Human-Induced Pluripotent Stem Cells and Herbal Small-Molecule Drugs for Treatment of Alzheimer’s Disease

Alzheimer’s disease (AD) is characterized by extracellular amyloid plaques composed of the β-amyloid peptides and intracellular neurofibrillary tangles and associates with progressive declines in memory and cognition. Several genes play important roles and regulate enzymes that produce a pathological accumulation of β-amyloid in the brain, such as gamma secretase (γ-secretase). Induced pluripotent stem cells from patients with Alzheimer’s disease with different underlying genetic mechanisms may help model different phenotypes of Alzheimer’s disease and facilitate personalized drug screening platforms for the identification of small molecules. We also discuss recent developments by γ-secretase inhibitors and modulators in the treatment of AD. In addition, small-molecule drugs isolated from Chinese herbal medicines have been shown effective in treating Alzheimer’s disease. We propose a mechanism of small-molecule drugs in treating Alzheimer’s disease. Combining therapy with different small-molecule drugs may increase the chance of symptomatic treatment. A customized strategy tailored to individuals and in combination with therapy may be a more suitable treatment option for Alzheimer’s disease in the future.


Introduction
Alzheimer's disease is a prevalent neurodegenerative disease that worsens over time and causes brain lesions that can result in behavioral changes. It accounts for the largest number of dementia cases [1]. In the United States, the Alzheimer's Association estimates approximately two new cases per year per 1000 people aged 65 to 74, 11 new cases per year per 1000 people aged 75 to 84, and 37 new cases per year per 1000 people aged 85 and older [2]. Alzheimer's disease is characterizing by short-term memory loss accompanied by depression. Impairment of language ability leads patients to avoid difficult or complex words; the ability to recognize written characters is greatly reduced as well [3]. In the older age group (>65 years), Alzheimer's disease is one of several dreaded neurodegenerative diseases that cause dementia. Others include frontotemporal dementia, dementia with Lewy bodies, vascular dementia, progressive supranuclear palsy, alcoholic dementia, Creutzfeldt-Jakob disease, and corticobasal degeneration (Figure 1) [4]. Alzheimer's disease has two forms: early-onset and late-onset. Early-onset Alzheimer's disease, also known as familial Alzheimer's disease, occurs mainly before the age of 65 in association with hereditary genetic mutations. Late-onset Alzheimer's with Alzheimer's disease. However, gene expression is most directly related to early-onset Alzheimer's disease [6]. Several gene mutations lead to increased accumulation of amyloid in the brain. The genes include amyloid precursor protein (APP), presenilin-1 (PSEN-1), and presenilin-2 . The APP gene sequence is located on human chromosome 21. APP protein was found in many tissues but is concentrated mainly in neuronal synapses. The functions of APP include the formation of regulatory synapses and the expulsion of iron atoms from cells [7]. APP protein exists in many different species and is highly conserved [8]. In humans, the PSEN-1 gene sequence is located on chromosome 14, and the PSEN-2 gene sequence is located on chromosome 1. PSEN-1 and PSEN-2 are highly homologous [9]. The function of PSENs is to form a complex with APP in cells and participate in the translocation and post-synthetic processing of APP [10].
Pathologically, AD is characterized by extracellular amyloid plaques composed of the β-amyloid peptides and intracellular neurofibrillary tangles that comprise the microtubule-associated protein tau [11]. Understanding the two pathologies that lead to neuronal death has been a focus of AD research. However, the generally used cell models of AD-PC12 and SH-SY5Y require the addition of Aβ1-42 peptides with induced cytotoxicity to mimic the symptoms of the disease [12][13][14][15][16][17]. However, the amyloid plaques are not the only symptoms of AD; hence, these in vitro research models are not comprehensive. The study of AD mechanisms in a human genetic context and natural display of AD symptoms in a cell model allow the understanding of AD pathology. Induced pluripotent stem cells (iPSCs) are a pathologically relevant in vitro model for the mechanistic studies and preclinical drug discovery of AD [18]. iPSCs are pluripotent stem cells formed by the lentiviral vector transfer of four transcription factors (i.e., Oct4, Sox2, c-Myc, and Klf4) into adult cells [19]. The benefits of iPSCs are that they differentiate into different types of neuronal cells from the stem cell stage and can be used to produce forebrain acetylcholine neurons, dopaminergic progenitor cells, and Purkinje cells [20][21][22][23]. Neuronal cells derived from iPSCs exhibit electrophysiological responses. Different types of iPSC-derived neurons have been used in the study of different diseases [24][25][26]. Basal forebrain cholinergic neurons Pathologically, AD is characterized by extracellular amyloid plaques composed of the β-amyloid peptides and intracellular neurofibrillary tangles that comprise the microtubule-associated protein tau [11]. Understanding the two pathologies that lead to neuronal death has been a focus of AD research. However, the generally used cell models of AD-PC12 and SH-SY5Y require the addition of Aβ1-42 peptides with induced cytotoxicity to mimic the symptoms of the disease [12][13][14][15][16][17]. However, the amyloid plaques are not the only symptoms of AD; hence, these in vitro research models are not comprehensive. The study of AD mechanisms in a human genetic context and natural display of AD symptoms in a cell model allow the understanding of AD pathology. Induced pluripotent stem cells (iPSCs) are a pathologically relevant in vitro model for the mechanistic studies and preclinical drug discovery of AD [18].
iPSCs are pluripotent stem cells formed by the lentiviral vector transfer of four transcription factors (i.e., Oct4, Sox2, c-Myc, and Klf4) into adult cells [19]. The benefits of iPSCs are that they differentiate into different types of neuronal cells from the stem cell stage and can be used to produce forebrain acetylcholine neurons, dopaminergic progenitor cells, and Purkinje cells [20][21][22][23]. Neuronal cells derived from iPSCs exhibit electrophysiological responses. Different types of iPSC-derived neurons have been used in the study of different diseases [24][25][26]. Basal forebrain cholinergic neurons degenerate in AD, Down syndrome, and dementia. Injectable acetylcholine agonists improve memory degradation in patients with AD [27]. Therefore, acetylcholine cells play a potentially regulatory role in memory and learning in the central nervous system [28]. The platelet-derived growth factor promoter driven amyloid precursor protein (PDAPP) transgenic mice exhibited amyloid deposition in the brain and memory dysfunction.
The cholinergic neurons derived from human iPSCs are transplanted to the hippocampus of PDAPP transgenic mice. Compared with vehicle-injected PDAPP mice, the transplantation of iPSC-derived differentiated neuronal precursor cells to PDAPP mice significantly improved memory dysfunction [29]. Purkinje cells are the only neurons in the cerebellar cortex that automatically produce two types of action potentials [30]. Damage to Purkinje cells is related to cognitive impairment and motor deficits in patients with AD [31]. The iPSC-induced increase in the number of Purkinje cell bodies is a potential novel strategy for AD treatment [22,32].
The AD cell model is useful for testing potential therapeutic drugs. Its use will increase the possibility of identifying an effective AD treatment through the collection of cells from patients with AD or Down syndrome and their reversal to iPSCs. The cells that have differentiated into neuronal cells and produce AD symptoms render them useful for exploring whether drugs are effective in the inhibition of these symptoms [33]. The use of iPSCs from patients with AD who have familial gene mutations will increase the possibility of developing effective therapeutic drugs, particularly those targeting the mechanisms of Aβ accumulation [34]. For example, a recent study using an iPSC platform exclusively derived from patients with AD was effective in the identification of small molecules and compounds with synergistic anti-Aβ effects [35]. Moreover, the collection of different early-onset AD gene mutations can produce different iPSCs based on AD symptoms. These approaches are helpful for understanding the mechanisms of AD and for developing drugs for this disease (Table 1).

APP Gene Mutations
Early-onset APP gene-mutated patient cells were used to produce a model of iPSCs overexpression, including a double mutation (KM670/671NL) of APP, which increases the total Aβ burden [36], followed by APP duplication to prepare iPSCs from patients with late-AD onset. The results revealed significantly higher levels of Aβ 1-40, phospho-tau (Thr 231), and active glycogen synthase kinase-3β (aGSK-3β) [37]. Interestingly, APP (A673T) mutations are protective against AD and cognitive decline [38].

Trisomy 21 Gene Mutation
Trisomy 21 is caused by the presence of an extra copy of the APP gene and is associated with increased Aβ peptide production. The accumulation of Aβ peptides is associated with a high risk of AD dementia in patients with Down syndrome [39]. iPSCs from individuals with trisomy 21 can induce Aβ aggregation, which causes increased expression of the tau protein and affects synapse formation and functionality [40].

Microtubule-Associated Protein Tau Gene Mutation
Microtubule-associated protein tau (MAPT) gene mutations are among the most common causes of familial frontotemporal dementia. MAPT (p.A152T) has been linked with frontotemporal dementia and Alzheimer's disease. iPSCs from individuals carrying the tau-A152T variant of MAPT neurons showed accumulation, redistribution, and decreased solubility of tau. Upregulation of tau was coupled with enhanced stress-inducible markers and vulnerability of cells to proteotoxic, excitotoxic, and mitochondrial stressors [43]. iPSCs with MAPT mutations N279K, P301L, and E10 + 16 had neuritis outgrowth deficiencies [44]. The iPSCs were prepared separately, and their phenotypes manifested accumulations of tau protein, mitochondrial dysfunction, and defects of nerve growth and differentiation [45]. Table 1. Alzheimer's diseases modeled with induced pluripotent stem (iPS) cells.

γ-secretase Inhibitors and γ-Secretase Modulators
Mutations in the PSEN-1 gene account for approximately 50% of the pathogenesis of early onset of AD and are the most common gene mutations in this disease [47]. Although gene mutations are responsible for early onset of AD, 95% of AD cases are late-onset diseases and are not necessarily caused by genetic mutations [48]. Late-onset AD can be caused by mutations in the PSEN-1 gene; such mutations account for 25% of cases of late-onset AD [49]. The PSEN-1 gene is an important regulator of γ-secretase in early-and late-onset AD ( Figure 2). γ-secretase cleaves >90 different substrates. Studies of the γ-secretase-mediated cleavage of APP and Notch have focused on AD and cancer [50]. Complete inhibition of γ-secretase, as achieved by nonselective γ-secretase inhibitors (GSIs), abolishes all γ-secretase activity, leading to decreased Aβ production and Notch signaling. A recent publication used cryo-electron microscopy to elucidate the structures of the γ-secretase basis of the APP-C83 fragment and the structures of the γ-secretase basis of the Notch 100 fragment [51]. This will help in the development of substrate-specific γ-secretase inhibitors or modulators. Therefore, we discuss the development of GSIs and γ-secretase modulators (GSMs) in clinical settings.

γ-Secretase Inhibitors
Most of the mutations in the PSEN-1 gene are single nucleotide substitution mutations, which are missense mutations. PSEN-1 mutations cause an increase in the production of amyloid-β protein (~90% 40-residue amyloid beta-protein and ~10% 42-residue amyloid beta-protein) [52][53][54]. PSEN-1 is one of the components of γ-secretase [55]. The drugs' development strategies are through inhibition cleaved by γ-secretase and reducing the production of amyloid-β protein [56,57]. However, the mechanism to treat Alzheimer's disease led to no improvement with cognitive function and accompanying adverse side effects led to failure [58]. Despite the initial failure of GSIs to treat Alzheimer's disease successfully, these inhibitors are still considered to have potential. γ-secretase suppresses the expression of the intracellular domain of the Notch protein (NICD), which represents a failure of this strategy, because the Notch protein plays a key role in the function of adult neural stem cells [59], affecting the differentiation and proliferation of neurons [60]. Development of GSIs must account for the effect on the Notch signals ( Figure 3), and GSIs can produce cell toxicity by increasing the levels of APP carboxy-terminal fragments (β-CTF) [61]. Beta-amyloid accumulation. In general, cleavage of beta amyloid precursor protein (APP) by alpha-secretase and γ-secretase cuts APP into P3 peptides (1-24~1-26 amino acids). However, once the gene generates mutations, such as APP and GSIS1, PSEN1 gene expression increases, and the wrong cleavage is likely to occur and produce not easily metabolized peptide fragments (1-40~1-42 amino acids).

γ-Secretase Inhibitors
Most of the mutations in the PSEN-1 gene are single nucleotide substitution mutations, which are missense mutations. PSEN-1 mutations cause an increase in the production of amyloid-β protein (~90% 40-residue amyloid beta-protein and~10% 42-residue amyloid beta-protein) [52][53][54]. PSEN-1 is one of the components of γ-secretase [55]. The drugs' development strategies are through inhibition cleaved by γ-secretase and reducing the production of amyloid-β protein [56,57]. However, the mechanism to treat Alzheimer's disease led to no improvement with cognitive function and accompanying adverse side effects led to failure [58]. Despite the initial failure of GSIs to treat Alzheimer's disease successfully, these inhibitors are still considered to have potential. γ-secretase suppresses the expression of the intracellular domain of the Notch protein (NICD), which represents a failure of this strategy, because the Notch protein plays a key role in the function of adult neural stem cells [59], affecting the differentiation and proliferation of neurons [60]. Development of GSIs must account for the effect on the Notch signals (Figure 3), and GSIs can produce cell toxicity by increasing the levels of APP carboxy-terminal fragments (β-CTF) [61]. . γ-secretase inhibitors. Notch and APP proteins are inactive when they are first synthesized and must have sections removed to become active. Furin and BACE first cleave these sections and activate the proteins. The second cleavage is by γ-secretase. GSIs can reduce γ-secretase but have a side effect of inhibiting Notch-signaling proteins. The function of this protein is related to nerve growth and differentiation. ADAM = A disintegrin and metalloproteinase, APP = amyloid precursor protein, BACE = beta-secretase, and GSI = γ-secretase inhibitors.

Semagacestat
Semagacestat (LY-450139) (Figure 4a) was the first GSI to enter a phase III trial for AD. The compounds showed beta-amyloid-lowering effects in the in vitro assays [62]. Semagacestat is a γsecretase inhibitor for Aβ42, Aβ40, and Aβ38 with IC50 of 10.9, 12.1, and 12.0 nM, respectively. It also inhibits Notch signaling with an IC50 of 14.1 nM in H4 human glioma cells [61]. The Tg2576 mouse model is an in vivo AD model. This model carries the APP KM670/671NL mutations. Oral administration of semagacestat (1 mg/kg) to 5.5-month-old APP-transgenic Tg2576 mice significantly improved the memory deficits [63]. Recently, reports have suggested that semagacestat is a pseudoinhibitor of γ-secretase [64]. A phase III trial showed that semagacestat did not improve cognitive status and those patients receiving the higher dose exhibited significant deterioration of their functional ability. Semagacestat has more adverse effects, including cancers and skin infections (clinical trial number NCT00594568) [65].  (Figure 4a) was the first GSI to enter a phase III trial for AD. The compounds showed beta-amyloid-lowering effects in the in vitro assays [62]. Semagacestat is a γ-secretase inhibitor for Aβ42, Aβ40, and Aβ38 with IC50 of 10.9, 12.1, and 12.0 nM, respectively. It also inhibits Notch signaling with an IC50 of 14.1 nM in H4 human glioma cells [61]. The Tg2576 mouse model is an in vivo AD model. This model carries the APP KM670/671NL mutations. Oral administration of semagacestat (1 mg/kg) to 5.5-month-old APP-transgenic Tg2576 mice significantly improved the memory deficits [63]. Recently, reports have suggested that semagacestat is a pseudo-inhibitor of γ-secretase [64]. A phase III trial showed that semagacestat did not improve cognitive status and those patients receiving the higher dose exhibited significant deterioration of their functional ability. Semagacestat has more adverse effects, including cancers and skin infections (clinical trial number NCT00594568) [65].

Avagacestat
Avagacestat (BMS-708163) (Figure 4b) was expected to enter a clinical phase II trial for AD. Avagacestat is a γ-secretase inhibitor of Aβ40 and Aβ42 with IC50 of 0.3 and 0.27 nM, respectively, and exhibits 193-fold selectivity against Notch in H4 human neuroglioma cells [66]. In vitro assays revealed that avagacestat reduced CSF Aβ levels without causing Notch-related toxicities [67]. The drug is administered orally to rats and dogs in the in vivo studies, and it significantly reduced Aβ40 levels

Avagacestat
Avagacestat (BMS-708163) (Figure 4b) was expected to enter a clinical phase II trial for AD. Avagacestat is a γ-secretase inhibitor of Aβ40 and Aβ42 with IC50 of 0.3 and 0.27 nM, respectively, and exhibits 193-fold selectivity against Notch in H4 human neuroglioma cells [66]. In vitro assays revealed that avagacestat reduced CSF Aβ levels without causing Notch-related toxicities [67]. The drug is administered orally to rats and dogs in the in vivo studies, and it significantly reduced Aβ40 levels for sustained periods in the brain, plasma, and cerebrospinal fluid [66]. A phase II clinical trial revealed that avagacestat was not effective (clinical trial number NCT00890890) [68].
γ-Secretase inhibitors (GSIs) block the γ-secretase-mediated cleavage of APP. However, these GSIs also block the γ-secretase-mediated cleavage of other substrates, including the critical signaling molecule Notch. For these reasons, drugs that do not inhibit the Notch signaling cleavage of γsecretase modulators (GSMs) may be candidate treatments for AD [69]. Given the amyloid hypothesis for Alzheimer's disease and ensuing cognitive decline, several treatments, including anti-betaamyloid (Aβ) small organic molecules, antigens, and monoclonal antibodies, have been developed to interfere with Aβ production or increase clearance, including β-site beta-secretase (BACE) inhibitors, to block the first enzymatic step of the formation of Aβ or GSIs. However, these clinical trials targeting amyloid plaques of Alzheimer's disease have failed in patients with mild-to-moderate Alzheimer's disease [70]. GSIs have problems about leading to critical issues of Notch-related neuron toxicity [57,58]. Therefore, GSMs may be a better and safer strategy for drug development against AD. GSMs can act through an allosteric binding site to increase the concentrations of the shorter non-Aβ 1-42/Aβ 1-40 peptides.

γ-secretase modulators
GSMs differ from GSIs in that they do not alter total Aβ production, do not increase APP, and do not alter the generation of γ-secretase substrates but rather decrease Aβ1-42 peptides and increase Aβ1-38 peptides ( Figure 5) [71]. Nonsteroidal anti-inflammatory drugs (NSAIDs), such as aspirin, is commonly used for pain and inflammation in adults. NSAIDs selectively reduce the formation of Aβ1-42 without inhibiting the proteolysis of Notch1. The first-generation GSMs include Rflurbiprofen and indomethacin (Figure 4c-g) [72]. γ-Secretase inhibitors (GSIs) block the γ-secretase-mediated cleavage of APP. However, these GSIs also block the γ-secretase-mediated cleavage of other substrates, including the critical signaling molecule Notch. For these reasons, drugs that do not inhibit the Notch signaling cleavage of γ-secretase modulators (GSMs) may be candidate treatments for AD [69]. Given the amyloid hypothesis for Alzheimer's disease and ensuing cognitive decline, several treatments, including anti-beta-amyloid (Aβ) small organic molecules, antigens, and monoclonal antibodies, have been developed to interfere with Aβ production or increase clearance, including β-site beta-secretase (BACE) inhibitors, to block the first enzymatic step of the formation of Aβ or GSIs. However, these clinical trials targeting amyloid plaques of Alzheimer's disease have failed in patients with mild-to-moderate Alzheimer's disease [70]. GSIs have problems about leading to critical issues of Notch-related neuron toxicity [57,58]. Therefore, GSMs may be a better and safer strategy for drug development against AD. GSMs can act through an allosteric binding site to increase the concentrations of the shorter non-Aβ 1-42/Aβ 1-40 peptides.

γ-secretase modulators
GSMs differ from GSIs in that they do not alter total Aβ production, do not increase APP, and do not alter the generation of γ-secretase substrates but rather decrease Aβ1-42 peptides and increase Aβ1-38 peptides ( Figure 5) [71]. Nonsteroidal anti-inflammatory drugs (NSAIDs), such as aspirin, is commonly used for pain and inflammation in adults. NSAIDs selectively reduce the formation of Aβ1-42 without inhibiting the proteolysis of Notch1. The first-generation GSMs include R-flurbiprofen and indomethacin (Figure 4c-g) [72].

R-flurbiprofen (Tarenflurbil, MPC-7869)
R-flurbiprofen was expected to enter a clinical phase III AD trial (clinical trial number NCT00105547). R-flurbiprofen has functioned based on downregulation by neuronal tryptophan 2, 3-dioxygenase (TDO2). TDO2 expression was increased by neuronal cyclooxygenase-2 (COX-2) activities. The overexpression of TDO2 in the hippocampus led to behavioral deficits. R-flurbiprofen has prevented behavioral deficits in APPSwe-PS1∆E9 mice [73]. In clinical trials of patients with Alzheimer's disease; R-flurbiprofen had low potency and brain penetration and was considered an unsuccessful treatment [74,75]. The shortcomings of the drug were not improvement by changes in the drug delivery route or the development of nanoparticle carriers [76]. R-flurbiprofen was expected to enter a clinical phase III AD trial (clinical trial number NCT00105547). R-flurbiprofen has functioned based on downregulation by neuronal tryptophan 2, 3-dioxygenase (TDO2). TDO2 expression was increased by neuronal cyclooxygenase-2 (COX-2) activities. The overexpression of TDO2 in the hippocampus led to behavioral deficits. R-flurbiprofen has prevented behavioral deficits in APPSwe-PS1ΔE9 mice [73]. In clinical trials of patients with Alzheimer's disease; R-flurbiprofen had low potency and brain penetration and was considered an unsuccessful treatment [74,75]. The shortcomings of the drug were not improvement by changes in the drug delivery route or the development of nanoparticle carriers [76].

Indomethacin
Indomethacin was expected to enter a clinical phase III AD trial (clinical trial number NCT00432081). Indomethacin showed more inhibitory effects against COX-1 than against COX-2 [77]. Indomethacin attenuates neuronal pathological changes and cognitive impairments in APP/PS1 transgenic mice [78]. However, a clinical trial revealed that indomethacin did not improve the slow cognitive decline in patients with AD [79].
First-generation GSMs have low potency and inefficient blood-brain barrier penetrance, which affects the treatment of AD [80]. In second-generation GSMs, which include EVP-0015962, BIIB042 exhibited improved brain penetration ability [81]. GSM treatment of Alzheimer's disease is considered feasible. BPN-15606 is a highly potent GSM that is a candidate for human clinical trials [82].

EVP-0015962
EVP-0015962 (Figure 4e) decreased Aβ1-42 peptides and increased the shorter Aβ1-38 peptides, thus lowering Aβ42 in the H4 cell model. The total levels of Aβ and the amyloid precursor protein were not changed. In vivo, EVP-0015962 decreased memory deficits and reduced Aβ plaque formation in Tg2576 transgenic animals [83]. EVP-0015962 will enter a clinical phase II trial for the treatment of early AD onset (clinical trial number NCT01661673).

BIIB042
BIIB042 (Figure 4f) is a potent γ-secretase modulator that lowered Aβ42, increased Aβ38, and did not affect Notch signaling in an in vitro model [84]. Tg2576 mice overexpress the human amyloid precursor protein and serve as a research model for AD. BIIB042 reduced Aβ42 levels and Aβ plaque burden in Tg2576 mice [85]. BIIB042 is a candidate drug for AD.

BPN-15606
BPN-15606 (Figure 4g) inhibited the production of the Aβ1-42 and Aβ1-40 peptides, while facilitating the concomitant increase in Aβ1-38 and Aβ1-37 peptides production without inhibiting the γ-secretase-mediated proteolysis of Notch. BPN-15606 can lower Aβ42 levels significantly in the central nervous system of rats and mice [82]. Recent reports indicate that BPN-15606 is effective when administered prior to the severe pathological manifestations of AD [86].

Small-Molecular Combination GSM Treatments of Alzheimer's Disease
GSMs have the opportunity to become candidate drugs for the treatment of AD, but a combined treatment with drugs involving different mechanisms might provide additional efficacy for ameliorating cognitive impairment in Alzheimer's disease. Many Chinese herbal medicines are considered to have the effect of treating dementia. We list small-molecule drugs extracted from commonly used Chinese herbal medicines and their mechanisms of treating Alzheimer's disease ( Table 2). The GSM structure is shown in Figure 6.
Peroxisome proliferator-activated receptor (PPAR) agonists inhibit the expression of interleukin 6, tumor necrosis factor α, and cyclooxygenase 2 in microglial cells and monocytes stimulated by Aβ and reduce the proinflammatory and neurotoxic substances that these cells produce [108]. Astragaloside IV (AS-IV), derived from Astragalus membranaceus, is a natural PPARγ agonist that suppresses the activity of BACE1 and reduces the generation of Aβ. Activation of PPARγ in microglia by PPAR may be of value in the treatment of Alzheimer's disease and associated inflammatory diseases [109].

n-Butylidenephthalide
n-Butylidenephthalide (n-BP) is derived from Angelica sinensis. It has been applied to the treatment of neurodegenerative diseases, including Parkinson's disease [110], amyotrophic lateral sclerosis [111], and Alzheimer's disease [40]. n-BP was shown to reduce Aβ, total tau protein and its hyperphosphorylated form, and cellular toxicity in iPSCs-derived neurons induced by Down syndrome [40].

Icariside
Icariside is an extract from Herba epimedii, an ancient herbal medicine. Aggregates of Aβ are the agents responsible for disease progression in Alzheimer's disease, and the accumulation of these aggregates is toxic to the brain. The major regulatory alpha-secretase in neurons, A disintegrin and metalloproteinase 10 (ADAM10), are responsible for cleaving APP. This cleavage generates a neuroprotective APP-derived short fragment (C83) and attenuates the production of Aβ peptides [87]. The mechanism of icariside in the treatment of Alzheimer's disease involves increased ADAM10 expression and decreased expression of both APP and BACE1, resulting in reduced production of Aβ [88].

Onjisaponin
Onjisaponin B (OB) decreases Aβ by degradation of APP without direct inhibition of BACE1 and γ-secretase activities. In vivo, oral administration of OB mitigates Aβ pathology and behavioral defects in APP/PS1 mice [89,90]. OB has effects of reduced IL-1β, IL-6, TNF-α, and malondialdehyde in the serum and hippocampus of lipopolysaccharide-induced cognitive-deficit rats. OB have beneficial effects by the antioxidant, anti-inflammatory, and anti-apoptosis in the lipopolysaccharide-induced PC12 cells [91].

Bajijiasu
Bajijiasu can inhibit neuronal inflammation, increase the levels of insulin degradation enzymes and neprilysin, and decrease BACE1 expression [92]. Bajijiasu have a function of neuron-protectiveness against Aβ-induced neurotoxicity in PC12 cells [93]. Animal studies also found the same function. In APP/PS1 mice models, oral of bajijiasu improved learning and memory ability. The effects of bajijiasu have protected neurons against apoptosis, elevated the expression of neurotrophic factors, and alleviated endoplasmic reticulum stress [94].

Asarones
Asarones, the active constituents of Acori Tatarinowii Rhizoma, can improve the proliferation and self-renewal of neural progenitor cells. A mechanistic study showed that asarones activated extracellular signal-regulated kinase (ERK), part of a critical kinase cascade for neurogenesis [95]. Asarone have decreased Aβ deposition in the cortex and hippocampus of APP/PS1 mice brains [96].

Catalpol
There is a significant positive correlation between brain-derived neurotropic factor (BDNF) content and memory. Catalpol can improve memory and reduce neurodegeneration by increasing BDNF expression [97]. In vitro, catalpol has been able to reduce the levels of soluble Aβ40 and Aβ42 in the cerebral cortex, and thus, inhibit the formation of senile plaques [98].

Liquiritin
Liquiritin have been extracted from Glycyrrhiza uralensis. Oxidative stress plays an important role in the pathogenesis of Alzheimer's disease. Histopathological hallmarks of Alzheimer's disease include formation of senile plaques composed of the Aβ peptides. Aβ can aggregate in formation with metal ions to catalyze production of reactive oxygen species [99]. Excessive oxidative stresses are generally regarded as a pathological condition. Malondialdehyde (MDA) and 8-hydroxy-2 deoxyguanosine (8-OHdG) are products of lipid peroxidation and are the most commonly detected biomarkers of oxidative stress [100]. Liquiritin reduces MDA and 8-OHdG in the hippocampus of rats with Alzheimer's disease and, through related mechanisms, can significantly improve spatial learning and memory [101].

Glabridin
Deterioration of cholinergic innervation in the human cerebral cortex leads to significant cognitive and behavioral disorders in Alzheimer's disease [102]. Glabridin, a small-molecule drug studied as a potential treatment for Alzheimer's disease, has anticholinesterase activity. Glabridin reduced the brain cholinesterase activity and memory improvement in amnesia mice. The U.S. Food and Drug Administration approved cholinesterase inhibitors for the treatment of Alzheimer's disease [103].

Tanshinone IIA and Cryptotanshinone
Tanshinone IIA (TIIA) and cryptotanshinone (CT) are extracted from Salviae miltiorrhizae as major lipophilic compounds. In vitro, TIIA and CT significantly reduced intracellular NO and Ca 2+ levels in H9c2 cells [104]. The gliosis-related and neuro-inflammatory markers in the hippocampal tissues revealed a remarkable reduction in the expression of glial fibrillary acidic protein (GFAP), S100β protein, COX-2, inducible nitric oxide synthase (iNOS), and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kBp65) after CRY treatment. TIIA and CT display anti-inflammatory and neuroprotective effects in a nongenetic mouse model of AD [105].

Ginsenoside Rg1
Ginsenoside Rg1 have been extracted from Panax ginseng. Ginsenoside Rg1 repaired hippocampal long-term potentiation and memory in an Alzheimer's disease mouse model. Through a mechanism that includes increased activity of the antioxidant enzymes glutathione peroxidase and superoxide dismutase and reduces astrocyte activation, it can increase the proliferation of hippocampal cells [106,107].

Astragaloside IV
Peroxisome proliferator-activated receptor (PPAR) agonists inhibit the expression of interleukin 6, tumor necrosis factor α, and cyclooxygenase 2 in microglial cells and monocytes stimulated by Aβ and reduce the proinflammatory and neurotoxic substances that these cells produce [108]. Astragaloside IV (AS-IV), derived from Astragalus membranaceus, is a natural PPARγ agonist that suppresses the activity of BACE1 and reduces the generation of Aβ. Activation of PPARγ in microglia by PPAR may be of value in the treatment of Alzheimer's disease and associated inflammatory diseases [109].

n-Butylidenephthalide
n-Butylidenephthalide (n-BP) is derived from Angelica sinensis. It has been applied to the treatment of neurodegenerative diseases, including Parkinson's disease [110], amyotrophic lateral sclerosis [111], and Alzheimer's disease [40]. n-BP was shown to reduce Aβ, total tau protein and its hyperphosphorylated form, and cellular toxicity in iPSCs-derived neurons induced by Down syndrome [40]. Table 2. The role and mechanisms of small-molecule drugs treated in the Alzheimer's disease model.

Extracts
Extracted from/ Chinese Name

Drug Doses & Experimental Model Possible Molecular Mechanism Effect Reference
Icariin II Herba Epimedii/淫羊藿 APP/PS1 transgenic mice were treated orally g icariside II 30 m/kg. Effectively ameliorated cognitive function deficits, but also inhibited neuronal degeneration and reduced the formation of plaque burden. [88] Onjisaponin B Radix polygalae/遠志 APP/PS1 transgenic mice were treated orally 200 µL of onjisaponin B 1 mg/mL.
Reductions of Aβ deposition and senile plaques and have higher levels of neurotrophic factors and inhibitory function on neuroinflammation in the brains of APP/PS1 mice. [92] Asarones Rhizoma Acori tatarinowii/石菖蒲 Adult hippocampal neural progenitor cells (NPC) cultures treated with 1 µM asarones.
The drugs improve Aβ1-42-induced spatial learning and memory impairment through inhibiting oxidative stress and neural apoptosis. [101]
That TIIA and CT display anti-inflammatory and neuron-protective effects in a nongenetic mouse model of AD. [105]. Cryptotanshinone
It prevents cognitive impairment and hippocampus senescence of D-galactose-induced aging in a rat model. [107] Astragaloside IV Astragalus membranaceus/黃耆 APP/PS1 mice that were treated orally AS-IV 10 mg/kg.

Conclusions
Recent studies report tau pathology and neurodegeneration are enhanced, but amyloid pathology is generally unaffected by concurrent tau pathology [113][114][115]. Through iPSCs, models get verified [116]. Using iPSCs that naturally produce Alzheimer's disease symptoms to explore the mechanisms of small-molecule drugs and combining this approach with GSM or other therapies is a strategy for developing Alzheimer's drugs. At the same time, using an iPSC platform for drug screening may be a good way to develop personalized medicine. Individual differences are perhaps the main cause of setbacks to the development of Alzheimer's drugs. Establishing the patient's iPSC to confirm the symptoms first and then accurately selecting the appropriate therapeutic drug may be a good way to provide effective treatment to patients while reducing the rate of failure of drug development. Through cryo-electron microscopy (cryo-EM) analysis, the structures for γ-secretase acting on the APP-C83 fragment and Notch-100 fragment will help develop substrate-specificity secretase inhibitors or modulators. Many small molecules extracted from Chinese herbal medicines, such as Curcumin, Asarones, and Astragaloside, have been used for the treatment of dementia [117][118][119]. However, these treatments have not achieved the expected results; thus, combinations of drugs with different mechanisms may be necessary. An increase in the number of choices may increase the chances of success in fighting Alzheimer's disease (Figure 7).

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.