Immune system roles in the body and brain in Alzheimer's disease development

CD4⁺ T cells exhibit remarkable plasticity, capable of differentiating into distinct effector subsets in response to specific cytokine environments [30]. These subsets can be broadly categorized into pro-inflammatory and anti-inflammatory phenotypes, each playing distinct roles in AD pathogenesis [31]. The pro-inflammatory subsets, primarily T helper type 1 (Th1) and Th17 cells, contribute to BBB disruption and enhanced microglial activation, thereby exacerbating AD-associated neuropathology [32, 33]. In contrast, anti-inflammatory subsets, including Th2 cells and regulatory T cells (Tregs), exhibit neuroprotective properties and can attenuate neuroinflammation [34, 35]. This functional dichotomy of CD4⁺ T-cell responses underscores the intricate immunological landscape in AD and suggests potential targets for intervention.

Macrophages, including border-associated macrophages (BAMs) and monocyte-derived macrophages (MDMs), play crucial roles in AD pathogenesis [49 –52]. BAMs, residing in cerebral border regions, are vital for maintaining brain homeostasis by facilitating glymphatic system clearance [50 –53]. However, age-related alterations in BAMs, such as enhanced major histocompatibility complex class II (MHC-II) expression and impaired extracellular matrix degradation, contribute to vascular and glymphatic dysfunction, thereby increasing the risk of AD [52, 54, 55]. In AD mouse models with amyloid pathology, BAM depletion exacerbates Aβ deposition and microglial activation [50].

MDMs accumulate with age and participate in Aβ clearance [56, 57], however, their capacity to take up Aβ diminishes with age and AD progression, partially due to decreased expression of phagocytic receptors, such as TREM2 [56]. Genetic variations in TREM2 are associated with an increased risk of AD onset [58]. TREM2⁺ macrophages cluster around amyloid plaques, while TREM2-deficient mice exhibit reduced inflammation and attenuated pathology [59]. In TREM2-deficient AD mouse models, the number of CD45^hiLy6C⁺ macrophages is substantially decreased, resulting in attenuated inflammation and ameliorated amyloid and tau pathology [60].

Targeting the immune checkpoint blockade by anti-PD-1 antibodies results in the recruitment of MDMs to the brain, which subsequently facilitates the clearance of amyloid plaques and restores cognitive function in 5×FAD mice [61]. Notably, the effects of PD-1/PD-L1 blockade extend beyond MDMs, as it also disrupts negative signaling between IFNγ-producing T cells and their target cells, resulting in broad enhancement of immune responses [61]. Building on this approach, IBC-Ab002, a humanized IgG1 antibody that inhibits the related immune checkpoint protein PD-L1, has now entered phase 1 clinical trials (NCT05551741) [45].

Transcriptomic analyses and immunohistochemical characterization of postmortem human hippocampal tissue revealed that monocyte-derived macrophages infiltrate the brain parenchyma in late-stage AD [62] (Fig. 2). Nonetheless, replacing brain-resident microglia with peripheral monocytes does not affect the amyloid plaque load in AD mouse models with amyloid pathology [63, 64]. This finding prompts inquiries regarding the distinct functions of peripheral monocytes and microglia in the clearance of Aβ.

Natural killer cells, key components of the innate immune system, play complex and seemingly paradoxical roles in AD pathogenesis. Depletion of NK cells in the 3×Tg-AD mouse model has neuroprotective effects, including attenuated neuroinflammation, enhanced neurogenesis, and improved cognitive function [65]. Conversely, augmentation of NK cell function in the APP/PS1 mouse model can mitigate cerebral Aβ deposition and enhance cognitive performance [66]. These contrasting findings underscore the multifaceted nature of NK cells in AD. Recent clinical investigations, notably the ASK-AD trial (NCT04678453), have explored the therapeutic potential of autologous NK cell therapy (SNK01) in AD patients. Preliminary results indicate that SNK01 treatment may lead to reductions in cerebral Aβ and tau protein levels, coupled with amelioration of neuroinflammation.

Neutrophils play a pivotal role in cerebral hypoperfusion, a well-established phenomenon in AD, by adhering to capillary segments, which leads to reduced and eventually obstructed cerebral blood flow (CBF) [67]. Additionally, neutrophils compromise BBB integrity by disrupting tight junction proteins such as occludin and claudins, thereby promoting increased BBB permeability [68]. In transgenic AD models, such as 5×FAD and 3×Tg-AD mice, neutrophils extravasate and accumulate in regions of amyloid deposition, accompanied by the formation of neutrophil extracellular traps (NETs) and expression of IL-17 [69]. Aβ42 peptide induces a high-affinity conformation of lymphocyte function-associated antigen 1 (LFA-1), promoting rapid neutrophil adhesion to integrin ligands, which subsequently exacerbates AD pathology and cognitive decline [69]. Therapeutic approaches targeting neutrophils have shown promising results. Transient neutrophil depletion during early disease stages results in sustained cognitive improvements, while administration of Ly6G antibodies or LFA-1 inhibitors enhances cerebral blood flow, alleviate AD-like neuropathology, and improve cognitive function [67, 69]. Moreover, LFA-1-deficient transgenic AD mice demonstrate resistance to cognitive deterioration and exhibit reduced gliosis [69]. LFA-1 is not only crucial for NK cell responses but is also expressed by other peripheral immune cells, including T cells, where it plays essential roles in activation, adhesion, and migration [68 –70]. The involvement of LFA-1 in these diverse immune cell functions suggests its potential significance in AD pathology and warrants further investigation.

B cells in AD patients became activated and accumulate in the periphery and brain parenchyma, where they produce immunoglobulins that target Aβ [70]. While B-cell-derived immunoglobulins that target Aβ may impede plaque formation and disease progression, they may concurrently impair microglial function and exacerbate AD pathology [70]. The role of B cells in AD extends beyond immunoglobulin production, as evidenced by the correlation between increased numbers of activated B cells in cervical lymph nodes from aged 3×Tg AD mice, which rise as the disease progresses [71]. B-cell depletion in 5×FAD and APP/PS1 mice is associated with attenuated disease progression, amelioration of cognitive and motor deficits, and a reduction in amyloid burden [71]. Paradoxically, B-cell depletion during the early disease stages accelerates cognitive decline and increases amyloid burden in AD mouse models, highlighting the complex and stage-dependent role of B cells in AD pathogenesis [72, 73].

Mucosal-associated invariant T (MAIT) cells, a subset of innate-like T cells, recognize microbial-derived metabolites, particularly those derived from vitamin B, through the major histocompatibility complex class I (MHC-I)-related protein MR1 [74]. In the 5×FAD mouse model, MAIT cell numbers progressively increase and display activation signatures [75]. Both AD patients and 5×FAD mice exhibit significantly elevated MR1 expression in microglia surrounding amyloid plaques [76]. Notably, amyloid plaque formation is markedly reduced in MR1-deficient mice [76].

Gamma delta T (γδT) cells, another subset of innate immune cells, serve as the primary source of interleukin-17 A (IL-17) in healthy meninges, where they support synaptic plasticity in CA1 hippocampal neurons and facilitate short-term memory [77]. However, in the 3×Tg AD mouse model, γδT cells accumulate substantially in the brain parenchyma and meninges, correlating with cognitive deterioration [37]. Notably, IL-17-producing γδT cells show marked increases in the CNS of these mice, with their accumulation associated with early short-term memory deficits and synaptic dysfunction [37]. These findings suggest that γδT cells may exacerbate neuroinflammation and contribute to the early AD pathological progression.

Pro-inflammatory cytokines, including interleukins (ILs), tumor necrosis factor (TNF) and interferons (IFNs), are predominantly secreted by peripheral immune cells [78]. In AD patients, these cytokines are markedly elevated in the plasma, facilitating their redistribution across a leakage-prone BBB associated with neurodegenerative progression [78].

Interferons are classified into Type I (IFN-I) and Type II (IFN-II, specifically IFN-γ) subtypes. IFN-I induces microglial activation and synaptic pruning, whereas selective inhibition of interferon α/β receptor (IFNAR) through specific blockers or neutralizing antibodies mitigates microgliosis and prevents synaptic loss in 5×FAD mouse model [79, 80]. IFN-II exhibits a biphasic role in neurodegenerative diseases. At physiological levels, IFN-γ enhances microglial phagocytosis, promotes Aβ clearance, and augments cognitive functions [81]. Additionally, IFN-γ-mediated immune responses recruit monocyte-derived macrophages to the CNS, facilitating amyloid plaque clearance and cognitive improvement [81]. However, chronic IFN-γ-induced inflammation compromises BBB integrity and allows peripheral inflammatory mediators to infiltrate the CNS, thereby exacerbating neuroinflammation, neuronal damage and cognitive decline [81, 82].

The pro-inflammatory milieu induces the upregulation and activation of NLRP3, forming the NLRP3 inflammasome and resulting in substantial production of pro-inflammatory cytokines [83]. In AD, NLRP3 inflammasome activation occurs predominantly in innate immune cells, particularly microglia in the CNS [83]. Hyperactivation of the inflammasome triggers an inflammatory cascade that exacerbates neuronal injury. Studies using 16-month-old APP/PS1 transgenic mice demonstrated that Nlrp3 deficiency markedly enhanced microglia-mediated Aβ phagocytosis, reduced amyloid plaques, and improved spatial memory [84]. Furthermore, in transgenic Tau22 mice, Nlrp3 deficiency attenuated tau hyperphosphorylation and aggregation while ameliorating Aβ-induced tau pathology [85].

B2M, a key component of MHC-I complex, exists in circulation as a non-covalently associated monomeric protein that is predominantly cleared through renal glomerular filtration [86]. Under pathological conditions, such as chronic kidney disease with long-term dialysis, B2M forms amyloid fibrils, leading to amyloidosis and related disorders [86]. Recent research has highlighted the significant involvement of B2M in AD and Down syndrome (DS), where it affects cognitive function through both central and peripheral pathways [10, 87].

In AD, elevated B2M levels promote amyloid pathology through co-aggregation with Aβ, facilitating plaque formation and propagation [10]. This aggregation process enhances neurotoxicity and neuroinflammation, thereby accelerating both neurodegeneration and the spatial expansion of Aβ pathology [10]. Experimental studies in AD mouse models have demonstrated that B2M reduction—achieved through genetic ablation, antisense oligonucleotide (ASO) administration, or antibody-based interventions—markedly attenuates cognitive deficits and reduces amyloid plaque burden [10]. Additionally, the ability of B2M to traverse the BBB presents a promising therapeutic opportunity [10]. Notably, peripheral B2M clearance via systemic administration of low-dose anti-B2M antibodies effectively reduces brain B2M levels and improves cognitive function in AD mouse models, offering a potential alternative to high-dose Aβ antibody therapies, which frequently induce adverse effects [10].

Individuals with DS frequently develop AD by middle age, suggesting shared pathogenic mechanisms between these conditions [88]. In DS, elevated plasma B2M crosses the BBB and suppresses N-methyl-D-aspartate (NMDA) receptor activity, disrupting synaptic excitatory/inhibitory homeostasis and precipitating cognitive dysfunction [87]. Mechanistic studies have revealed that B2M directly interacts with the GluN1 subunit of NMDA receptors, attenuating receptor function. The development of GluN1-P2, a therapeutic peptide that competitively inhibits this B2M-GluN1 interaction, has demonstrated significant improvements in synaptic plasticity and cognitive performance in both DS and aging mouse models [87]. Through parabiosis studies and plasma transfer experiments, elevated circulating B2M has been identified as a critical mediator of cognitive impairment in both DS patients and mouse models [87]. Therapeutic interventions targeting peripheral B2M, including antibody neutralization and parabiotic circulation exchange, successfully reverse these cognitive deficits.

As a systemic aging factor, B2M levels increase with age in the plasma and CSF, contributing to cognitive decline in aging and neurodegenerative disorders [89]. The co-aggregation of B2M with Aβ enhances neurotoxicity, necessitating a refinement of the amyloid cascade hypothesis. Furthermore, B2M's pathological interaction with NMDA receptors establishes it as a promising therapeutic target. These mechanistic insights provide a foundation for developing small-molecule inhibitors and antibody-based therapeutics targeting B2M-mediated pathology [10, 87].

The APOE4 gene variant encoded on human chromosome 19, is the most important genetic risk factor for late-onset AD [90]. APOE4 is characterized by arginine residues at positions 112 and 158 (Arg112/Arg158), distinguishing it from APOE3 (Cys112/Arg158) and APOE2 (Cys112/Cys158) [90]. APOE4 impacts both the CNS and peripheral immunity. In cognitively normal APOE4 carriers, the presence of plasma proteins in brain tissue and CSF suggests the presence of compromised BBB integrity prior to cognitive decline [91]. APOE4 potentially weakens endothelial tight junctions, increases BBB permeability and facilitates neuroinflammation [92], and enhances BBB breakdown by reducing the capillary basement membrane area and increasing thrombin levels in vessel walls and perivascular spaces [93, 94]. APOE4 also affects pathogenic changes in cerebral vasculature by promoting Aβ deposition and exacerbating CAA, further compromising BBB integrity and AD neuropathology [13]. Moreover, APOE4 impairs BBB-mediated Aβ clearance, shifting Aβ40/42 clearance from rapid clearance pathways mediated by low-density lipoprotein receptor-related protein 1 (LRP1) to slower clearance pathways mediated by the very low-density lipoprotein (VLDL) receptor pathway [95]. Conversely, APOE2 and APOE3 facilitate faster Aβ clearance at the BBB [95]. CAA progression alters Aβ clearance mechanisms and vascular pathology, including increased vascular pulsatility and reduced vascular smooth muscle cell coverage [13]. APOE4 also attenuates Aβ clearance by promoting premature meningeal lymphatic vessel atrophy [96].

APOE4 accelerates immunosenescence in neutrophils, resulting in the infiltration of immunosuppressive IL-17-producing neutrophils in the brains of female APOE4 carriers with AD. In human APOE4 knock-in APP/PS1 mice, genetic ablation of APOE4 in neutrophils diminishes their immunosuppressive phenotype and reduces IL-17 signaling. This alteration restores the microglial response, alleviates amyloid pathology, and enhances cognitive function in these AD transgenic mice [97]. Studies using mouse models expressing human APOE3 or APOE4 in the liver have demonstrated that liver-expressed APOE4 exacerbates cerebral amyloid pathology, whereas APOE3 reduces it. These findings suggest that peripheral APOE4 impairs cerebrovascular function, affecting synaptic plasticity and cognition, thus providing a rationale for targeting peripheral APOE4 in AD therapy [92]. Therapeutic strategies targeting APOE4, such as angiotensin receptor blockers, vascular endothelial growth factor A (VEGFA), and epidermal growth factor (EGF), can potentially improve BBB integrity and function and slow AD progression [98].

APOE is primarily synthesized and secreted by astrocytes and activated microglia in the CNS [90, 99]. In the PS19-APOE4 mouse model, astrocyte-specific APOE4 deletion reduces disease-associated gene expression across multiple cell types and suppresses aberrant synaptic phagocytosis [100, 101]. Neuron-specific APOE4 deletion markedly reduces tau pathology and ameliorates neurodegeneration [102], while microglial APOE4 deletion enhances Aβ clearance in APP/PS1 mice [103]. These cell type-specific effects reveal distinct contributions of APOE4 to AD pathogenesis.

TREM2, a member of the immunoglobulin receptor superfamily, is predominantly expressed in osteoclasts, macrophages, and microglia [104]. It activates downstream signaling pathways, including Ca²⁺ mobilization, MAPK, and mTOR, by binding to various ligands, such as phospholipids, APOE, Aβ oligomers, and apoptotic neurons [105 –107]. In mouse models of AD, TREM2 exhibits neuroprotective effects primarily through its expression in microglia, where it serves as a critical regulator of immune responses in neurodegenerative diseases [108]. Studies in AD mouse models have also shown that Trem2 knockout (Trem2^−/−) or the presence of the TREM2 R47H mutation results in reduced plaque-associated microglia and increased dystrophic neurites [109]. Microglial reduction exacerbates Aβ toxicity and tau seeding and spreading [110]. Conversely, TREM2-activating antibodies enhance microglial metabolism and function, which leads to reduced amyloid plaque formation [111].

TREM2 undergoes ectodomain shedding at histidine 157 by α-secretases ADAM10 and ADAM17, generating soluble TREM2 (sTREM2) [112 –114]. sTREM2 promotes microglial survival and plaque-associated clustering while enhancing both Aβ clearance and cytokine production [115]. Both TREM2 and sTREM2 significantly affect BBB function [116, 117]. Loss of TREM2 function or mutation can exacerbate neuroinflammation and neurodegenerative progression by increasing BBB permeability, thereby allowing peripheral immune cells and other inflammatory factors to enter the CNS [118].

As a potential link between the CNS and peripheral immune system, changes in sTREM2 levels can reflect the inflammatory state and severity of neurodegenerative changes within the CNS [119]. In AD patients, elevated sTREM2 levels in CSF and plasma are associated with early disease stages and mild cognitive impairment [120]. Increased sTREM2 correlates with memory decline and hippocampal atrophy [121]. Additionally, sTREM2 levels positively correlate with CSF tau and p-tau levels, indicating its potential applicability as a biomarker for AD progression [122].

CD22, which was originally characterized as a B-cell receptor, has been found to exert important regulatory effects on microglia within the CNS [123]. Microglia maintain CNS homeostasis through phagocytic clearance of protein aggregates and cellular debris; however, this function deteriorates with age and in various neurodegenerative disorders associated with cognitive decline [123]. Studies have demonstrated that CD22 functions as a negative regulator of microglial phagocytosis and is upregulated in aged microglia [123]. CD22 is associated with exacerbated neuroinflammation and neurodegenerative changes through its anti-phagocytic effects derived from CD22 interactions with α2,6-linked sialic acid, which inhibits microglial phagocytosis and promotes the formation and propagation of amyloid plaques [123].

Soluble CD22 (sCD22) serves as a marker of inflammation and microglial dysfunction [124]. Elevated sCD22 levels have been detected in the peripheral blood of AD patients, are negatively correlated with CSF Aβ42 levels and Aβ42/Aβ40 ratios, and are positively correlated with phosphorylated tau levels and amyloid burden in the brain [125]. Furthermore, elevated plasma sCD22 levels are associated with accelerated cognitive decline, suggesting that sCD22 may accelerate the progression of various neurodegenerative disorders [125].

In the periphery, CD22 regulates B-cell activation and immune responses through interactions with B-cell receptors [126]. The upregulation of CD22 in AD patients may lead to dysregulation of the peripheral immune system, exacerbating inflammation in the CNS. Studies have shown that inhibition of CD22 activity can increase microglial phagocytosis activity and improve clearance of myelin debris, Aβ oligomers, and α-synuclein fibrils, thereby improving cognitive function in murine models [123]. Consequently, CD22 and its associated pathways represent potential therapeutic targets and offer novel strategies for mitigating or halting the progression of various neurodegenerative disorders.