2024 Volume 72 Issue 7 Pages 618-629
Alzheimer’s disease (AD) is a common form of dementia. Although the causal mechanisms of AD are not fully understood, intracerebral accumulation of amyloid beta (Aβ) and tau aggregates seems to play an important role in disease development. Therefore, numerous experimental and clinical studies targeting the Aβ and tau proteins have been performed. However, these treatments have not achieved good clinical results. Additionally, recent findings have indicated that immune abnormalities contribute to the pathogenesis of AD. Several immune- and microglia-related genes have been identified as putative causative genes for the disease. Microglia, which are resident immune cells in the central nervous system (CNS), are key players that maintain brain homeostasis by communicating with other cells, such as astrocytes and immune cells, in or around the CNS. Furthermore, dysfunction of microglia and the immune system of the CNS could lead to chronic neuroinflammation and impairment of protective neuroimmune responses, which have been associated with the pathogenesis of AD and other forms of dementia. In this review, we assemble information regarding genetic evidence, imaging and biofluid biomarkers, and the pathophysiology of AD, especially highlighting bilateral (protective or detrimental) microglial functions, thus connecting neuroimmune dysfunction and AD. We also introduce candidate drugs to target neuroimmune dysfunction in AD. Finally, we discuss future therapeutic precision medicine approaches for AD, which could be achieved by identifying and targeting signals critical for AD pathogenesis through analyses of interactions between genetic risk factors, as well as identifying and modulating disease-relevant immune cell populations.
The number of people with dementia worldwide is predicted to increase by 3-fold by 2050.1,2) Forms of dementia include Alzheimer’s disease (AD), frontotemporal dementia, vascular dementia, and dementia with Lewy bodies. AD, which is a major type of dementia, is estimated to contribute to 60–70% of dementia cases.3) AD pathology is characterized by the accumulation of protein aggregates such as amyloid beta (Aβ) plaques and tau neurofibrillary tangles.4) Both genetic and neuropathological evidence strongly supports the amyloid cascade hypothesis for AD, which suggests that deposition of Aβ peptide initiates the disease.5) In addition, results from clinical and non-clinical studies have indicated interactions between the Aβ and tau proteins.4) Therefore, clinical trials targeting these proteins have been conducted (e.g., β-secretase inhibitors, anti-Aβ or tau antibodies). However, these trials showed limited efficacy or clinical benefit and produced safety concerns.6)
An increasing number of reports have highlighted the critical involvement of immune responses in AD, and immune-targeted interventions are considered alternative approaches for treatment. Inflammation has been observed in the brains of patients with AD, as indicated by the presence of inflammatory proteins and activated microglia, which are central nervous system (CNS)-resident immune cells.7) As an easy-to-understand example, genetic analysis identified that polymorphisms of inflammatory genes such as interleukin (IL)-6 were associated with an increased risk of developing AD.8) Genetic studies also identified several immune-related genes, including triggering receptor expressed on myeloid cells 2 (TREM2), as AD-associated genes, indicating that immune system dysfunction is involved in the clinical features of AD.9) Immune responses in the CNS are mainly regulated by CNS-resident microglia, astrocytes, and peripherally derived immune cells as well as immune cells that reside at the CNS borders. Neuroinflammation is considered to promote both the initiation and progression of AD.10,11) In addition, recent findings indicate the involvement of protective microglial activities mediated by TREM2 in AD.12) Identification of these new mechanistic factors that contribute to AD pathologies mediated by neuroimmune responses has offered the possibility of novel therapeutic approaches. Indeed, many drugs targeting neuroinflammation and protective microglial functions in AD have been tested in clinical trials.13) In addition to AD, recent reports have suggested the involvement of immune abnormalities in other forms of dementia including frontotemporal dementia,14) vascular dementia,15) and dementia with Lewy bodies.16)
In this review, we first provide genetic evidence linking immune dysfunction and AD. Next, we present the functions of major cell types that control immune reactions in the CNS and summarize recent updates of altered neuroimmune interactions involved in AD as well as other forms of dementia. We also provide information on the therapeutic pipeline for AD, which is believed to be deeply and directly related to immune function. Finally, we discuss perspectives of dementia treatment by highlighting recent studies with novel technologies including comprehensive genetic interaction analysis and single-cell genomics, which could lead to precision medicine for dementia by identifying and targeting disease- or symptom-relevant immune dysfunction.
No easy method exists to identify the causes of diseases that have multiple causative genes and are closely related to environmental factors, such as sporadic AD. Several solutions have been proposed to overcome this difficulty, and examples are given below. The aim of genome-wide association studies (GWASs), which differ from other specific candidate-driven studies, is to identify genome-wide localizations associated with specific phenotypes.17) Single nucleotide polymorphisms (SNPs) are the most widely used positional data, and information on copy number variants or sequence variations in the human genome is also available. This method can capture the responsible genes with high sensitivity by using many patient specimens.
AD is classified as either early-onset AD, with an age of onset ≤65 years, or late-onset AD (LOAD), with an age of onset >65 years.18) Genetic studies of early-onset AD and its familial characteristics have revealed three rare single-gene variants known to cause the disease: amyloid precursor protein (APP), presenilin 1 (PSEN1), and presenilin 2 (PSEN2).19) However, LOAD is more common and has a more complex genetic basis.19) Apolipoprotein E (APOE), which has three isoforms (APOE2, APOE3, and APOE4), is the most well-known gene associated with LOAD. The APOE4 allele increases, while the APOE2 allele decreases, the AD risk.20) Many genetic loci associated with AD risk were found over several decades, and these studies identified dozens of loci harboring genetic variants that influence inflammatory pathways.21,22) Expression analyses of AD-related genes in human and mouse brain cells revealed that many genes are preferentially expressed in microglia compared with other brain cell types.23,24) One of those genes is TREM2, which encodes a key protein that shifts the microglial state from homeostatic to disease-associated.12) The TREM2 R47H variant is a loss-of-function mutation that increases the odds ratio of AD by 2- to 4.5-fold.25,26) The contribution of immune responses to AD is further supported by the identification of an immune-mediated disease haplotype of the human leucocyte antigen (HLA)-DR15 as a risk factor for LOAD.24) In addition, quantitative trait locus analysis has been performed to identify DNA regions associated with specific phenotypes (histone acetylation, mRNA/protein expression, and mRNA splicing). Cell type-specific enhancer–promoter interactions were identified in the human cortex to analyze their association with GWAS-identified risk variants. These analyses showed that most AD risk variants are located within the microglia-specific enhancer; in other words, many variants related to AD impact gene expression only in microglia. One causal variant, the rs6733839 SNP, is located in the BIN1 locus on the microglia-specific enhancer. Deletion of the enhancer region abolished BIN1 expression in microglia but not in other cell types.27) AD risk variants have been shown to be predominantly located in active enhancers of myeloid (monocyte/macrophage/microglia) genes.28) Therefore, the Microglial Genomic Atlas was established to understand how genetic variation affects the microglial transcriptome.29) Genomic data were searched using gene expression data from primary microglia from 100 brain donors. The data showed that genetic variants were correlated with gene expression and splicing of thousands of genes, such as USP6NL and CD33, in microglia. CD33 expression was elevated in brains with AD,30) and the role of CD33 in microglial phagocytosis has been reported.31) These genetic studies indicate the importance of microglia-mediated immune responses in the development or progression of AD.
Immune responses are regulated by coordinated and sequential activation of the innate and adaptive immune systems to eliminate pathogens. Cells of the innate immune system, including microglia, monocytes/macrophages, and dendritic cells, act as the first line of defense and induce rapid and non-specific immune responses by sensing pathogen invasion and subsequent tissue damage. T and B cells, two major populations of the adaptive immune system, react in a slow but specific manner to mount antigen-specific immune responses.32) Homeostasis of the CNS environment is maintained by harmonious interactions among CNS-resident cells such as neurons, microglia, astrocytes, and oligodendrocytes. Peripheral immune cells and immune cells residing in the CNS borders participate in the maintenance of brain functions.33) Critical involvement of abnormal neuroimmune communications, including neuroinflammation, has been reported in several neurodegenerative and neurological diseases such as AD34) (Fig. 1), Parkinson’s disease,35) amyotrophic lateral sclerosis,36) and depression.37)
Immune responses in the central nervous system play both protective (left) and detrimental (right) roles in AD pathogenesis. Microglia exert protective functions such as phagocytosis and the clearance of amyloid beta (Aβ) protein. However, detrimental effects, including tau propagation, synaptic dysfunction, and neurotoxicity, can be induced by microglia-mediated neuroinflammation. Recent findings indicate pathological interactions among microglia, astrocytes, and T cells in AD. Created with Biorender.com.
Microglia, specialized macrophages that reside in the brain and spinal cord, are a type of glial cell that play critical roles in maintaining homeostasis in the CNS.38) Signaling through the colony stimulating factor (CSF)-1 receptor is critical for the development and maintenance of microglia.39) Microglia control synaptic plasticity and synaptogenesis, which are critical for brain development and function. Synaptic pruning, a complement-dependent process used to sculpt immature neuronal circuits by engulfing axons and dendritic spines, is an important function of microglia during development40,41) and in the postnatal phase.42,43) Apoptosis of CNS cells occurs during development, under both homeostatic and pathological conditions. Phagocytosis of apoptotic cells in the CNS is regulated by microglia via the TREM2-dependent pathway.44–46) Microglia, which are critical regulators of innate immune reactions in the CNS, play central roles in CNS immune responses, and their activation is critical for protecting the brain from pathogenic microorganisms. Infection and subsequent tissue damage are sensed by microglia through receptors that recognize pathogen-associated molecular patterns (i.e., viral RNA and DNA, bacterial proteins) and damage-associated molecular patterns (i.e., self-DNA, ATP). Inflammatory cytokines, such as IL-6, tumor necrosis factor (TNF)-α, and IL-1β, produced by microglia are critical mediators of protective antiviral immune responses.38,47–49) Activated microglia also serve as antigen-presenting cells and regulate antigen-specific adaptive immune responses against invading pathogens by activating T cells.50–52) Overall, microglia play critical roles in maintaining CNS homeostasis. Furthermore, under neurodegenerative conditions, microglia dynamically change their state from a homeostatic to a disease-associated phenotype, which is relevant to neurodegeneration through both TREM2-dependent and -independent pathways. This microglial phenotype is considered to protect the brain by removing amyloid plaques and damaged neurons.53,54) However, a similar microglial phenotype induced by TREM2 signaling may exacerbate disease pathologies under some circumstances.55,56) Additionally, other disease-associated microglial states have been identified that may contribute to disease pathogenesis and progression.57) Therefore, current therapeutic approaches targeting neuroimmune dysfunction mainly focus on microglia, considering their central roles in controlling immune responses in the CNS and the abundant information indicating their contributions to CNS diseases. We provide detailed information on the mechanistic links between microglial dysfunction and AD pathogenesis in the following sections.
3.2. AstrocytesAstrocytes are the most abundant type of glial cell and maintain brain homeostasis by controlling neurotransmitter recycling, ionic balance, and synaptic connectivity.58) Glial fibrillary acidic protein, expressed in astrocytes, has been associated with AD.59) The maintenance of blood–brain barrier (BBB) integrity is an important function of astrocytes. Astrocytes are a major producer of APOE proteins, which have been shown to impact BBB permeability. Importantly, APOE4, which is the protein that generates the greatest genetic risk for AD, is produced by astrocytes and induces impairment of BBB integrity via the induction of matrix metalloproteinase 9 in pericytes.60,61) In addition, reactive astrocytes, which are characterized by increased intermediate filament protein expression and cellular hypertrophy, have been implicated in AD progression.62,63) Furthermore, crosstalk between astrocytes and microglia can modulate AD-associated neuroinflammation in a complement-dependent manner.64–66) Collectively, an increasing number of studies have indicated that astrocyte dysfunction is involved in the pathogenesis of AD. Novel findings from recent studies, including single-cell or nuclear RNA sequence analyses,67–69) have revealed the diversity of astrocytes in disease conditions, which may offer novel astrocyte-targeted strategies for AD treatment.70)
3.3. Peripheral Immune Cells and Immune Cells at the CNS BordersSeveral reports have suggested the involvement of adaptive immune cells such as T cells and B cells, which control antigen-specific immune responses by activating cell-mediated and humoral responses, in AD progression.71–73) Importantly, recent findings highlight the critical roles of the CNS borders, including the meninges, cerebrospinal fluid, choroid plexus, and skull bone marrow, in neuroimmune interactions.33) The identification of the meningeal lymphatic vessels, which facilitate drainage of brain-derived antigens from the cerebrospinal fluid to cervical lymph nodes to prime adaptive immune responses,74) and their contribution to brain function strongly indicates that an interaction occurs between the CNS and the peripheral immune system to maintain brain homeostasis.75,76) Interestingly, recent reports have indicated the involvement of meningeal lymphatic functions,77) as well as T cells and their interactions with microglia,78–80) in the pathogenesis of AD. In addition, perivascular macrophages have been implicated in controlling cerebrospinal fluid flow81) and microglial phagocytic functions,82) which may be associated with AD. These new findings provide novel mechanistic insights into CNS diseases and could be linked to future therapeutic strategies for AD and other forms of dementia.
Post-mortem studies have shown that morphologically activated microglia exist in the brains of patients with AD,83,84) and microglial activation-associated markers such as CD68 and HLA-DR are also increased compared with those in healthy controls.85) These studies indicated that microglia participate in the development of AD. Longitudinal neuroimaging and cerebrospinal fluid biomarker studies in living patients have been conducted to better understand the contributions of microglia to AD pathologies and symptoms during disease progression. Mitochondrial translocator protein (TSPO), a protein upregulated in activated glial cells,86) was used as a target of a positron emission tomography tracer to identify activated glial cells including microglia. These studies indicated bilateral aspects of microglia in AD pathogenesis; in other words, dual peaks of microglial activation were observed during the mild cognitive impairment (MCI) and AD stages.87,88) Greater microglial activation was found in patients with MCI and AD than in controls. During the follow-up period, activated microglia were decreased in patients with MCI, whereas they were increased in patients with AD. In patients with MCI, the TSPO signal was positively associated with cognitive function and brain volume, suggesting that early microglial activation exerts protective effects against cognitive decline.89,90) However, a high TSPO signal in the AD stage was negatively correlated with cognitive scores and brain volume, indicating that microglia play a detrimental role in AD pathophysiology after disease onset.91,92) Cerebrospinal fluid levels of the soluble forms of receptors such as TREM2, AXL, and Tyro3, which contribute to the phagocytotic function of microglia,93,94) are related to the preservation of cognitive performance in patients with MCI and AD.95–97) Cerebrospinal fluid biomarker studies have indicated that microglial activation in the early stage of AD exerts protective effects against cognitive decline via phagocytosis of pathological aggregates, in contrast to the disease-promoting effects of their activation during the later phases of AD.
Many studies have indicated that increased reactive astrocyte levels near amyloid plaques represent a response to cellular stress in patients with AD. These cells produce large amounts of inflammatory cytokines and are involved in the pathophysiology of AD and other neurodegenerative diseases.98) Reactive astrocytes overexpress monoamine oxidase B, which can be used to distinguish this type of cell as a potential surrogate marker.99) Reactive astrocytes could be detected at Aβ accumulation sites prior to AD onset using this biomarker, which was highly correlated with the Aβ burden. Furthermore, monoamine oxidase B has recently been suggested as a potential therapeutic target in AD because of its association with aberrant γ-aminobutyric acid production in reactive astrocytes.100)
Recently, adaptive immune cells were also suggested to play an important role in AD pathogenesis. Epidemiological research showed that viral and bacterial infections increase the risk of AD and other CNS disorders. Notably, an association was identified between viral encephalitis and AD,101) indicating the possible involvement of both innate and adaptive immunity in AD. Indeed, post-mortem studies indicated that T cells infiltrate the brain parenchyma and are increased in patients with AD.102,103) Furthermore, CD8+ T effector memory CD45RA+ (TEMRA) cells, a subset of T cells with potent effector functions including cytotoxicity and proinflammatory cytokine production, were increased in cerebrospinal fluid from patients with AD, and the percentage of TEMRA cells in peripheral blood mononuclear cells was negatively correlated with cognitive function.78,104) Interestingly, clonally expanded CD8+ T cells from patients with AD showed specificity against Epstein–Barr viral antigens, suggesting a link between Epstein–Barr viral infection and AD pathogenesis.78,104)
The interaction of activated microglia with amyloid plaques was observed in post-mortem brains of patients with AD.105) Non-clinical studies demonstrated that microglia play an important role in clearing Aβ protein by phagocytosis.93,106,107) Microglial Aβ engulfment was also observed after Aβ immunotherapy with aducanumab, indicating the importance of the phagocytic function of microglia in Aβ pathology and immunotherapy—including lecanemab, the novel AD treatment.108–110) In addition, risk genes for sporadic AD, such as TREM2 and CD33, were shown to regulate microglial phagocytic activity. TREM2-deficient mouse microglia and human induced pluripotent stem cell-derived microglia showed impaired chemotaxis and phagocytic activity, resulting in decreased microglial accumulation around amyloid plaques and abnormal amyloid deposition.93,111) In contrast, TREM2 overexpression or activation induced the accumulation of microglia around plaques and reduced abnormal amyloid deposits in AD mouse models.112) A deficiency of CD33, which acts upstream of TREM2, also increased microglial activity and ameliorated amyloid pathology in Aβ transgenic mice.106,113) CD33-expressing microglia were found around amyloid plaques in post-mortem brains of patients with AD, and the number of microglia was associated with the amyloid burden.106) On the basis of findings indicating the critical role of TREM2 signaling-induced microglial phagocytic activity in AD, the TREM2 agonistic antibody (AL002), for which the mouse variant (AL002c) antibody was shown to activate microglia in AD mouse models expressing human TREM2, was tested in a phase 2 clinical trial114) (Table 1).
Stage | Drug name | Mechanism of action | AD model** | Results*** |
---|---|---|---|---|
Phase 3 | Masitinib | Tyk inhibitor | APPPS1 | A synaptoprotective effect associated with mast cell inhibition was observed.207) |
NE3107 | ERK/NF-κB inhibitor | |||
Semaglutide | GLP-1R agonist | |||
Phase 2 | AL002 | Anti-TREM2 receptor mAb | 5xFAD | A significant and beneficial impact on risk taking and fear conditioning via modulation of microglia was observed.114) |
BCG vaccine | Immunomodulator | APPPS1 | A beneficial effect on cognitive decline via modulation of inflammation-resolving monocytes and Tregs was observed.181) | |
APPPS1 | 4Aβ1-15/BCG combination treatment resulted in more effective synaptic preservation and improved learning efficiency through upregulation of IL-10 and recruitment of macrophages.182) | |||
Baricitinib | JAKs inhibitor | — | — | |
Canakinumab | Anti-IL1-β mAb | — | — | |
Daratumumab | Anti-CD38 mAb | — | — | |
Dasatinib + quercetin | Tyk inhibitor (dasatinib) and flavonoid (quercetin) | APPPS1 | Ameliorated cognitive deficits were observed via selective removal of senescent cells from the plaque environment.208) | |
L-Serin | Dietary amino acid; inhibits toxic misfolding | — | — | |
Lenalidomide | Reduces inflammatory cytokines | — | — | |
Montelukast | CysLT-1 receptor antagonist | Aβ injection | Aβ1–42-induced memory impairment was reduced via inhibition of neuroinflammation and apoptosis.209) | |
5xFAD | Cognitive decline was attenuated via modulation of microglia and CD8+ T-cells.210) | |||
Pegipanermin | TNF inhibitor | 5xFAD | Impaired LTP was rescued via modulation of MHCII+ microglia/macrophages and CD4+ T cells in the brain.143) | |
TgCRND8 | Synaptic deficits were prevented before the onset of amyloid plaque formation.211) | |||
Pepinemab (VX15) | Anti-sema4D mAb | — | — | |
Proleukin | IL-2 immunomodulator | |||
Sargramostim | GM-CSF | APPPS1 | Reduced brain amyloidosis, increased plasma Aβ, and rescue of cognitive impairment via increased hippocampal expression of calbindin and synaptophysin and increased doublecortin-positive cell levels in the dentate gyrus were observed.212) | |
Senicapoc | KCa3.1 blocker | 5xFAD | Although no improvement in cognitive function was indicated, reduced neuroinflammation, decreased cerebral amyloid load, and enhanced hippocampal neuronal plasticity via modulation of microglia were observed.213) | |
TB006 | Anti-galectin 3 mAb | APPSwe, 5xFAD | This efficacious therapeutic entity acted via degradation of toxic oligomers, and blocking or even reversal of AD progression was observed.214) | |
Tdap vaccine | Immune reaction to diphtheria, pertussis, tetanus vaccine | — | — | |
Valacyclovir | Antiviral against HSV-1 and -2 infection | |||
Phase 1 | CpG1018 | TLR9 agonist | ||
Emtricitabine | NRTI | — | — | |
IBC-Ab002 | Anti-PD-L1 mAb | 5xFAD, DM-HTau | A single antibody dose improved cognitive performance after 1 month in a mouse model.215) | |
Salsalate | NSAID | PS19 | Memory deficits and hippocampal atrophy were rescued by inhibiting tau acetylation.216) | |
Rapamycin | mTOR inhibitor | 3xTg-AD, PS19, J20, Tg2576, APPPS1, etc. | Improvements in pathology were observed in animal models when dosing was initiated before or after disease symptoms were present.217) | |
VT301 | Tregs |
* The drug candidates described here were extracted from reference 13 based on their immunomodulatory roles. ** APPSwe: a transgenic mouse that expresses the 695-amino acid isoform of human Alzheimer’s amyloid beta precursor protein derived from a large Swedish family with early-onset Alzheimer’s disease; APPPS1 a double transgenic mouse that expresses a chimeric mouse/human amyloid precursor protein (Mo/HuAPP695swe) and a mutant human presenilin 1 (PS1-dE9); Aβ injection: intrahippocampal injection of amyloid beta peptide (Aβ (1–42)) was used to construct the model; 5xFAD: a transgenic mouse with overexpression of APP and PSEN1 that contains five mutations; DM-HTau: a transgenic mouse that expresses the human tau gene with two mutations (K257T/P301S); J20: a transgenic mouse that expresses human amyloid protein precursor with both the Swedish (K670N/M671L) and the Indiana (V717F) mutations; P301S: a transgenic mouse that expresses human tauP301S; PS19: a transgenic mouse that expresses the T34 isoform of microtubule-associated protein tau with one N-terminal insert and four microtubule-binding repeats encoding the human P301S mutation; Tg2576: a transgenic mouse that expresses human amyloid precursor protein (APP695) with the Swedish mutation (KM670/671NL)’ 3xTg-AD: a transgenic mouse that expresses human presenilin 1 (PSEN1-M146V), human amyloid precursor protein (Swedish mutation), and the P301L mutation of human tau; TgCRND8: a transgenic mouse that expresses amyloid precursor protein (APP695) with two mutations. *** Although the results of clinical experiments using the drug and preclinical (in vivo) experiments using model animals are shown, for example, when the modality is an antibody, the results of the formulation when considering species differences may also be shown. Abbreviations: Aβ: Amyloid beta, AD: Alzheimer’s disease, APP: Amyloid precursor protein, BCG: Bacillus Calmette–Guérin, CD: Cluster of differentiation, CSF: Colony stimulating factor-1, CysLT-1: Cysteinyl leukotriene type 1, ERK: Extracellular signal-regulated kinase, GLP-1R: Glucagon-like peptide-1 receptor, GM-CSF: Granulocyte macrophage colony stimulating factor, HSV: Herpes simplex virus, IL: Interleukin, JAKs: Janus kinase, KCa3.1: Calcium-activated potassium channel 3.1, LTP: Long-term potentiation, mAb: Monoclonal antibody, MHCII: Major histocompatibility complex class II molecules, mTOR: Mechanistic target of rapamycin, NF-κB: Nuclear factor kappa-light-chain-enhancer of activated B cells, NRTI: Nucleoside reverse transcriptase inhibitor, NSAID: Non-steroidal anti-inflammatory drug, PD-L1: Programmed cell death ligand 1, PS1: Presenilin 1, Sema4D: Semaphorin 4D, TLR: Toll-like receptor, TNF: Tumor necrosis factor, Tregs: Regulatory T cells, TREM2: Triggering receptor expressed on myeloid cells 2, Tyk: Tyrosine kinase.
Granulocyte-macrophage (GM)-CSF is a cytokine that modulates microglial functions, including phagocytosis. An ex vivo co-culture model of amyloid plaque phagocytosis showed that GM-CSF induced microglial proliferation and reduced plaque size.115) In addition, intrahippocampal injection of GM-CSF into AD model mice reduced amyloid accumulation and improved cognitive function, indicating that GM-CSF stimulation of microglia leads to treatment effects via the promotion of Aβ clearance.116) A synthetic form of GM-CSF, known as sargramostim, has therefore been developed and is currently undergoing phase 2 clinical trials117) (Table 1). Overall, the enhancement of Aβ clearance by promoting microglial phagocytic activity is considered to exert therapeutic effects.
5.2. Mechanisms of Disease-Promoting Immune Functions5.2.1. Tau PropagationMisfolded and abnormally shaped tau proteins are found in the brains of patients with AD or tauopathy. Tau proteins can spread among cells and neuroanatomically connected regions,118) resulting in the accumulation of neurofibrillary tangles in AD and related disorders.119) Microglia can both protect against and contribute to the spread of pathological tau proteins. Clinical and non-clinical studies have indicated that microglia can internalize and degrade tau seeds, mitigating their spread.120–125) However, when microglia fail to degrade these tau seeds, they secrete tau or small extracellular vesicles (EVs), called exosomes, containing tau that can spread to neurons. Microglia derived from tau transgenic mice and patients with AD have been shown to release seed-competent tau proteins.125) A study analyzing a mouse model exhibiting rapid tau propagation indicated that microglia spread tau via exosome secretion, and inhibiting exosome synthesis significantly reduced tau propagation in vitro and in vivo.126) In addition, plaque-associated microglia expressed high levels of EV markers and hyper-secreted EVs containing the phosphorylated form of tau.127) Furthermore, clinical research showed that spatial propagation of microglial activation and tau accumulation were colocalized in AD patients.128) Collectively, microglia may contribute to AD pathology by spreading tau, suggesting the inhibition of microglia-dependent tau propagation as a possible therapeutic strategy.
5.2.2. Synaptic DysfunctionNeurons communicate with each other via microstructures called synapses, which can change their strength in response to neuronal activity. This synaptic plasticity is reflected by long-term potentiation (LTP), a process that involves persistent strengthening of synapses and is an important neurochemical foundation of learning and memory.129,130) The number of synapses is decreased in brains from patients with AD, and synapse loss is highly correlated with impaired cognitive performance.131–133) Many studies have shown that Aβ and tau induce synaptic toxicity, and glial cells/chronic inflammation cause synapse dysfunction.134) Pro-inflammatory cytokines, such as TNF-α and IL-1β, have been shown to disrupt synaptic plasticity in rodent hippocampus slices and AD model mice. Interestingly, LTP dysfunction was observed after the application of oligomeric Aβ and in hippocampal slices from Aβ transgenic mice, and this impairment was recovered by treatment with a TNF-α antagonist, infliximab.135) IL-1β application to hippocampal slices also induced LTP deficits.136) Synaptic plasticity requires the synthesis of proteins such as c-Fos,130) brain-derived neurotrophic factor, and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)137,138) after neuronal activation. IL-1β inhibited activity-dependent protein synthesis in neurons via p38 mitogen-activated protein kinase activation and the reduction of protein kinase B phosphorylation, resulting in decreased AMPA expression.139) This is possibly the mechanism of IL-1β-induced LTP impairment. Additionally, an IL-1 receptor antagonist was reported to ameliorate LTP deficits in senescence-accelerated mice.140) These results indicate that TNF-α or IL-1β inhibition may improve synaptic function in AD. In a 6-month-long exploratory investigation, supraspinal TNF-α antagonist administration improved cognitive scores in 15 patients with AD.141) An anti-IL-1β agent, canakinumab, and an anti-TNF-α agent, Xpro1595, were developed in each clinical trial (Table 1). Xpro1595 was tested on Aβ transgenic mice and rescued LTP deficits.143) These results suggest that anti-cytokine therapy may improve cognitive function in AD patients by recovering synaptic plasticity.
In addition to indirect mechanisms via cytokines, microglia and astrocytes directly eliminate synapses. During CNS development, glial cells engulf synapses and regulate the synapse number. During this process, synapses are opsonized by complement factors such as C1q and C3 and recognized by glial cells via complement receptors. Although glial synaptic pruning is important during development and homeostatic states, excessive pruning is considered to induce synapse loss and cognitive dysfunction in AD. C1q and C3 upregulation was observed in brain lysate, synaptosomes, and cerebrospinal fluid from patients with AD and in non-clinical AD models.144–146) C1q accumulated at synapses in soluble Aβ oligomer-injected mice, and microglial synapse engulfment increased. C1q antibody treatment or C3 receptor knockout ameliorated this excessive microglial synapse engulfment and aberrant synaptic function.147) Recently, astrocytes were shown to contribute to both excitatory and inhibitory synapse loss in a C1q-dependent manner.146,148) These studies demonstrated that complement-dependent pruning of excessive synapses by glial cells plays an important role in synaptic dysfunction in AD. In clinical settings, microglial activation has been reported to reduce brain functional connectivity independent of Aβ and tau accumulation, which could be an effect of cytokine-induced synaptic dysfunction and synaptic phagocytosis, as described above.149)
5.2.3. Neuronal DeathMicroglia induce neuronal death through various mechanisms including reactive oxygen species (ROS)-mediated cell death.150,151) Microglia generate ROS in response to Aβ and tau, which are the pathological hallmarks of AD.142) In turn, ROS generated by microglia damage cellular components, including DNA, proteins, and lipids. ROS can also induce oxidative stress, which is a hallmark of neurodegeneration, including AD.152) Other microglia-derived factors, such as proinflammatory cytokines, matrix metalloproteinases, and neurotrophic factors, also modulate neuronal cell death.153,154) In particular, the proinflammatory cytokine TNF-α binds to TNF receptor 1 expressed on neurons and leads to neuronal cell death by inducing nuclear factor-kappaB (NF-κB) signaling-mediated apoptosis or necroptosis.155,156)
In addition to microglia, other inflammatory cells such as astrocytes and T cells can contribute to neuronal death in neurodegenerative diseases. Neurotoxic reactive astrocytes are induced by microglia-derived IL-1α, TNF-α, and C1q in neuroinflammatory contexts.157) These reactive astrocytes were found in post-mortem brains of patients with AD and other neurodegenerative diseases.157,158)
Recent studies suggest the involvement of CD8+ T cells in neurodegeneration in AD and tauopathies.80) Cytotoxic CD8+ T cells infiltrated into the brain parenchyma in APOE4/tau transgenic mice, especially in areas with tau pathology. Surprisingly, T cell depletion in model mice ameliorated neurodegeneration, suggesting that T cells also contribute to neuronal death in AD and tauopathy. Although the exact mechanism of CD8+ T cell-mediated modulation of the neuronal landscape is not fully understood, T cell depletion inhibited microgliosis and tau accumulation in APOE4/tau transgenic mice. Therefore, T cells may exert cytotoxic effects against neurons via microglia- and tau-dependent mechanisms. Additionally, microglial depletion inhibited both T cell infiltration into the brain and neurodegeneration, suggesting a pathological interaction between microglia and T cells in AD and tauopathies.
5.2.4. Drug Candidates for Treatment of NeuroinflammationTable 1 shows 24 drug candidates in clinical development that have mechanisms to regulate immune responses.13) The mechanisms of action are indicated for each drug candidate, but interesting facts are noticed when considering microglia, which play a central role in neuroinflammation, as a starting point. For example, some immune responses that relate neuroinflammation to microglia are regulated via certain types of kinases, such as Bruton’s tyrosine kinase, extracellular signal-regulated kinase 1/2, and Janus kinase.159–166) Next, we focused on inflammatory humoral factors. Inflammatory cytokines such as IL-1, IL-6, and TNF-α, which modulate microglial responses to damage-associated molecular patterns,167) can be toxic to neurons and other glial cells.168) Furthermore, CD38-, cysteinyl leukotriene type 1-, calcium-activated potassium channel (KCa3.1)-, galectin 3-, CSF1 receptor-, semaphorin 4D receptor-, and PD-1 receptor-bearing microglia seem to participate in neuroinflammation,70,169–180) and it is reasonable for these cells to be considered as therapeutic targets for dementia. Interestingly, recent studies have investigated the potential of the Bacillus Calmette–Guérin vaccine to modify microglial activation phenotypes and reduce neuroinflammation.181,182) Finally, salsalate, which is a non-steroidal anti-inflammatory drug, has been found to have anti-inflammatory effects on microglia in vitro and in vivo.183,184)
In addition to AD, immune dysfunction has been associated with other forms of dementia such as frontotemporal dementia, vascular dementia, and dementia with Lewy bodies. Frontotemporal dementia is a group of neurodegenerative disorders that induce the loss of nerve cells in the frontal and temporal lobes of the brain and are characterized by alterations in cognition and behaviors. Genetic studies identified the involvement of immune-related genes including C9orf72 as well as the HLA locus in patients with frontotemporal dementia.185,186) In addition, TREM2 mutations have been linked to frontotemporal dementia as well as AD.187) Autoimmunity188,189) and microglial activation also contribute to frontotemporal dementia.190) Vascular dementia, which is often associated with stroke, is a type of dementia that affects cognition because of cerebrovascular pathologies.191,192) Although robust genetic risk factors have not been identified for vascular dementia,193–195) clinical and non-clinical studies have suggested that age-related immune abnormalities and cerebrovascular inflammation are associated with the development of age-related diseases, including stroke and vascular dementia, indicating the possible involvement of neuroinflammation in vascular dementia.196) Dementia with Lewy bodies is a type of progressive dementia characterized by abnormal α-synuclein deposits in the brain, which are known as Lewy bodies. Genetic studies have identified risk factors, including APOE, which has also been implicated in neuroinflammation in AD, in patients with dementia with Lewy bodies.197,198) Imaging and peripheral biomarker studies found the activation of glial cells and increased peripheral blood inflammatory cytokine levels in the mild or prodromal stages of dementia with Lewy bodies.199,200) Additionally, recent reports have identified the contributions of T cells, potentially by recognizing α-synuclein, to neurodegeneration in patients with dementia with Lewy bodies and Lewy body dementia, which consists of dementia with Lewy bodies and Parkinson’s disease dementia.201,202)
Collectively, neuroimmune dysfunction has also been implicated in frontotemporal dementia, vascular dementia, and dementia with Lewy bodies, indicating the potential of therapeutic strategies targeting neuroinflammation. Furthermore, common characteristics among these diseases, such as disease-linked TREM2 mutations in AD and frontotemporal dementia as well as the identification of APOE as a common risk factor in AD and dementia with Lewy bodies, indicate possible shared mechanisms between AD and other forms of dementia. Further studies will be required to uncover the similarities and differences of the contributions of immune abnormalities to various forms of dementia.
Genetic and positron emission tomography imaging analyses have suggested a link between changes in microglial functions and AD. In particular, proinflammatory cytokine release is considered an important function of microglia, which play central roles in immune responses in the CNS, and their effects on disease pathogenesis are being elucidated by numerous clinical and non-clinical studies. TNF-α is known to cause neuronal dysfunction, such as abnormal synaptic plasticity and TNF receptor-mediated neuronal death, and its suppression may not only improve short-term cognitive dysfunction, but also inhibit the long-term progression of dementia. Therefore, Xpro1595, a TNF-α inhibitor, and senicapoc and NE3107, which suppress TNF-α production by microglia, may be promising candidates to ameliorate neuronal dysfunction and cognitive impairment.
However, the functions of microglia include not only proinflammatory cytokine release, but also phagocytosis of Aβ, tau propagation via exosomes, synaptic pruning, and ROS production, all of which are associated with AD pathology. Therefore, agents that regulate both these functions and the production of proinflammatory cytokines are considered to exert therapeutic effects. However, it is difficult to determine which functions need to be regulated and to what extent because the contribution of each function to AD pathology is unknown. Therefore, it is necessary to identify promising targets through approaches obtained from new perspectives. One such approach is to directly select genes that are highly related to AD symptoms and pathology using a genetic approach. These selected factors may maximally contribute to AD pathology by strongly regulating specific functions of microglia or broadly regulating multiple functions. Although AD-related factors have been identified by GWASs, the impact of each gene on AD is small, and genes with a large impact result in only a small number of patients with that genotype. One possible solution is to analyze the interactions of multiple SNPs from a major AD population, which may identify pathologically important signals based on combinations of SNPs with a large impact. Another approach is to identify microglial subsets that are closely related to AD pathology and to find the regulators of these subsets. Clinical and non-clinical studies using novel technologies, such as single-cell genomics and spatial transcriptomics, have dramatically advanced our understanding of AD.203) Importantly, recent studies utilizing these new technologies have identified disease-associated glial cell populations with unique transcriptional signatures.204,205) In the future, it may be possible to identify and regulate microglial subsets that are closely associated with AD symptoms and pathology. These efforts may enable the amelioration of microglial dysfunctions that significantly contribute to AD pathology. In addition, novel cerebrospinal fluid or peripheral blood biomarkers that reflect such disease- or symptom-relevant microglial phenotypes—and the classification of patients with AD according to these biomarkers—will help identify patient populations that can benefit from microglia-targeted interventions. Furthermore, disease- or symptom-associated biomarkers could also be utilized to monitor therapeutic effects during treatments, potentially leading to shorter and smaller clinical trials. Recent studies strongly indicate that neuroimmune communication at the CNS borders is critically involved in CNS diseases.206) Therefore, in addition to CNS-resident glial cells, other immune cells, such as T cells, B cells, and monocytes/macrophages, will be candidate therapeutic immune targets of AD. Overall, novel findings linking neuroimmune dysfunction and AD as well as other forms of dementia highlight the potential of immune-targeted therapeutic approaches. Further studies will be required to achieve immune-based precision therapies for dementia and to identify and target disease- or symptom-specific neuroimmune dysfunctions.
We thank Lisa Kreiner, Ph.D., for editing a draft of this manuscript.
All authors are employees of Shionogi & Co., Ltd. The authors declare no conflict of interest.