Approaches to experimental modeling of neuroinflammation in neurodegenerative diseases: part 2 — genetic models

Introduction

The advantage of genetic models compared to classical animal models of diseases is their high specificity, allowing intervention in a single gene or biochemical pathway, which provides a promising platform for testing targeted biotherapeutics (anti-cytokine antibodies, small molecules). However, a number of limitations must be considered:

— Species specificity of the human and mouse immune systems: microglia differ significantly in gene expression profiles and responses to stimuli;
— Simplification of pathology: human neurodegenerative diseases are multifactorial, and a single mutation cannot reproduce the entire spectrum of the pathological condition;
— The problem of “superpathology”: excessive gene expression can lead to artifacts not relevant to humans;
— In many models, it is difficult to determine whether inflammation is a trigger or a consequence of neurodegeneration;
— Lack of aging as a key risk factor: most studies are conducted on young animals, whereas in humans neuroinflammation (NI) progresses with age.

In this article (part 2), we discuss the main genetic models of NI and describe their features.

Transgenic models reproducing key pathologies

Models targeting the expression of mutant forms of the amyloid precursor protein (APP) and presenilins (PS). These reproduce amyloidosis, microglial activation (disease-associated microglia — DAM phenotype), and astrogliosis.

PDAPP mice — one of the first transgenic models of Alzheimer’s disease (AD), characterized by overexpression of human amyloid precursor protein (APP) with the V717F mutation under the control of the PDGF-β promoter. This results in an 18-fold increase in APP gene RNA levels and a 10-fold increase in APP protein concentration, which in turn leads to Aβ accumulation. A distinctive feature of this model is a multiple increase in the concentration of Aβ42 — the form of β-amyloid most prone to aggregation in the cortex and hippocampus, followed by activation of astrocytes and microglia, which in turn induces NI and neurodegeneration [1].

APP/PS1 mice express a chimeric human/mouse APP with mutations found in familial AD. These mutations promote rapid and sustained Aβ accumulation, leading to a prolonged neuroinflammatory response and synaptic dysfunction in areas of senile plaque deposition and amyloid aggregates [2].

5xFAD — a line of transgenic mice generated by introducing two human genes with five mutations. These animals overexpress mutant human APP (A4) 695 with the Swedish (K670N, M671L), Florida (I716V), and London (V717I) familial AD mutations, as well as human PS1 containing two FAD mutations — M146L and L286V. Expression of both transgenes is regulated by neuron-specific elements of the mouse thy1 promoter [3]. 5xFAD mice exhibit increased APP expression, correlating with accelerated deposition of amyloid fragments Aβ1-40 and Aβ1-42 in the brain and cerebrospinal fluid, which rapidly increases with age. Histological analysis of the cerebral cortex and hippocampus revealed significant numbers of plaques and the formation of neurofibrillary tangles accompanied by strong NI. These pathological features also intensify significantly with age. Around the third month of life, astrogliosis and microgliosis begin to develop simultaneously with plaque formation. 5xFAD mice show reduced levels of synaptic marker protein, elevated p25 levels, and neurodegenerative and cognitive impairments [4].

APP23 mice express human APP751 with the Swedish double mutation (K670N/M671L) under the control of the neuron-specific mouse thy1 promoter, resulting in a seven-fold increase in human APP expression compared to endogenous APP levels in wild-type mice. APP23 mice are characterized by the formation of more stable amyloid plaques in the hippocampus and neocortex, the production of which increases exponentially with age. These deposits are accompanied by NI, synaptic dysfunction, neuronal loss, and tau hyperphosphorylation [5].

Tg2576 — another model overexpressing human APP with the Swedish double mutation, regulated by the hamster prion protein gene promoter. This mutation increases the production of both forms of β-amyloid — Aβ42 and Aβ40. Tg2576 mice exhibit a relatively slow rate of amyloid plaque formation, accompanied by a deficiency of antioxidants in the neocortex — glutathione peroxidase 1 (GPX1), superoxide dismutase-1 and -2 — as well as gliosis, astrocytosis, impaired glucose metabolism, and neurodegenerative changes [6].

3xTg-AD represents one of the most biologically significant animal models described to date, as it reproduces all the histopathological and behavioral features of AD. 3xTg-AD mice contain three genetic loci associated with AD: human PS1 M146V, human APP SWE K670N/M671L, and human tau P301L. These mice exhibit both plaque and tangle pathologies. Aβ deposition [7].

Models expressing mutant forms of tau protein (MAPT-transgenic mice) and the necroptosis protein MLKL (Mixed-lineage kinase domain-like)

P301S — a genetic mouse model with the P301S mutation in the tau gene (MAPT), expressing the 383 isoform of human tau protein under the control of the mouse thy1 promoter. It is characterized by accumulation of tau aggregates, neurofibrillary tangles, brain atrophy, cognitive impairment, and motor system dysfunction. As early as the third month of life, these mice show synaptic loss in the hippocampus due to pathological microglial activation and NI [8, 9].

rTg4510 mice express human tau protein containing the P301L mutation associated with frontotemporal dementia and mimic features of human tauopathy, including tau hyperphosphorylation, neuronal loss, and memory impairment. This model demonstrates sustained tau aggregation and neurodegeneration; however, several factors beyond hTau overexpression complicate data interpretation. Potential side effects resulting from transgene insertion into the Fgf14 locus should be considered. Disruption of this gene, critical for neuronal excitability, may independently contribute to some observed behavioral deficits, particularly those related to motor coordination and exploratory behavior [10].

PS19 transgenic mice overexpress the human tau isoform T34 and the 4 microtubule-binding domain (1N4R) tau protein with the P301S mutation under the regulatory control of the mouse prion promoter. PS19 mice are a popular model for studying tau pathology, tau aggregates, and other AD-related symptoms such as age-related cognitive impairment. Neurodegenerative disease in PS19 mice is driven by p16INK4a-expressing endothelial cells and microglia [11].

Tg-Mlkl−/− were generated by crossing MLKL knockout mice (MLKL — mixed-lineage kinase domain-like protein, involved in necroptosis) with transgenic mice carrying the SNCA A53T mutation. This novel model accurately mimics progressive features of Parkinson’s disease. In vitro experiments showed that MLKL inhibition reduces cell death induced by 6-hydroxydopamine and TNF-α or by toxic preformed α-Syn fibrils (PFF). Moreover, MLKL reduction led to improved motor symptoms, reduced NI, and decreased expression of phosphorylated α-Syn in the substantia nigra (SN), cerebral cortex, and striatum of A53T transgenic mice [12].

Models targeting CNS immune cells

CX3CR1-GFP/+ (and analogues) — a genetic model used to study the role of the chemokine receptor CX3CR1 in NI. The model serves as a tool for visualizing monocyte activity in the brain and illustrates the potential of the multifunctional fluorescence method based on Cx3cr1(gfp/+) for analyzing monocyte function in vivo [13].

hM3Dq/hM4Di (DREADD) in microglial cells allows chemogenetic activation or suppression of specific immune cell populations and the study of consequences for neurons and behavior. DREADD technology is based on the creation of molecules that activate muscarinic receptor cell signaling. A modified form of the human M3 muscarinic receptor — hM3Dq (Gq-coupled DREADD) — is used to enhance neuronal activity, while hM4Di (Gi/o-coupled DREADD) is used to suppress neuronal activity [14].

TREM2-ko — a genetic model of NI in which the TREM2 gene is knocked out. TREM2 is a transmembrane protein expressed exclusively in brain microglia. Rare variant mutations R47H, R62H, and H157Y in the TREM2 gene increase the risk of late-onset AD. The TREM2 R47H knockout mouse model is used to study demyelination. The TREM2 H157Y model is used to study the development of amyloidosis [15].

Models of systemic inflammation affecting the brain

Transgenic animals with a proinflammatory phenotype (IL-1β, p25, NGF)

IL-1β[XAT] — transgenic mice overexpressing human IL-1β. A characteristic feature of NI is increased IL-1β production, leading to the development of microgliosis and astrogliosis with chronically elevated levels of proinflammatory cytokines. A distinctive feature of this model is the absence of neurodegenerative changes in the brain; synthesis of the β-amyloid precursor remains unchanged despite the presence of cognitive impairment. This model can be used to reproduce NI without neurodegenerative changes [16].

Transgenic model with hyperproduction of the p25 subunit of cyclin-dependent kinase 5 (CDK5). Normally, brain cells express the p35 subunit, which, forming a complex with CDK5, participates in corticogenesis, regulation of synaptic vesicle metabolism, neurotransmitter release, and signal transduction. Under pathological conditions, cleavage of the p35 subunit to p25 is activated by calcium-dependent kinases, leading to CDK5 dysregulation and the development of neurotoxic effects. Mice with p25 hyperproduction develop NI, tau hyperphosphorylation, cognitive deficits, and amyloid accumulation. Studies using this model have shown that genetic knockdown of CDK5 can prevent the formation of insoluble tau in the hippocampus and the impairment of spatial memory in p25-overproducing mice [17].

Models with NGF deficiency (NGFR100W knockouts) — this model is based on the generation of transgenic animals carrying a mutation in the nerve growth factor (NGF) gene. Such animals develop neurodegenerative processes characterized by deficits in visual recognition and spatial memory, neuronal degeneration, cholinergic deficit, tau hyperphosphorylation, and the appearance of β-amyloid plaques. At the biochemical level, there is also expression of several proinflammatory cytokines, such as IL-1β, TNF-α, and IFN-γ-induced ATPase, due to the development of an autoimmune response [18].

Conclusion

Modern genetic models are a powerful but still imperfect tool for studying the neurobiology and pharmacological correction of NI. Expert analysis requires a clear understanding of which specific aspect of NI is being studied and at what stage, taking into account species limitations and model artifacts. Therefore, to reflect a more complete picture of the multifactorial process of neuroinflammation, a combination of genetic approaches with other methods is necessary. A promising direction in modeling NI is the combination of genetic and chemical triggers, for example, the induction of neuropathology by systemic toxin exposure in animals that have undergone genetic correction.