Alzheimer’s disease (AD) maintains the highest prevalence and morbidity among neurodegenerative diseases. Memory loss is the key symptom of AD, the aggravation of which will lead to other symptoms associated with the disease development, and eventually lead to death of the patients. AD brings suffering to the patients and their families - a heavy burden on society.
Figure 1: Common Symptoms of Alzheimer’s Disease (AD)
Apart from the combination drug targeting the cholinergic system and glutamate receptor that was approved by the FDA, in last 20 years, all the remaining clinical trials on AD therapeutics have ended in failure. Scientists and pharmaceutical companies still carry on with their development of AD drugs. As of 2019, there were 2,173 ongoing clinical trials on AD.
Although the clinical trials focusing on amyloid hypothesis account for the largest proportion (about 22.3%), currently many of the drugs targeting Aβ do not improve the cognitive function of patients after clearing Aβ; the clinical trials on neurotransmitter hypothesis occupy the second largest proportion (about 19.0%). The currently approved drugs are all related to this hypothesis, but their effect is barely satisfactory and will wear off soon; the tau hypothesis accounts for about 12.7%, but because of its clinical outcome analogous to that of the drugs targeting Aβ, it cannot improve cognition; about 17.0% of clinical trials focus on the mitochondrial cascade hypothesis and related hypotheses. This type of drug is not well-targeted, and there is no promising drug for AD at present. This fully shows that from a social perspective, the increase in the incidence of AD caused by aging has become an inevitable problem; from an economic point of view, all the major pharmaceutical research institutes recognize the enormous potential benefits of AD drug development, so AD drug discovery will continue despite the obstacles currently faced.
Figure 2: Global clinical trials on AD in 2019
At present, there are about 194 kinds of genetically modified Alzheimer disease (AD) mice available. If the unverified or non-transgenic models are excluded, there are at least 161 types of gene-edited mice supported by literature - most of these are single gene-edited, but the multiple gene-edited AD models are the most popular among other AD mouse models.
Figure 3. Types and number of gene-edited AD mouse models (data source: Alzforum). In the figure, pink indicates a single gene altered AD model; blue indicates a double gene-altered model; purple indicates a model with multiple genes altered; the numbers represent the number of model types.
In terms of model type, among the single gene edited mice, the largest number of models targeting APP, followed by the tau coding gene MAPT, while presenilin (PS) gene and β-cleavage enzyme BACE also occupy a tiny space. However, researchers are mostly focused on PS1 or BACE1 and there appears to be no interest in PS2 or BACE2 transgenic mice. It should be noted that both APP and PS1 are mutations found in familial AD patients, so properly speaking, only a model designed for these two genes is a standard AD model. However, since BACE1 does not mutate to induce AD, BACE1-related mice are not suitable AD models, but the study of their function still helps to better understand the pathogenesis of AD.
There are dozens of APP family mutations, but only four are commonly used (marked with red stars in Figure 4). A familial APP mutation is often named according to the location of the first family in which the mutation is discovered. For example, the SWE mutation was discovered in a family in Sweden. Moreover, APP mutated at two sites in this family, and the SWE mutation is one of the most commonly used in all the AD gene-edited animals. The gene that has the second largest number of mutations in familial AD is PS1, and five of its mutations are used frequently.
More than 100 kinds of transgenic AD models can be obtained by arranging, combining, and expressing these mutant genes. Some famous ones are the single-transgenic APPswe Tg2576, double-transgenic APP/PS1△E9 and 5×FAD, of which 5×FAD mice have the strongest pathological phenotype and early time of onset. This is because three APP mutations and two PS1 mutations are inserted into the mice, which are currently the AD mice carrying the largest number of inserted mutations. It should be noted that although 5×FAD mice have obvious Aβ pathological phenotype, there are no neurofibrillary tangles (NFT) formed around Aβ spliced protein or precursor protein APP in any of the single-transgenic mice. Therefore, the MAPT mutation in frontotemporal dementia (FTD) (the MAPT mutation enhances the phosphorylation of tau protein) is introduced, developing 3×TG mice based on the APPswe mutation and the PS1M146V mutation. The 3×TG mice have all the pathological phenotypes of AD.
Figure 4: Mutation and disease model mice
A literature review published in 2012 contains a summary of common transgenic AD mice. Figure 5 shows four kinds of AD mouse models separated by thick black lines, while the electrophysiological experiment is distinguished from the behavioral experiment by the background color. The electrophysiological experiment, which is in the upper portion of the figure, includes basal synaptic transmission (BST), long-term potentiation (LTP) and paired pulse facilitation (PPE); the behavioral experiment, which is in the lower portion of the figure, is divided into spatial memory (including Morris water maze, Barnes maze, etc.), working memory (T/Y maze, etc.) and fear memory (fear conditioning). In addition, owing to the impact of age on the course of AD, the age of the experimental animals is further differentiated in this paper, with the experimental animals divided into the young group (under 6 months of age), the middle-aged group (7-12 months old), and the elderly group (above 12 months of age). Every red and green box represents an experiment. There may be several experiments in the same paper. In fact, the six pieces of electrophysiological data of APP/PS1 come from the same paper, while electrophysiological data of 3×TG are also from the same paper. Red indicates that the function of the transgenic mice is impaired compared with the control, while green indicates that there is no significant difference from the wild type control.
Figure 5: Most common AD mice models
Source: Tara Spires-Jones and Shira Knafo, 2012
In general, there are relatively few electrophysiological experiments. The only information that can be obtained is that PPF is damaged in APP/PS1 while it is not altered at all in the rest three kinds of mice; on the other hand, both BST and LTP in 5×FAD are significantly affected with age. So for electrophysiological experiments, we may have to consider whether the synaptic function of various transgenic mice will change at different ages.
Compared with electrophysiology, there are much more data from behavioral experiments, such as the Morris water maze, to study spatial memory, and T-maze, to study working memory and fear conditioning. As can be seen from Figure 5, the spatial memory and working memory of Tg2576 are basically impaired to an equal extent at the age of 12 months. In addition, considering that the original table has been simplified, the fear memory is originally divided into hippocampus-dependent environmental fear memory and hippocampus-independent auditory fear memory. The two green blocks in the figure both represent hippocampus-independent experiments. If we look at the hippocampus-dependent experiments only, the four listed experiments of AD mice aged above 2 months all indicate memory impairment. For the behavioral experiment on the mice aged 7-12 months, the results of single-gene overexpressed Tg2576 are quite contradictory. In contrast, the overall memory of the remaining multi-transgenic mice is significantly impaired.
The data of mice aged 1-6 months are quite interesting. AD progresses fast in mice during this period of time, so it is not surprising that there are various contradictory results. In the future, the statistical time should be further subdivided to discover more relevant patterns.
Here we can see that if a fear memory experiment is conducted, Tg2576 will accumulate sufficient data, suggesting that its contextual fear conditioning (FC) at all ages is severely affected. This indicates that the hippocampus is seriously impaired, but on the other hand, the electrophysiological phenotype is not very obvious.
The incidence of AD is closely related not only to APP, PS1/2, MAPT, etc., but also to TREM2 in the immune signaling pathway and APOE in the cholesterol metabolism pathway, as is shown by many GWAS studies; also, the aggregation of Aβ or the phosphorylation of tau is affected by the alteration of many genes in pathways such as cytoskeleton positioning, energy metabolism and gene expression regulation. These signaling pathways, as well as the relationship between AD and the proteins in the signaling pathways, remain worth studying extensively. There are currently no successful clinical cases for the Aβ antibody drugs and tau dephosphorylation drugs of greatest concern, seeming to indicate that the key to treating AD may not lie in the direct pathological characteristics. Aβ and tau are probably not the “cause” of AD, but the “effect”. Therefore, the study of genes in non-traditional disease models may be the key to AD treatment.
Table 1: Pathways influencing AD and representative genes
The use of custom mouse models could provide much more clarity to these disease states by more accurately representing the human system(s) of significance for each study. This is especially critical for the broader range of invasive therapeutics; establishing an effective animal model for a particular study can provide significant savings (time & budget) over the long-term by quickly providing in vivo proof of concept & feasibility data.
Cyagen provides researchers from around the world with transgenic, knockout, knockin, conditional knockout models and also offers a comprehensive series of stem cell products for research use, including cell lines, media, and differentiation kits. Cyagen’s proprietary embryonic stem (ES) cell-based TurboKnockout® Gene Targeting is free of patent disputes, providing precise mouse models on a comparable timeline to CRISPR/Cas9-based techniques. TurboKnockout® can provide you with custom conditional knockout (cKO/floxed), conventional knockout (KO), ROSA26 knockin, point mutation knockin (KI), reporter knockin, and humanization mouse models in as fast as 6 months.
Feel free to contact us for more information about our services and how we may work with you to provide a free targeting strategy design for your next project.
In addition to gene editing models, AD animal models may be obtained in other ways, of which the aging model and chemical induction model are commonly used. Before the maturation of gene editing technology and methods, animal models were mostly generated through administration of drugs or viruses. The ways that substances enter the animal body include, but are not limited to, breathing, food and water intake, intragastric administration and injection. This results in mouse models which suffer from learning and memory disorders analogous to those caused by AD, without simulating the actual pathology of AD onset. The advantage of the above method is that it is easy to quickly build a model and only wild-type mice are needed.
Below, Table 2 shows several most common non-gene-edited AD models as well as their advantages and disadvantages. It is worth noting that these models do not contain senile plaques or NFT, the two most important pathological phenotypes of AD. So, many scholars argue that these models can only be used as memory impairment models and cannot truly be called AD models.
|Model Type||Model Name||Method||Advantages||Disadvantages|
|Aging Model||Natural Aging||Raise the mice till the age of 12-32 months, and conduct an experiment at time points during the period||Abiogenesis;
Taken for injury intervention
|No Aβ deposition; no NFT deposition; long period.|
|Aging Model||Quick Aging||SAMP mice||There are some AD-related inflammatory lesions;
There are learning and memory disorders.
|High cost; short experiment period.|
|Chemical Model||Aβ Injection||Inject Aβ into the hippocampus or ventricle||The pathological mechanism of Aβ can be fully studied.||No Aβ deposition; no NFT deposition; long period.|
|Chemical Model||Scopolamine Induction||Inject scopolamine into the abdominal cavity||Easy to inject medicine; the cholinergic hypothesis can be studied.||No Aβ deposition; no NFT deposition.|
|Chemical Model||Streptomycin Induction||Intracerebroventricular injection of streptomycin||Some characteristics of sporadic AD can be simulated.||No Aβ deposition; no NFT deposition; mice die easily.|
Table 2: AD chemical induction & viral vector-mediated models
In addition to drug selection, the way of administration is also an issue that needs considering. Although easy-to-operate, intravenous or intraperitoneal injection causes a great loss to drug efficacy, and it is uncertain whether the drug can penetrate the blood-brain barrier. Thus, the delivery of many drugs requires support from brain stereotaxic injection. With the help of this technology, we can directly inject the drug into the target brain region with precision. On the one hand, the drug is prevented from generating an extensive effect on the nervous system. On the other hand, the amount of usage is reduced, laying a foundation for conducting an elaborate experiment on the central nervous system. However, this injection method requires investment in fixed instruments and a cost of long-term learning.
Moreover, we can also inject a viral vector to directly edit the genes in the specific brain region of animals. This kind of gene editing applies to both wild-type animals or animals that have been gene-edited. For example, when using AAV for conditional expression, we can either inject a specific promoter + Cre gene into Loxp rats or inject genes into Cre animals. In general, the former is more common, and this is because there are usually more Loxp animals available than Cre animals. Additionally, owing to capacity issues, AAV is easier to operate in the first method. Since the viral vector is injected, we can accurately control gene expression or knockout through the injection site, Cre promoter, and AAV serotype - so as to improve the reliability of the experiment.
Figure 6: Surgery/drug-assisted modeling and brain stereotaxic injection
Animal Drug Evaluation & Disease Model Services
Cyagen can provide you with a variety of drug evaluation models along with phenotype analysis services – delivering reliable and expedient data reporting for your project.
There are many types of behavioral experiments used in AD research, which are roughly divided into four categories: learning and memory, motion balance, mental emotion, and social group. Although these four types of behavioral experiments are all used for AD phenotype detection and there are also these types of behavioral changes in AD patients, the changes and impairments in memory function are most obvious in human, primate and rodent models. Therefore, behavioral detection methods such as Morris water maze, novel object recognition, fear memory box, radial maze, etc. are the most important behavioral experiments for evaluating the AD modeling effect and drug efficacy. Table 3 summarizes the types of behavioral experiments related to AD and different memory indexes tested in various experiments. During the process of research, we often need to conduct more than one behavioral experiment to evaluate the memory function of animals.
Table 3: Overview of AD-related behavioral experiments (picture source, or presented in a table)
For many behavioral experiments, different periods are associated with different research focuses. In the early studies, various sophisticated mazes usually dominated neurobiological research. But over time, standardized experimental procedures gradually became dominant. Moreover, the T/Y maze which is simplified from the traditional mazes is gradually lost its share. The Morris water maze experiment and novel object recognition invented in the early 1980s accounted for half of the AD-related behavioral experiments.
Figure 7: Changes in the major behavioral tests related to AD
An animal behavior experiment needs to be designed systematically. For example, the Y-maze experiment used for memory testing that is slightly irritating animals and related to the early memory impairment induced by AD can be placed first in terms of time; novel object recognition, a type of cognitive ability testing, is also one of such experiments. The strongly irritating water maze and radial arm water maze experiments should be conducted at the intermediate stage. However, there should be a certain time interval between different types of experiments. The fear memory box not mentioned in the figure needs to be used carefully in accordance with the experimental purpose. It can be used for the final behavioral experiment or there should be a longer time interval between it and other experiments, or it can be used for another group of animals. When the animals reach old age, the slightly irritating active choice experiment or novel object recognition experiment can be conducted in the Y-maze.
Figure 8: AD behavior experiment process setting
There are different detection indexes for different behavioral experiments. The detection index for the Morris water maze is the time required by the animals to find the escape platform. The shorter the time, the better the memory of the mice; the Y-maze is used to count the ratio of the number of correct choices made by the mice to the total number of choices, as well as the accuracy of choices. The higher the accuracy, the better the memory of the mice; the planetary maze is similar to the Y-maze, but there are more countable data; in novel object recognition, the characteristic that mice are more interested in novel objects is used to test whether the mice can correctly distinguish between new and old objects. To be specific, if the mice focus on exploring a new object, it indicates that they have a stronger memory.
Figure 9: Detection indexes for AD behavior experiments
In addition to AD, vascular dementia, dementia with Lewy bodies (DLB), frontotemporal dementia (FTD), and even Huntington’s disease (HD) can all cause memory impairment. Therefore, an accurate AD model must have the unique pathological characteristics of AD — deposition of amyloid-β protein (i.e., senile plaques) — in addition to the behavioral features of the disease. Taking the APP single gene mutation Tg2576 mouse model as an example, the components that constitute senile plaques begin to increase at the age of 6 months, and senile plaques begin to appear at the age of 9 months; this marks successful model construction at the pathological level. On the other hand, because there is no MAPT gene alteration in Tg2576, another pathological marker of AD — NFT — will not appear, but there will be a certain degree of increase in tau phosphorylation.
Figure 10: Pathological progression of AD
On the upper left of Figure 13 is a brain tissue slice of APP/PS1 double transgenic mice at the age of 18 months. We can observe that the brain is filled with senile plaques, which induce inflammation and therefore cause the death of neurons nearby, exerting an evidently destructive effect on the function of the brain. However, there are no visible senile plaques in the brains of wild-type mice; On the right side of Figure 13 are NFTs appearing in the brains of 3×Tg mice.
Figure 11: Senile plaques and NFT in AD mice
Source: Wendy Liu, 2016
When APP is spliced into Aβ, there will be many intermediates generated, and there may even be a difference in the size of Aβ itself. Different sizes of Aβ also exert different degrees of pathological effects. For instance, the toxicity of Aβ42 is stronger than that of Aβ40. APP has two different destinies, one of which is non-amyloid precipitation splicing and the other is amyloid splicing, which may cause AD. The former type of APP is first spliced by ADAM, an α-secretase, which cuts off APP from the middle of the Aβ region, eradicating the possibility of Aβ generation from the very beginning. Special products, such as APPSα, α-CTF and P3, may be generated during this process. For the second splicing method, APPSβ, β-CTF and Aβ are generated; the increase of these products has a characteristic effect on the severity of AD. Besides, considering that different forms of Aβ affect the pathogenesis of AD, we can get a general understanding of the pathology of AD by detecting the relative proportions of these components.
Figure 12: Aβ splicing
(Cuadrado-Tejedor and Garcia-Osta, 2014)
Taking the study in Figure 12 as an example, the author tested the various APP splicing products mentioned above, discovering that the APP gene in gene-edited rats tended to be β-spliced, while a protective mutation occurred in PS1 or this trend was suppressed, thereby reducing the generation of β splicing products.
Figure 13: Detection of Aβ splicing intermediates
Source: Jill Calentline, 2020
Due to the small molecular weight of Aβ-related products, it is hard to achieve good results by using Western Blot (WB) for detection, so the ELISA kit is needed in many cases.
Behavior and pathology have verified the effectiveness of AD models at the macro and micro levels, so it is necessary to conduct research from the perspective of signaling pathways. The human body is a sophisticated “device”. Any change may affect the whole body. Thus, the information about the pathological process of AD must be studied to effectively develop medicines or identify the cause of any change. In other words, it is imperative to explore the signaling pathway. This part is not only most complex, but also most modular.
Figure 14: Signaling pathway is the focus of phenotype research
|Research Workflow||How Cyagen Support Your Research|
● GRK5 regulates GSK3 β, but it has not been found to be associated with AD before
● Based on APP / PS1, GRK5 was mutated to make it lose its function, and a new gene disease model was constructed.
● At the same time, sh-5ysy cell line was used for genetic modification to construct cell model
● Species: Mouse and Rat
● Method: KO/ cKO / KI / cKI / TG / Humanize
● Technology: TurboKnockout®, CRISPR/Cas9, Transgenic, BAC, Cre-loxP
● KO, KI, Overexpression, Point mutations
● Neural Stem Cells
● Protein inhibition was performed in APP / ps1-grk5ko model, and pathway changes were detected (intraperitoneal injection).
● Sh-5ysy-grk5ko was used for protein inhibition and RNA interference to detect the changes of pathway proteins.
● Virus Injection
● Drug Induction
● Water maze was used to detect the changes of memory function in 3-month-old, 9-month-old and 14 month old mice
● The detection index is the distance to the escape platform。
AD Pathological Detection
● The changes of pathological molecules (GSK3 β, pgsk3 β, tau, PtAu) under different knockout conditions were detected
● Including in vivo mouse detection and cell line detection.
● Metabolic Phenotyping
● Gene and Protein Expression
● Cell Function Testing
Signal Pathway Anlaysis
● The changes of GRK5 were detected by ELESA
● WB to detect of downstream GSK3 beta changes
● The cell localization of GRK5 was detected by fluorescence
● The distribution of GRK5 overexpression related proteins induced by virus was detected by fluorescence.
With more than 15 years’ experience in genetic engineering, Cyagen offers a complete range of services to support Alzheimer's disease research, including: custom genetically engineered models, ready-to-use catalog models, drug development models, colony management, and phenotype analysis - allowing researchers worldwide to create, manage, and monitor models with ease.
Custom Model Generation
● TurboKnockout® Gene Targeting Mice
● CRISPR-Pro Technology - Capabilities for Mouse & Ratl Models
● Transgenic Capabilities for Mouse & Rat Models
● Custom Rat Model Generation
Drug Development Mouse Models
● Metabolic Disease Models
● CDX/Syngeneic Models
● Immunodeficient Mice
In addition to animal model generation, Cyagen has established a range of supporting services, such as animal breeding, embryo/sperm cryopreservation, histology, and transcriptome profiling. Outsourcing these services to Cyagen is a more cost-effective way of conducting your research. All these services are performed by experienced specialists following standardized procedures, so we can guarantee delivery of high quality services and data acquisition with unbeatable price and turnaround.
Animal Model Supporting Services
- Breeding & Model Management
- Cryopreservation & Cryorecovery
- Phenotype Analysis
- Surgical Model Services
Contact us with your project details to request your complimentary model generation project consultation, strategy, and quote.