the manuscript. XW constructed the pharmacological networks. Jing Shi revised the manuscript. All authors were responsible for reviewing data. All authors read and approved the nal manuscript.