Generation and Characterization of a Novel Knockin Mouse Model Expressing PSEN1 D385A: Implications for Investigating Herbal Drug Effects in γ-Secretase Activity

Abstract

Background:

Presenilin (PSEN, PS) is essential for γ-secretase function, and mutations can disrupt amyloid-β (Aβ) production in familial Alzheimer’s disease. Targeting γ-secretase is complex due to its broad involvement in physiological processes.

Objective:

Our aim was to create a novel knockin (KI) mouse model expressing PSEN1 D385A mutation and investigate the efficacy of a Geniposide and Ginsenoside Rg1 combination (NeuroProtect modified formula, NP-2) in restoring γ-secretase activity.

Methods:

Using gene manipulation, we established the PS1 D385A KI mouse model and confirmed the mutation, mRNA, and protein levels using Southern blotting, northern blotting, and western blotting, respectively. In vitro γ-secretase assay was conducted to measure γ-secretase activity, while histological analyses examined neurogenesis effects. NP-2 administration evaluated its impact on γ-secretase activity.

Results:

The PS1 D385A KI homozygotes displayed severe cerebral hemorrhage, postnatal lethality, developmental disorders, reduced proliferation of neural progenitor cells, and disrupted γ-secretase function. The mutation abolished PS1 protein self-shearing, leading to compromised γ-secretase activity. NP-2 intervention effectively restored γ-secretase activity in the heterozygous mice.

Conclusions:

PS1 D385A mutant disrupted PS1 protein self-cleaving, impairing γ-secretase activity in KI mice. NP-2 restored γ-secretase function, offering potential for novel AD treatment strategies despite the challenges posed by γ-secretase’s complex role in physiological processes.

Keywords

Alzheimer’s disease amyloid-β D385A geniposide ginsenoside Rg1 presenilin γ-secretase

INTRODUCTION

Presenilin gene (PS for protein, PSEN for DNA, Psen for mRNA) is the primary genetic factors in familial Alzheimer’s disease (FAD) [1]. Mutations in the PSEN gene account for approximately 90% of FAD cases [2]. The Presenilin-1 (PS1) protein is a key component of γ-secretase, an enzyme involved in the cleavage of type I transmembrane proteins such as amyloid-β protein precursor (AβPP), Notch, and N-cadherin [3]. The aspartate residues within the transmembrane domains of PS1 are critical for the catalytic activity of γ-secretase [4]. Mutations in these aspartate residues, particularly the D257 and D385 within PS1, have been found to disrupt the normal processing of the AβPP and lead to the accumulation of amyloid-β (Aβ) peptide fragments, which are the major components of amyloid plaques in Alzheimer’s disease (AD) brains [5, 6].

To elucidate the function of the aspartate mutation in presenilin, specifically to the D385A mutation, we generated PS1 D385A knockin (KI) mice to introduce the specific aspartate-to-alanine mutation at position 385 in the PSEN1 gene. This genetic modification results in born lethal with impaired skeletal development, brain hemorrhage, neuronal progenitor cell reduction in homozygotes, and abnormal cerebrovascular structure, amyloid vascular plaques in the adult heterozygotes [7]. D385A mutation disrupted PS1 protein self-shearing activity, and altered γ-secretase function. This PSEN1 D385A KI mice provides a valuable model for studying the functional consequences of presenilin in AD.

Given the involvement of γ-secretase in the production of Aβ peptides and neurogenesis, it has been explored as a potential therapeutic target for AD [8]. Several γ-secretase inhibitors and modulators have been developed and investigated for their potential to reduce Aβ production and mitigate AD-related pathology [9]. However, the development of γ-secretase-targeted drugs for AD has faced significant challenges [10], since γ-secretase is involved in the proteolytic processing of numerous substrates beyond AβPP, including Notch receptors, which play critical roles in cellular signaling and development [11].

In the recent years, natural products and herbal medicines have been explored for their potential in treating AD and modulating γ-secretase activity with safe [12 –15]. Previously, we reported that a traditional Chinese medicine formula— NeuroProtect (NP), the combination of Geniposide (GP) and Panax notoginseng saponins (PNS), improved learning memory impairment in multiple AD animal models [16, 17], and their main bioactive ingredient Ginsenoside Rg1 (Rg1) modulated the protein levels of γ-secretase [18]. Using this novel KI mouse model, we further illustrated the potential mechanism of GP and Rg1 combination (NeuroProtect modified formula, NP-2) in control of γ-secretase activity.

METHODS

Generation of PS1 D385A Mice

A mouse PS1 D385A mutant plasmid targeting vector (PGKneolox2DTA plasmid vector) was constructed using polymerase chain reaction (PCR) amplification and targeted mutagenesis techniques. Initially, a fragment spanning from exon 8 to exon 12 was amplified from the mouse PSEN1 gene (RP23-330F11, Children’s Hospital Oakland Research Institute, Boston, MA, USA) via PCR, resulting in a 2.2-kb left arm fragment and a 6.7-kb right arm fragment. Subsequently, the 6.7-kb right arm fragment was mutated from aspartic acid to alanine at position 385 using a point mutation kit. The mutated right arm fragment and the 2.2-kb left arm fragment were then ligated into the plasmid PGKneolox2DTA. The targeting vector was single-digested to generate single-stranded fragments, which were transfected into MKV6.5 embryonic stem cells obtained from a cross between C57BL/6 and 129 strains. Electrotransfection was utilized to introduce the targeting vector into the embryonic stem cells, aiming to promote the integration of the targeting vector and subsequently screen for cells that underwent the desired homologous recombination events. Confirmation of correct targeting was achieved through Southern analysis. Once the embryonic stem cells carrying the desired genetic modifications were confirmed, they were injected into blastocysts. The resulting chimeric mice were bred, and selected male chimeras were crossed with C57BL6/J-129 mice. Positive mice in the first generation were identified and further bred with Camk2a-Cre transgenic mice to excise the screening gene PGK-neo from male germ cells, derived from the target plasmid. Southern analysis was performed to verify the success of genetic manipulation and confirm the intended genetic modifications. For subsequent experiments, PS1 D385A heterozygous (KI/+) mice, PS1 D385A heterozygous (KI/KI) mice, and wild-type (+/+) mice were utilized. Animal experiments were approved by the Animal Ethics Committee of the Beijing University of Chinese Medicine (BUCM-4-2016103101-1008).

Southern blot analysis

Genomic DNA was extracted from either embryonic stem cells or mouse tail samples measuring 1–3 mm in length. To ensure complete dissolution of DNA, 50–100μl TE buffer was utilized, and the DNA concentration and purity were assessed using a NanoDrop spectrophotometer (Amersham, ND-1000). BamHI and ApaI restriction endonucleases were employed to fragment the genomic DNA, which was subsequently transferred overnight onto an HN membrane (Amersham, RPN303N) following 0.7% agarose gel (Invitrogen, 16500-100) electrophoresis. The HN membrane was crosslinked under ultraviolet light (UV) light, then immersed in 2XSSC solution (SB, 19629) and agitated at 60 rpm for 2–3 min. Simultaneously, a hybridization cartridge containing 30 ml of 2XSSC was preheated to 50°C. The HN membrane was rolled onto a nylon mesh, inserted into the preheated hybridization cartridge at 50°C for 15 min. After removing the 2XSSC solution, 15 ml of hybridization closure solution was added at 42°C for 3–5 h.

Amplification and purification of isotope phosphorus 32-labeled probes (PerkinElmer, BLu513Z500uC) were performed in advance. 150μl of the probe was thermally denatured at 95°C for 10 min, followed by a quick transfer to an ice bath for 5 min before being addition to the hybridization cartridge. The Hybridization process was conducted at 50°C for 14–24 h. Subsequently, the membrane was washed and exposed for 1–3 days for imaging purposes.

For this Southern assay, primers for the 5’ and 3’ ends of the PSEN1 gene fragment probes were prepared as follows. For 5’ end, forward primer: 5’-AACAGGCTCTCTCCCTAGTAAG-3’, reverse primer: 5’-GCACTACATACATGTAGGCC-3’; for 3’ end, forward primer: 5’-TCACAAGACATGGACCATCG-3’, reverse primer: 5’-AGGTTCAAGGTCATCCTTGG-3’.

Northern blot analysis

Fetal mouse brains from embryonic day 18.5 (E18.5) were weighed, and total RNA was extracted. To dissolve the RNA, 50–100μl DEPC-treated water (Sigma, D5758) was utilized, and the RNA concentration and purity were assessed using a NanoDrop spectrophotometer. The RNA samples were denatured and separated by 1.2% agarose gel electrophoresis to fractionate the RNA molecules. Subsequently, the RNA was transferred to a HN membrane (GE Healthcare, RPN303N) and cross-linked using intense UV light.

The HN membrane was immersed in 2XSSC solution and agitated at 60 rpm for 2–3 min. It was then briefly stained with methylene blue dye (Fluke, 50484) for 10 s and washed three times with DEPC-treated water for 10 min each. The membrane was visually inspected to confirm the presence of two labelled RNA bands, namely 18 S and 28 S. Meanwhile, a hybridization cartridge containing 30 ml of 2XSSC was preheated to 50°C, and the HN membrane was positioned on the cartridge.

Amplification and purification of isotope phosphorus 32-labeled probes (PerkinElmer, BLu513Z500uC) were performed in advance. 150μl of the purified probe was thermally denatured at 95°C for 10 min, followed by rapid cooling in an ice bath for 5 min before being added to the hybridization cartridge. The reverse primer sequences used were as follows: for PSEN 1, forward primer: 5’-TGCACCTTTGTCCTACTTCC-3’, Reverse primer: 5’-ATGACACTGATCATGATGGC-3’; for GAPDH, forward primer: 5’- ACCACAGTCCATGCCATCAC-3’, reverse primer: 5’-TCCACCACCCTGTTGCTGTA-3’. The hybridization process was conducted at 50°C for 14–24 h. Following hybridization, the membrane was washed and subsequently exposed for 1–3 days for imaging purposes.

Western blot analysis

Whole fetal mouse brains were harvested, weighed, and total protein was extracted from the brain tissue. The protein concentration was determined using a bicinchoninic acid (BCA) protein quantification kit (Thermo Scientific, 23227). The extracted proteins were denatured and subjected to electrophoresis. During electrophoresis, an antioxidant solution (Novex, NP0005) was added at a concentration of 1 ml/L of running buffer (NuPAGE, NP0002). The proteins were subsequently transferred onto a membrane. Following electrophoresis, the membrane was blocked using 5% skimmed milk and incubated at room temperature for 1 h. Subsequently, the membrane was probed with a primary antibody at a dilution of 1 : 500–1 : 5,000. After incubation, the membrane was washed three times with PBST (phosphate-buffered saline with Tween-20). For detection, a secondary antibody was applied at a concentration of 1 : 10,000. The membrane was incubated with the secondary antibody and then washed three times with PBST. Finally, the protein bands on the membrane were visualized by exposure to a film using a film sweeper (LI-COR, Odyssey). Details of the antibodies utilized in this experiment can be found in Table 1.

Table 1

Antibodies in western blotting experiments

Antibody name	Dilution	Vendor	Product number
Aph1a	1 : 1000	Zymed	38–3600
Aph1b	1 : 500	Abcam	ab116657
AβPP	1 : 4000	Sigma	A8717
β-actin	1 : 5000	Sigma	A5060
CD31	1 : 1000	Abcam	ab9498
N-cadherin	1 : 2500	BD Biosciences	610920
Nicastrin	1 : 800	Sigma	N1660
NICD	1 : 1000	Cell signaling	2421
Notch1	1 : 1000	Cell signaling	3608
Pen2	1 : 4000	Zymed	36–7100
Presenilin 1 CTF	1 : 1000	EMD Millipore	PC267
Presenilin 1 NTF	1 : 1000	EMD Millipore	PC244
Goat anti-rabbit 800CW	1 : 10000	LI-Cor	926–32211
Goat anti-mouse 800CW	1 : 10000	LI-Cor	926–32210
Goat anti-rabbit 680RD	1 : 10000	LI-Cor	926–68071
Goat anti-mouse 680LT	1 : 10000	LI-Cor	926–68020

γ-secretase substrates preparation

The preparation of γ-secretase substrates was carried out by culturing Escherichia coli (E. coli) overexpression C99 or Notch ΔE FLAGmycHis (N102-FmH) substrate. Initially, a single colony was carefully picked and inoculated it into 2 ml of lysogeny broth (LB) medium, then shaken at 250 rpm and incubated at 37°C for 3– 6 h. Subsequently, the 2 ml bacterial culture was transferred into 30–50 ml of LB medium, shaken at 250 rpm, and cultured at 37°C overnight. The following day, the bacterial culture was transferred into 250 ml of LB medium, shaken at 250 rpm, and incubated at 37°C until the OD600 value reached between 0.35 and 0.5.

To induce protein expression, IPTG was added to a final concentration of 0.1 mM, and cultivation under the same conditions continued for 3 h. After cultivation, the culture was centrifuged, the bacterial pellet was collected at 3,000 g for 20 min, and the supernatant was discarded. The pellet was then resuspended in 10 ml of PB solution containing protease inhibitors (Complete Mini (EDTA-free), Roche, 11836170001) with thorough shaking. Lysozyme (Sigma, L6876) was added to a final concentration of 0.3 mg/ml, and the solution was incubated at 4°C for 30 min. Sonication was performed 2–5 times for 30 s each. Subsequently, DNaseI was added and incubated at 4°C for 30 min. 1% TritonX-100 was introduced and incubated at 4°C for 1 h.

Afterward, a cold centrifugation was performed at 4°C, 30,000 rpm for 45 min using a SW41T rotor. The supernatant was collected, filtered through a 0.45μm filter, and prepared for protein purification. Before purification, the purification column (Amersham, 17-0851-01) was set up, and the solution was passed through with a series of different reagents including deionized water, 50 millimolar nickel sulfate (Sigma, 31483), and 50 millimolar imidazole (Sigma, 15513).

Subsequently, the purified protein solution was run through the column, eluted with varying concentrations of elution buffer, and varying concentrations of elution buffer, and fractions were collected as follows: filtrate (fraction A), 5μM imidazole (fraction B), 50μM imidazole (fraction C), 100μM imidazole (fraction D), 300μM imidazole (fraction E), 50μM EDTA (fraction F). 10μl of each eluted fraction (labeled A–F) was extracted for western blot analysis. The eluates with the highest protein concentration and purity were selected for cold salt precipitation at 4°C for 2–3 h or overnight. After precipitation, the purified proteins were aliquoted, stored at –80°C, and the protein concentration was determined using a BCA protein quantification kit (Thermo Scientific, 23227) following the manufacturer’s instructions. Finally, the molar concentration of the purified protein was calculated based on the quantification results.

γ-secretase assay extraction

The γ-secretase activity was evaluated using a cell-free in vitro assay that involved CHAPSO-solubilized brain fractions and bacterially expressed recombinant proteins as substrates [19]. Fresh brain tissue was collected and homogenized in 1 ml of homogenization buffer (20 mM PIPES, pH 7.0, 140 mM KCl, 0.25 M sucrose, 5 mM EGTA), followed by extensive grinding and centrifugation at 800 g for 10 min at 4°C. The pellet was then collected and suspended in a 0.1 M sodium bicarbonate solution (pH 11.3), and centrifuged at 100,000 g for 1 h at 4°C. After discarding the supernatant, the pellet was treated with 1% CHAPSO solution (50 mM PIPES, pH 7.0, 0.25 M Sucrose, 1 mM EGTA) at a ratio of 0.25μl per 1 mg of tissue. Following a 1-hour incubation on ice, the sample was centrifuged at 100,000 g for 1 h at 4°C. The resulting supernatant was collected for subsequent in vitro experiments.

In vitro γ-secretase assay

For the γ-secretase assay, the enzyme was incubated with the C99 or N102-FmH substrate for 16 h at 37°C in a reaction system as follow:

Reagents	Volume or concentration
γ-secretase in 1%	15μl (0.3%
CHAPSO solution	CHAPSO solution)
C99 or N102-FmH	1–2μM
2 X Assay buffer	25μl
20% Phosphatidylcholine	0.25μl
Deionized water	Up to 50μl

The 2X Assay buffer contained 20 mM HEPES, pH 7.3, 300 mM NaCl, 10 mM EDTA, 2X Complete, 20μg/ml Phosphoramidon, 10 mM 1,10-phenanthroline. The reaction was terminated by centrifugation at 95°C for 5 min, and the supernatant was collected by centrifugation at 4°C for 10 min at maximum speed.

Aβ₄₀ and Aβ₄₂ generated in the in vitro assays were quantified using the 11A50-B10 (Covance SIG-39140)/4G8 (Covance, SIG-39240) and 12F4 (Covance, SIG-39142)/4G8 sandwich ELISA, respectively [20]. On the first day, the capture antibody (20μl/well) was coated in a 384-well plate (Corning, 3577) as follows:

Assay	Capture antibody	Coating concentration
Aβ₄₀	11A50-B10	2μg/ml
Aβ₄₂	12F4	5μg/ml

The plate was tightly sealed with a plate sealing film and incubated overnight at 4°C. On the second day, the antibody solution was aspirated, and the wells were blocked with 40μl of 1% BlockAce (AbdSerotec, BuF029, diluted in PBS) at room temperature for 2 h. Standards and samples were combined with 4G8 antibody solution (diluted 1 : 2,000 for Aβ₄₀ and 1 : 200 for Aβ₄₂) on ice, added to the plate, and incubated overnight at 4°C. The standards were diluted in a gradient as shown below:

Dilution	Concentration	Standard	Standard
	(ng/ml)	Aβ₄₀ (pM)	Aβ₄₂ (pM)
1 : 400	2.5	576	554
1 : 2	1.25	288	277
1 : 2	0.625	144	138.5
1 : 2	0.32	72	69.3
1 : 2	0.16	36	34.6
1 : 2	0.08	18	17.3
1 : 2	0.04	9	8.7
1 : 2	0.02	4.5	4.3
1 : 2	0.01	2.3	2.2

On the third day, after washing with PBST for three times, Streptavidin (Promega, V5591, diluted 1 : 5,000 in PBS) was added and incubated at room temperature for 1 h for Aβ₄₀ and 2 h for Aβ₄₂. Following another wash with PBS three times, the AttoPhos reaction reagent was sequentially added and incubated for 3–5 min. The luminescence was measured and analyzed.

Immunohistochemistry

Fetal mouse brains from E12.5 and E16.5 were carefully collected and divided into the left hemisphere for western blotting and the right hemisphere for immunochemistry. The cerebral cortex and hippocampus of the right hemisphere were meticulously separated and fixed in 4% paraformaldehyde at 4°C for 3 h. Subsequently, the fetal mouse brains were dissected, embedded in paraffin, and transverse sections were serially cut at 10μm thickness. These sections were then dewaxed, stained with Hematoxylin & Eosin (HE) (Fisher Scientific, SE23-500D, SE26-500D), dehydrated, and sealed with neutralresin.

To label the dividing cells, BrdU (Sigma, B5002) was intraperitoneally injected 30 min prior to tissue retrieval. Paraffin sections of the brain were also made, each 5–8μm thick. The number of neuronal cells exhibiting BrdU-positive signals was counted using immunohistochemical analysis. Antigen restoration was performed on the paraffin sections to be stained, and the sections were incubated at room temperature for 1 hour using a blocking solution containing 0.1% TritonX-100, 5% goat serum (Vector, S-1000), and 1% bovine serum (Sigma, A3294). The section was then incubated with specific antibodies, followed by color development using 3,3’-diaminobenzidine (Vector, SK-4100). Finally, the section was sealed by dehydration, and images were captured using an inverted microscope (Olympus, CKX41) for further analysis.

Drug administration

PS1 D385A heterozygous (KI/+) mice and wild-type (+/+) mice from the same litter were raised for the experiment. Gavage administration of the drug started when the mice reached 9 months of age and continued for a duration of 1 month. In this experiment, Rg1, the main active ingredient of Panax notoginseng, and GP, the active ingredient of Gardenia jasminoides Ellis, were combined in the ratio of GP:Rg1 = 5 : 1 [7 , 21]. The dosage administered was Rg1 (purity 98%) (Sigma, 00370580) at a concentration of 3.85 mg/kg/d and geniposide (purity 98%) (Sigma, SML0153) at a concentration of 18.68 mg/kg/d.

The above drugs were dissolved proportionally in saline and mixed thoroughly into a clarified liquid. The +/+ blank control group (5 male mice) and the KI/+ model group (4 male mice) were administered saline via gavage. The KI/+ dosing group (4 male mice) received the combination of geniposie and Rg1 (NP-2) via gavage.

Statistical analysis

All data are presented as mean±SEM (standard error of the mean). Statistical analyses were performed using Prism software. Paired or unpaired Student’s t-test were used for comparing means between two groups, while one or two-way analysis of variance was employed for multiple group comparisons. Post-hoc tests such as the Tukey test, Scheffe’s test, or Bonferroni test were applied as appropriate to determine significant differences between groups. The significance level for all analyses was set at p≤0.05, and two-sided tests were conducted to assess statistical significance.

RESULTS

Establishment of a KI mouse model for the PS1 D385A mutation

A plasmid containing the D385A mutation gene was genetically engineered and transfected into MKV6.5 embryonic stem cells via electrotransfection. The schematic diagram of the wild-type allele and the constructed targeting vector is illustrated in Fig. 1A, with vertical dashed lines outlining the PSEN1 genomic sequences integrated into the targeting vector. The D385A mutation is situated in exon 11. 5’ and 3’ probes, 0.5 kb fragments utilized for screening homologous recombinants, are positioned outside of the 5’ and 3’ arms as indicated by gray boxes (Fig. 1A). Genomic DNA was extracted and digested with restriction endonucleases BamHI and BamHI/ApaI, respectively. The 5’ end fragment was cleaved by BamHI, resulting in 22.8 kb and 2.8 kb fragments representing the wild type (WT) and the target gene (T), respectively; the 3’ end fragment was cut by BamHI/ApaI, resulting in 9.9 kb and 8.6 kb fragments representing the WT and the target gene (T), respectively (Fig. 1B).

Fig. 1

Molecular characterization and phenotypic features of PS1 D385A KI mouse model. A) Schematic representation of the targeting strategy used to generate the PS1 D385A KI mouse model. B, C) Southern analysis of genomic DNA isolated from ES cells (B) and tails (C) to confirm correct targeting. D) Genomic sequencing analysis of tail genomic DNA validating the A to D substitution in D385A KI/KI mice. E) Lateral view of D385A KI/KI embryos at E18.5 displaying phenotypic characteristics such as hemorrhage, shortened tail, and altered rostrocaudal body axis.

Positive clones were microinjected into C57BL/6 mouse blastocysts, and PS1 D385A KI mice were generated by mating to eliminate the screening gene PGK-neo from the male germ cells for tail genomic DNA assessment (Fig. 1A). Results indicated that the 5’ end fragment, cleaved by BamHI, generated two fragments of 22.8 kb and 2.8 kb, representing the WT and the target gene (T) respectively; the 3’ end fragment, cut by BamHI/ApaI, yielded two fragments of 9.9 kb and 6.9 kb, representing the WT and the target gene (T) respectively (Fig. 1C).

Gene sequencing confirmed that in +/+ mice, the PSEN1 gene featured three codons of AGA for aspartate (D); in KI/KI mice, the corresponding three codons were AGC for alanine (A) at the same position. The black arrow denotes the mutation site (Fig. 1D). Moreover, the observation of postnatal phenotypes in PS1 D385A KI mice reveals that KI/KI mice were not viable, presenting impaired skeletal development and short tails (Fig. 1E). Brain hemorrhage was visibly observed in the head (Fig. 1E), and the phenotype was consistent with PS1 knockout (– /–) mice [22].

The D385A mutation does not impact Psen1 mRNA expression levels but disrupts PS1 protein self-cleavage activity

Using the northern blot method, Psen1 mRNA levels were assessed in PS1 D385A mice. The analysis indicated that mRNA levels were similar to those of wild-type (+/+) and heterozygous (KI/+) littermates, showing no significant differences (Fig. 2A, Supplementary Figure 1). This suggests that the D385A mutation has no effect on Psen1 mRNA expression levels.

Fig. 2

Impact of the D385A mutation on PS1 mRNA and protein processing. A) PSEN1 mRNA levels in both KI/KI and KI/+ embryos at E18.5 were comparable to their wild-type littermate controls. Representative blotting images (A1) and corresponding statistical analysis (A2) are shown. B) At E18.5, the expression of PS1 FL protein was significantly elevated in both KI/KI and KI/+ embryos, up to 16-fold and 11-fold higher, respectively, compared to wild-type littermates. Notably, PS1 NTF and CTF were undetectable, similar to PS1– /– mice. Representative western blotting images (B1) and statistical analysis (B2) are shown. C) even with a prolonged exposure compared to PS1– /– brains, residual PS1 NTF or CTF was not observed. D) At E14.5, KI/KI and KI/+ embryos exhibited a significant increase in PS1 FL protein expression compared to wild-type littermates. PS1 CTF and NTF were not prominently detected in KI/KI embryos. Representative western blotting images (D1) and statistical analysis (D2) are included. Data is presented as mean±SEM. ^***p < 0.001, ^**p < 0.01.

Embryos at day E18.5 (Fig. 2B) were collected, and whole brains were extracted for western blot analysis. The results revealed a notable increase in the expression of full-length PS1 protein (PS1 FL) in PS1 D385A KI/KI and KI/+ mice, which was 18 and 16 times higher, respectively, than in the +/+ littermates. Additionally, there was a significant decrease in the expression of PS1 C-terminal fragments (PS1 CTF) and N-terminal fragments (PS1 NTF) in KI/+ mice compared to +/+ littermates. PS1 CTF and PS1 NTF were not detectable in KI/KI, with PS1– /– mice used as controls [22].

Moreover, even with increased exposure intensity (Fig. 2C), PS1 CTF and PS1 NTF remained undetectable in KI/KI mice compared to PS1– /–. These findings suggest that the D385A mutation results in the loss of PS1 protein self-shearingfunction.

Similarly, a significant elevation in PS1 FL levels in PS1 D385A KI/KI and KI/+ embryos was observed at E14.5 (Fig. 2D). In this analysis, PS1 CTF and PS1 NTF showed significantly lower expression levels in KI/+ mice compared to +/+ littermates.

Loss of γ-secretase activity due to the D385A mutation

AβPP protein serves as a substrate of γ-secretase, which normally cleaves AβPP CTFs to generate Aβ₄₀ and Aβ₄₂. Western blot analysis indicated that the D385A mutation did not impact the expression of the γ-secretase substrate AβPP FL protein. However, there was a significant increase in the levels of AβPP CTFs in KI/KI mice (Fig. 3A).

Fig. 3

D385A mutations abolished γ-secretase-dependent cleavage of AβPP, Notch, and N-cadherin, leading to changes in the content of the different catalytic subunits of γ-secretase. A) Expression of AβPP CTFs in KI/KI at E18.5 or P0 is significantly elevated compared to wild-type littermate controls. Representative western blotting images (A1) and statistical analysis (A2) are shown. B) Expression of N-cadherin CTF1 in KI/KI at E18.5 or P0 is significantly increased compared to wild-type littermate controls. Representative western blotting images (B1) and statistical analysis (B2) are shown. C) In vitro γ-secretase assay using CHAPSO-solubilized E18.5 whole brain fractions and recombinant Notch substrate N102-FmH followed by western analysis. NICD production is reduced in KI/+ mice and abolished in KI/KI mice. SNAP25 is used as internal loading control for membrane fractions. Representative western blotting images (C1) and statistical analysis (C2) are shown. D) ELISA measurements of Aβ₄₀ and Aβ₄₂ following in vitro γ-secretase assay that de novo generation of Aβ₄₀ and Aβ₄₂ in KI/+ brains at E18.5 is reduced by 42.4% and 45.8%, respectively. E, F) The content of the catalytic subunit of γ-secretase was examined at E18.5 and E14.5, respectively. Significant differences in Aph-1b and Pen-2 were observed in KI/KI as early as E14.5, while Nicastrin showed significant differences at E18.5. KI/+ showed significant differences in Pen-2 as early as E14.5, and Aph-1b also showed a significant difference at E18.5. Representative western blotting images (E1 and F1) and statistical analysis (E2 and F2) are shown. Data are represented as mean±SEM. ^***p < 0.001, ^**p < 0.01, ^*p < 0.05.

N-cadherin, another substrate of γ-secretase, undergoes cleavage to produce CTF1 from N-cadherin under normal condition [23]. Western analysis showed that the D385A mutation had no effect on the expression of γ-secretase substrate N-cadherin. Nevertheless, there was a notable elevation in the levels of N-cadherin CTF1 in both KI/+ and KI/KI mice (Fig. 3B).

The Notch substrate N102-FmH was utilized to assess γ-secretase activity, with the detection of Notch intracellular domain (NICD) expression through western assays. The expression of NICD was reduced by 41% in KI/+ mice compared to +/+ littermates, and no NICD expression was observed in KI/KI mice. SNAP25, a membrane protein, served as an internal control in this analysis (Fig. 3C).

In addition, using C99 as a substrate for γ-secretase, the production of Aβ₄₀ and Aβ₄₂ was quantified via ELISA assays. In KI/+ mice, there was a significant reduction of 48% and 43% in Aβ₄₀ and Aβ₄₂ production, respectively, compared to +/+ littermates. Notably, Aβ₄₀ and Aβ₄₂ were undetectable in KI/KI mice (Fig. 3D).

The levels of γ-secretase subunits were examined at E18.5 (Fig. 3E) and E14.5 (Fig. 3F), with notable changes observed in Aph-1, Nicastrin, and Pen-2.

These experimental findings suggest that the D385A mutation can lead to a loss of γ-secretase function in vivo. In KI/KI mice, a complete loss of γ-secretase activity was evident, while KI/+ mice exhibited a reduction of approximately 45% in γ-secretase activity.

Severe cerebral hemorrhage and reduced proliferation of neuronal progenitor cells were observed during embryonic development in PS1 D385A KI/KI mice

PS1 D385A KI/KI mice displayed severe brain hemorrhage at birth, visually confirmed. To investigate the timing of brain hemorrhage in KI mice, embryos were collected from day E11.5 to P0 (Table 2). At E11.5, no cerebral hemorrhage was detected in KI/KI mice; by E12.5, 34% of KI/KI mice exhibited cerebral hemorrhage; and by E13.5, all KI/KI mice showed cerebral hemorrhage. Subsequently, an expanding area of cerebral hemorrhage was observed in KI/KI mice, and 100% affected (Table 2).

Table 2

The D385A alteration causes severe hemorrhage as early as E12.5

Age	Hemorrhage observed in KI/KI	Non-hemorrhage observed in KI/KI	Hemorrhage percentage in KI/KI
E11.5	0	4	0%
E12.5	2	4	34%
E13.5	3	0	100%
E14.5	9	0	100%
E16.5	8	0	100%
E18.5	26	0	100%
P0	8	0	100%

To further assess the location of the brain hemorrhage, its distribution, and its impact on brain structure, embryos were collected at E12.5 and E16.5 (Fig. 4A). The brain slice locations are illustrated in Fig. 4A2 and 4A5. HE staining revealed that the brain structure of KI/+ mice remained intact and comparable to +/+ littermates, while KI/KI mice displayed severe structural abnormalities (Fig. 4B). At E12.5, the lateral ganglionic eminence (LGE) in KI/KI mice exhibited poor development, hindering hippocampal formation (Fig. 4B1). At E16.5, the hippocampal region was completely destroyed, the midbrain development (previously forming hypothalamus) was significantly impaired, showing signs of incomplete neuronal cellular and structural development, and a reduced midbrain volume (Fig. 4B2). Additionally, severe cerebral hemorrhage was evident, starting in the lateral ventricle at E12.5 and extending to various brain regions by E16.5, including the lateral ventricles, midbrain, and the cerebral cortex (Fig. 4B).

Fig. 4

PS1 D385A KI/KI mice exhibit severe cerebral hemorrhages and decreased neuronal progenitor cell proliferation during embryonic development. A) Schematic representation of the embryo at E12.5 (A1), and at E16.5 (A4), the transverse brain section of the embryo (A2, A5) and the corresponding brain atlas (A3, A6) for reference, sourced from the PRENATAL MOUSE BRAIN ATLAS (Library of Congress Control Number: 2008924612). B) The laminar architecture of the developing brains in D385A KI/+ is identical to their wildtype littermate, while KI/KI mice display disorganization and lower neural progenitor cell density compared to wildtype mice. The LGE is less prominent in KI/KI mice (highlighted in box in B1). At E12.5, the hippocampal formation in KI/KI mice is disturbed, contrasting with KI/+ mice (B1, arrows); by E16.5, the hippocampus is completely disrupted in KI/KI mice (B2, arrows). The diencephalon (midbrain) exhibits evident disruption with reduced neural cell density and incomplete morphology compared to wildtype littermate, further shrinking by E16.5. An abnormally enlarged lateral ventricle is observed at E12.5, expanding by E16.5. Severe hemorrhage is visible in the lateral ventricle as early as E12.5, extending into the subcortical region of the temporal lobe and mesencephalon in KI/KI brain at E16.5 (B, arrowheads). C, D) Proliferating (S-phase) neural progenitor cells with BrdU at E12.5 (C). The quantity of proliferating neural progenitor cells in the lateral vertical zone (LV) and diencephalon (DC) area is significantly reduced in KI/KI brains compared to wildtype controls (D). Data are represented as mean±SEM. ^***p < 0.001.

HE staining analysis of D385A’s effect on brain structure in KI mice revealed a notable decrease in neuronal cells in the lateral ventricular and midbrain regions. To investigate whether the D385A mutation affected the proliferation of neuronal progenitor cells during embryonic development, BrdU was intraperitoneally 30 min before sampling. Analysis of neuronal progenitor cell proliferation via immunohistochemistry showed a significant reduction in BrdU-positive neuronal cells in the lateral ventricle and midbrain region (diencephalon) of KI/KI mice (Fig. 4C). In contrast, KI/+ and +/+ mice exhibited similar levels of proliferation without significant difference (Fig. 4D).

NP-2 enhances γ-secretase activity in PS1 D385A model mice

Analysis of PS1 FL, NTF, and CTF contents in the cortical and hippocampal brain regions revealed that NP-2 decreased PS1 FL levels in the cortex but did not show a similar trend in the hippocampus (Fig. 5AB, EF). There were no significant changes in the levels of PS1 NTF and CTF in the hippocampus and cortex under NP-2 administration compared to KI/+ mice (Fig. 5A, CD, E, GH).

Fig. 5

The combination of Rg1 and GP (NP-2) restores γ-secretase activity partially in the cortex of 10-month-old KI/+mice. A-D) NP-2 significantly decreased PS1 FL expression in the cortex of 10-month-old KI/+ mice. The representative western blotting images (A), the statistical analysis of PS1 FL expression (B), NTF expression (C), and CTF expression (D) are shown. E– H) NP-2 did not impact PS1 FL expression in the hippocampus of 10-month-old KI/+ mice. The representative western blotting images (E), the statistical analysis of PS1 FL expression (F), NTF expression (G), and CTF expression (H) are shown. I, J) ELISA measurements of Aβ₄₀ and Aβ₄₂ post in vitro γ-secretase assay with cortex tissues. After a 4-h reaction, there is a significant decrease in de novo generation of Aβ₄₀ in KI/+ mice compared to WT mice; with the administration of NP-2, de novo generation of Aβ₄₂ significantly increases compared to KI/+ mice. Data are represented as mean±SEM. ^***p < 0.001, ^*p < 0.05.

In PS1 D385A model mice, the impaired PS1 protein self-shearing function led to a disruption in γ-secretase activity (Fig. 5IJ). To investigate whether the combination of GP and Rg1 (NP-2) could enhance PS1 protein self-shearing function and potentially restore γ-secretase activity in the cortex, mice were administrated with NP-2 (KI/+NP-2). We followed the in vitro γ-secretase assay, and the data found that NP-2 significantly increased the production of Aβ₄₂ compared with KI/+ mice (Fig. 5IJ). These results indicated that NP-2 may modulate γ-secretase function.

DISCUSSION

γ-Secretase carries out a sequential cleavage of the substrate C99 to generate Aβ peptides [24]. Mutations in PS1 can lead to alterations in γ-secretase activity, which disrupts the metabolic processing of AβPP. This, in turn, selectively promotes the production and accumulation of insoluble Aβ peptides, a major component of amyloid plaques found in the brains of patients with AD [25].

Aspartate mutations in PS1 such as D385A and D257A, are particularly crucial for the self-shearing function of PS1 [5]. These specific aspartate residues play a vital role in maintaining the structural integrity and catalytic activity of the PS1 protein within the context of γ-secretase function [5, 6]. In this study, we generated a novel KI mouse model expression PSEN1 D385A. To confirm the successful generation of PS1 D385A KI mice, we employed the Southern method and gene sequencing to analyze genomic DNA extracted from embryonic stem cells and newborn mice. In embryonic stem cells, we detected an 8.6 kb KI genotype fragment using the BamHI/ApaI digestion system, which was attributed to the insertion of a screening fragment (neo, 1.7 kb). In the genomic DNA of newborn mice, the neo fragment was excised, revealing a KI genotype fragment of 6.9 kb. These findings confirmed the presence of the mutation site within the PSEN1 gene, as anticipated. Furthermore, gene sequencing confirmed the successful genetic modification in the mice.

PS1 D385A KI/KI mice exhibited postnatal lethality. The KI/KI mice displayed: (i) abnormal growth retardation and significantly smaller size compared to wild-type +/+ littermates; (ii) impaired skeletal development and abnormally short tails; and (iii) severe head hemorrhage. Developmental abnormalities were observed as early as E12.5, when the hippocampus initiates formation, and by day E16.5, the hippocampus was completely absent. Similarly, the midbrain displayed evident developmental abnormalities at day E12.5, and by day E16.5, the midbrain structure was severely compromised. These phenotypes seemed to be more severe damage compared to PS1– /– mice, PS1 L435F, and C410Y KI/KI mice [26, 27]. Notably, PS1 D385A KI/+ mice exhibited a phenotype similar to wild-type mice, with no observable abnormalities.

A previous study demonstrated that brain hemorrhage in PS1– /– mice was linked to diminished brain capillary branching, expanded vascular diameter, and abnormal proliferation and necrosis of vascular endothelial cells, implicating the PSEN1 gene in the regulation of vascular development and angiogenesis [28]. Likewise, PS1 D385A KI/KI mice exhibited severe cerebral hemorrhage, prompting an investigation into the impact of the D385A mutation on angiogenesis. The examination of CD31 expression, a marker for cerebral vascular endothelial cells at various embryonic stages, revealed a significant increase over time in KI/KI and KI/+ mice (Supplementary Figure 1) [7], underscoring the role of the D385A mutation in inducting abnormal angiogenesis. Moreover, studies have indicated that PS1 plays a crucial role in angiogenesis, with proteins involved in this process, such as vascular endothelial growth factor receptor 1 (VEGFR-1), Notch, ErbB-1, and IGFI-R, serving as substrates of γ-secretase [29]. Deactivation of VEGFR-1 phosphorylation is observed in PS1– /– cells, while overexpression of PS1 full-length protein activates VEGFR-1 dephosphorylation, hinting at the significance of PS1 full-length protein in regulating VEGFR-1 dephosphorylation and its involvement in vascular neogenesis [30]. Furthermore, we observed a significant incidence of tumors in the KI/+ model mice (Supplementary Figure 2A). The first tumors were detected in 4-month-old male mice, presenting with skin cancer and severe liver damage. As the mice aged, an increasing number developed skin cancer, and cases of lymphoma were also noted (Supplementary Figure 2B). Post-mortem examinations revealed abnormally enlarged spleens and liver damage in individual mice. In contrast, no tumors were observed in the +/+ littermates. These findings indicate a loss of PS1 protein self-excision and shearing function, disruption of γ-secretase activity, and loss of shearing substrate function. Notably, Notch protein, a substrate of γ-secretase associated with carcinogenesis, is primarily affected. These results provide insight into why γ-secretase inhibitors used in clinical settings may lead to serious side effects such as skin cancer.

Experimental evidence suggests that FAD-related PSEN1 gene mutations impair the proteolysis, leading to the accumulation of full-length PS1 protein and a decrease in cleavage fragments. In our study, we found that the D385A mutation completely abrogates the self-shearing function of PS1 protein (Fig. 3). Compared to PS1 L435F and PS1 C410Y KI/KI mice [27], PS1 D385A KI/KI mice brains showed no expression of PS1 CTF and NTF, similar to PS1– /– brain tissue samples. Unlike PS1– /– samples, however, PS1 full-length protein was significantly accumulated in the D385A KI mice. This difference suggests that the more severe phenotype observed in KI/KI mice may be influenced by several factors. Firstly, it is possible that the PSEN1 gene is regulated in a dominant inactivation. Secondly, the retained functionality of the full-length PS1 protein and its continuous accumulation could exacerbate toxic effects. Thirdly, we observed a marked increase in the expression of other γ-secretase components (e.g., Nicastrin, Pen-2) following the rise in full-length PS1 protein expression (unpublished data). This contrasts with the decrease in component expression seen in PS1– /– mice, suggesting that the accumulation of full-length PS1 protein may contribute to the more severe phenotype in KI/KI mice. Additionally, we observed a significant decrease in the proliferation of neural progenitor cells at day E12.5 due to the D385A mutation, potentially contributing to the extensive loss of neural cells and abnormal enlargement of brain ventricles. Furthermore, Notch protein, a crucial γ-secretase substrate involved in the development, may be affected by the loss of γ-secretase activity, potentially contributing to the developmental abnormalities observed in KI/KI mice [31].

Furthermore, does the D385A mutation further affect the γ-secretase activity, and to what extent? We investigated the γ-secretase activity using various methods. Firstly, we assessed the expression of γ-secretase substrates, such as AβPP and N-cadherin proteins, in the brain tissue of model mice through western blotting. Our findings revealed a significant increase in the expression of CTF fragments of AβPP and N-cadherin proteins (Fig. 3A, B), indicating a reduction or disruption in γ-secretase activity. Secondly, through an in vitro γ-secretase activity assay, we directly incubated γ-secretase extracted from brain tissue with substrates C99 and N102-FmH in vitro to detect the shearing products. We observed that the products Aβ₄₀ and Aβ₄₂ from C99 were undetectable in the KI/KI mice, and the levels of both products were notably reduced in KI/+ mice (Fig. 3C, D). Similarly, no NICD expression was detected in KI/KI mice in the reaction with N102-FmH, and NICD expression was significantly decreased in KI/+ mice (Fig. 3C, D). These results suggest that the PS1 D385A mutation led to a loss of γ-secretase activity, affecting the cleavage of γ-secretase substrates AβPP and Notch. This experimental evidence further explains the severe side effects often associated with γ-secretaseinhibitors [10].

In recent years, the development of small molecule inhibitors and modulators targeting γ-secretase has emerged as a potential therapeutic approach for AD, indicating the potential advantages of small molecules derived from Chinese medicine in this context. The herbal formula combines GP with PNS (NP), with GP acting as the principal component and PNS as the assistant one. GP, a natural product found in Gardenia jasminoides Ellis [32], has gained attention for its potential therapeutic effects in neurodegenerative diseases [33, 34]. PNS, a bioactive mixture derived from Panax notoginseng, has demonstrated neuroprotective properties in dementia [35], and Rg1 as a main bioactive compound of PNS has been extensively investigated in the remedy of brain disorders, especially dementia and depression [36], and it has the potentials to prevent AD by alleviating depression, obesity, diabetes, and hypertension [36 –38].

The combination of GP and PNS (NeuroProtect formula, NP) has shown efficacy in reducing amyloid plaque accumulation in the brains of APP23 [39], improving synaptic density in APP/PS1 mice [21], ameliorating memory deficits, and exerting neuroprotective effects [21]. PNS was also observed to protect primary rat hippocampal neurons and brain microvascular endothelial cells from cell death under normal conditions and during oxygen/glucose-deprivation [40]. Furthermore, in vitro experiments, we found the main component GP and Rg1 (NP-2) regulated AβPP processing by modulating γ-secretase [18]. In vivo experiments with PS1 D385A KI mice, NP-2 significantly reduced the Aβ₄₀ accumulation along with the cerebral blood vessels, up to 80% compared with littermate heterozygous mice (Supplementary Figure 4) [7], which demonstrated that the imbalance production of Aβ occurred in PS1 D385A KI mice, and NP-2 showed the ability to reduce Aβ deposition in the context of cerebral blood vessels.

Moreover, in this study, we investigated NP-2 in the modulation of γ-secretase function. We found that NP-2 decreased PS1 FL levels in the cortex, and restored its disrupted γ-secretase activity in PS1 D385A KI mice, which may be attributed to the loss of PS1 protein self-shearing function. These findings lay an experimental foundation for exploring pathways to restore γ-secretase function and offering a treatment strategy for AD.

Conclusion

We successfully generated PS1 D385A mutant KI mice using gene editing technology. Our experimental findings indicated that the D385A mutation did not impact the expression of PS1 mRNA; however, it disrupted the self-shearing function of the PS1 protein, leading to the loss of γ-secretase activity. The KI/KI mice exhibited a phenotype similar to PS1– /– mice, including postnatal lethality, developmental abnormalities, brain hemorrhage, decreased neural progenitor cell proliferation, and increased cerebral angiogenesis. Upon comparison with PS1– /– mice, the D385A KI/KI mice displayed a more severe phenotype, particularly in terms of heightened brain hemorrhage and more pronounced brain developmental disorders. This observation may be attributed to the overexpression of the full-length protein. Additionally, our study revealed that the combination of Geniposide and Ginsenoside Rg1 (NP-2) effectively restored γ-secretase activity, offering a potential avenue for exploring new drug targets and their mechanisms of action in AD.

AUTHOR CONTRIBUTIONS

Chengeng Deng (Data curation; Formal analysis; Software); Qingyuan Cai (Formal analysis; Methodology; Visualization); Jiani Zhang (Investigation); Kexin Chang (Data curation); Tiantian Peng (Visualization); Xiaoge Liu (Methodology); Feng Cao (Investigation); Xinyuan Yan (Software); Junshi Cheng (Visualization); Xu Wang (Investigation); Yan Tan (Conceptualization; Data curation; Funding acquisition; Project administration; Supervision; Validation; Writing – original draft; Writing – review & editing); Qian Hua (Funding acquisition; Project administration; Supervision; Writing – review & editing).

Footnotes

ACKNOWLEDGMENTS

We would like to express our gratitude to Professor Jie Shen and Principal Scientist Dan Xia from Harvard Medical School for their invaluable technical guidance during this study. Principal Scientist Dan Xia provided expert assistance with the experiments while Dr. Yan Tan was a Joint Cultivation Doctoral Program student at Harvard Medical School, supported by the China Scholarship Council (CSC).

FUNDING

This study was supported by grants from the National Natural Science Foundation of China (Grant No. U21A201401, 82374175).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The data supporting the findings of this study is available on request from the corresponding author. The data are not publicly available due to privacy.

The supplementary material is available in the electronic version of this article: .

References

Watanabe

, Shen

(2017) Dominant negative mechanism of Presenilin-1 mutations in FAD. Proc Natl Acad Sci U S A 114, 12635–12637.

Hardy

, Higgins

(1992) Alzheimer’s disease: The amyloid cascade hypothesis. Science 256, 184–185.

Berezovska

, Jack

, McLean

, Aster

, Hicks

, Xia

, Wolfe

, Kimberly

, Weinmaster

, Selkoe

, Hyman

(2000) Aspartate mutations in presenilin and gamma-secretase inhibitors both impair notch1 proteolysis and nuclear translocation with relative preservation of notch1 signaling. J Neurochem 75, 583–593.

De Strooper

, Annaert

, Cupers

, Saftig

, Craessaerts

, Mumm

, Schroeter

, Schrijvers

, Wolfe

, Ray

, Goate

, Kopan

(1999) A presenilin-1-dependent gamma-secretase-like protease mediates release of Notch intracellular domain. Nature 398, 518–522.

Wolfe

, Xia

, Ostaszewski

, Diehl

, Kimberly

, Selkoe

(1999) Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and gamma-secretase activity. Nature 398, 513–517.

Bagaria

, Bagyinszky

, An

SSA

(2022) Genetics, functions, and clinical impact of presenilin-1 (PSEN1) gene. Int J Mol Sci 23, 10970.

Chen

, Hua

, Zhang

, Peng

, Liu

, Deng

, Si

, Chang

, Wang

, Yan

(2023) Effect of ginsenoside Rg1 and gardenoside on cerebral vascular Aβ of PS1 D385A mice study on the impact of level. Prog Modern Biomed 23, 2001–2007.

Hur

(2022) γ-Secretase in Alzheimer’s disease. Exp Mol Med 54, 433–446.

Nie

, Vartak

, Li

(2020) γ-Secretase inhibitors and modulators: Mechanistic insights into the function and regulation of γ-Secretase. Semin Cell Dev Biol 105, 43–53.

10.

Tan

, Zhang

, Wong

, Hua

(2016) Anti-Alzheimer therapeutic drugs targeting γ-secretase. Curr Top Med Chem 16, 549–557.

11.

Wolfe

(2019) Structure and function of the γ-secretase complex. Biochemistry 58, 2953–2966.

12.

Hage

, Stanga

, Marinangeli

, Octave

, Dewachter

, Quetin-Leclercq

, Kienlen-Campard

(2015) Characterization of Pterocarpus erinaceus kino extract and its gamma-secretase inhibitory properties. J Ethnopharmacol 163, 192–202.

13.

Wuli

, Tsai

, Chiou

, Harn

(2020) Human-induced pluripotent stem cells and herbal small-molecule drugs for treatment of Alzheimer’s disease. Int J Mol Sci 21, 1327.

14.

Calderaro

, Patanè

, Tellone

, Barreca

, Ficarra

, Misiti

, Laganà

(2022) The neuroprotective potentiality of flavonoids on Alzheimer’s disease. Int J Mol Sci 23, 14835.

15.

Nakajima

, Ohizumi

(2019) Potential benefits of nobiletin, a citrus flavonoid, against Alzheimer’s disease and Parkinson’s disease. Int J Mol Sci 20, 3380.

16.

Liu

, Hua

, Lei

, Li

(2011) Effect of Tong Luo Jiu Nao on Aβ-degrading enzymes in AD rat brains. J Ethnopharmacol 137, 1035–1046.

17.

Yang

, Tan

, Wang

, Zhang

, Sun

, Zhang

, Yao

, Zhao

, Wang

, Fan

, Hua

(2014) The improvement of spatial memory deficits in APP/V717I transgenic mice by chronic anti-stroke herb treatment. Exp Biol Med (Maywood) 239, 1007–1017.

18.

Tan

, Zhang

, Yang

, Xu

, Zhang

, Chen

, Peng

, Wang

, Liu

, Wei

, Li

, Zhang

, Liu

, Hua

(2022) Anti-stroke Chinese herbal medicines inhibit abnormal amyloid-β protein precursor processing in Alzheimer’s disease. J Alzheimers Dis 85, 261–272.

19.

Takahashi

, Hayashi

, Tominari

, Rikimaru

, Morohashi

, Kan

, Natsugari

, Fukuyama

, Tomita

, Iwatsubo

(2003) Sulindac sulfide is a noncompetitive gamma-secretase inhibitor that preferentially reduces Abeta 42 generation. J Biol Chem 278, 18664–18670.

20.

Watanabe

, Xia

, Kanekiyo

, Kelleher

3rd , Shen

(2012) Familial frontotemporal dementia-associated presenilin-1 c.548G>T mutation causes decreased mRNA expression and reduced presenilin function in knock-in mice. J Neurosci 32, 5085–5096.

21.

Tan

, Wang

, Zhang

, Li

, Peng

, Chen

, Wei

, Liu

, He

, Li

, Ding

, Li

, Wang

, Zhang

, Hua

(2022) NeuroProtect, a candidate formula from traditional Chinese medicine, attenuates amyloid-β and restores synaptic structures in APP/PS1 transgenic mice. Front Pharmacol 13, 850175.

22.

Shen

, Bronson

, Chen

, Xia

, Selkoe

, Tonegawa

(1997) Skeletal and CNS defects in Presenilin-1-deficient mice. Cell 89, 629–639.

23.

Shoval

, Ludwig

, Kalcheim

(2007) Antagonistic roles of full-length N-cadherin and its soluble BMP cleavage product in neural crest delamination. Development 134, 491–501.

24.

Selkoe

(2001) Alzheimer’s disease: Genes, proteins, and therapy. Physiol Rev 81, 741–766.

25.

Xia

(2000) Role of presenilin in gamma-secretase cleavage of amyloid precursor protein. Exp Gerontol 35, 453–460.

26.

Karlawish

, Jack

Jr. , Rocca

, Snyder

, Carrillo

(2017) Alzheimer’s disease: The next frontier-Special Report 2017. Alzheimers Dement 13, 374–380.

27.

Xia

, Watanabe

, Wu

, Lee

, Li

, Tsvetkov

, Bolshakov

, Shen

, Kelleher

3rd (2015) Presenilin-1 knockin mice reveal loss-of-function mechanism for familial Alzheimer’s disease. Neuron 85, 967–981.

28.

Nakajima

, Yuasa

, Ueno

, Takakura

, Koseki

, Shirasawa

(2003) Abnormal blood vessel development in mice lacking presenilin-1. Mech Dev 120, 657–667.

29.

Boulton

, Cai

, Grant

(2008) gamma-Secretase: A multifaceted regulator of angiogenesis. J Cell Mol Med 12, 781–795.

30.

Cai

, Chen

, Ruan

, Han

, Liu

, Qi

, Boye

, Hauswirth

, Grant

, Boulton

(2011) γ-Secretase and presenilin mediate cleavage and phosphorylation of vascular endothelial growth factor receptor-1. J Biol Chem 286, 42514–42523.

31.

Handler

, Yang

, Shen

(2000) Presenilin-1 regulates neuronal differentiation during neurogenesis. Development 127, 2593–2606.

32.

Zhang

, Zhang

, Hu

, Xiao

, Ou

, Chen

, Luo

, Cheng

, Jiang

, Ma

, Zhao

(2021) The emerging possibility of the use of geniposide in the treatment of cerebral diseases: A review. Chin Med 16, 86.

33.

Zhang

, Liu

, Shi

, Xie

, Deng

, Chen

, Li

(2021) Therapeutic potential of catalpol and geniposide in Alzheimer’s and Parkinson’s diseases: A snapshot of their underlying mechanisms. Brain Res Bull 174, 281–295.

34.

Dinda

, Dinda

, Kulsi

, Chakraborty

, Dinda

(2019) Therapeutic potentials of plant iridoids in Alzheimer’s and Parkinson’s diseases: A review. Eur J Med Chem 169, 185–199.

35.

Zhou

, Chen

, Sapkota

, Long

, Liao

, Jiang

, Liang

, Wei

, Zhou

(2020) Pananx notoginseng saponins attenuate CCL2-induced cognitive deficits in rats via anti-inflammation and anti-apoptosis effects that involve suppressing over-activation of NMDA receptors. Biomed Pharmacother 127, 110139.

36.

Yang

, Wang

, Cheng

, Chen

, Hu

, Zhu

(2023) Ginsenoside Rg1 in neurological diseases: From bench to bedside. Acta Pharmacol Sin 44, 913–930.

37.

, Zhang

, Liu

, Xia

, Yang

, Tang

, Chen

, Ao

, Peng

(2024) Ginsenoside Rg1, lights up the way for the potential prevention of Alzheimer’s disease due to its therapeutic effects on the drug-controllable risk factors of Alzheimer’s disease. J Ethnopharmacol 318, 116955.

38.

, Yang

, Wan

, Xia

, Xu

, Zhang

, Liu

, Chen

, Tang

, Ao

, Peng

(2022) New insights into the role and mechanisms of ginsenoside Rg1 in the management of Alzheimer’s disease. Biomed Pharmacother 152, 113207.

39.

, Li

, Hua

, Liu

, Staufenbiel

, Li

, Shen

(2013) Chronic administration of anti-stroke herbal medicine TongLuoJiuNao reduces amyloidogenic processing of amyloid precursor protein in a mouse model of Alzheimer’s disease. PLoS One 8, e58181.

40.

, Hou

, Sun

, Li

, He

, Liu

, Zhao

, Hua

(2012) Neuroprotective effects of tongluojiunao in neurons exposed to oxygen and glucose deprivation. J Ethnopharmacol 141, 927–933.