Controlling brain immune cell inflammation in Alzheimer's disease: how it works and possible treatments

TREM2, a type I transmembrane receptor predominantly found on microglia within the CNS, stands as arguably the most critical genetic and therapeutic target in AD pathogenesis. Genetic variants, notably the R47H mutation, significantly elevate the risk of LOAD, acting primarily as loss-of-function mutations that impair microglial defense mechanisms (70). TREM2 is indispensable for numerous essential microglial functions, including survival, proliferation, chemotaxis, phagocytosis, and maintaining lipid metabolic fitness (71). Mechanistically, TREM2 engagement is crucial for driving the transition from homeostatic microglia to the protective DAM phenotype (specifically Stage 2 activation). This state acquires robust phagocytic capabilities essential for engulfing pathological proteins like Aβ and tau, thereby restricting plaque seeding and spread. This signaling axis also regulates inflammation and microglial phenotype shift via pathways including SYK and PI3K/AKT/mTOR cascades (57). Consequently, therapeutic agonism aimed at boosting TREM2 signaling constitutes a major strategy to enhance phagocytosis, clear pathology, and restore neuroimmune homeostasis.

This therapeutic direction has fueled the intensive development of agonistic mAbs. Leading candidates such as AL002 (including variants AL002a/c), Ab-T1, Ab18 and 4D9 have been investigated to enhance TREM2 function (72). These agonistic antibodies typically function by promoting receptor cross-linking to amplify the downstream DAP12/SYK signaling cascade, consequently boosting microglial proliferation and phagocytic activity. A related strategy employed by antibodies like AL002c and 4D9 involves binding near the ADAM10/ADAM17 cleavage site on the extracellular stalk region. This binding inhibits ectodomain shedding, thereby stabilizing full-length TREM2 on the cell surface and terminating the generation of soluble TREM2 (sTREM2). Preclinical studies demonstrated that these TREM2 agonists reduced plaque burden and ameliorated cognitive deficits in various AD mouse models by promoting migration and phagocytosis. However, the recent Phase II INVOKE-2 trial evaluating AL002 (NCT04592874↗) did not meet its primary endpoints, failing to significantly slow clinical progression or reduce amyloid burden, despite achieving target engagement (73). This outcome, coupled with preclinical findings that chronic TREM2 activation can potentially exacerbate Aβ-associated tau seeding and spreading, underscores the complexities and context-dependent nature of TREM2-targeted therapy.

To circumvent the limitations associated with large therapeutic biologics, particularly poor blood-brain barrier (BBB) penetration, researchers are actively developing brain-penetrant small-molecule TREM2 agonists. VG-3927 (Vigil Neurosciences) represents a key small-molecule candidate currently undergoing Phase I clinical evaluation (NCT06343636↗) (56). This orally bioavailable compound is notable for functioning as a molecular glue, selectively potentiating the function of membrane-bound TREM2 in response to natural ligands without directly binding to the decoy sTREM2 fragment (74). Initial pharmacokinetic/pharmacodynamic (PK/PD) studies for VG-3927 demonstrated brain penetrance and pharmacological activity in the CNS (75). Beyond synthetic compounds, preclinical data also supports the potential of natural compounds, such as the plant alkaloid Hecubine and extracts from Pyrolae Herba, which directly activate TREM2 signaling and mitigate neuroinflammation (76). Ultimately, optimizing the precise therapeutic window for TREM2 agonism—likely initiated during early, protective pathological stages—is crucial to maximize protective phagocytic clearance while mitigating the risk of inadvertent pro-inflammatory or detrimental activation observed later in the disease course. Crucially, recent breakthrough evidence provides a mechanistic explanation for this detrimental late-stage activation, revealing that chronic engagement of the TREM2-APOE signaling axis can paradoxically drive microglia into a dysfunctional, senescent state. This suggests that sustained TREM2 activation over time may shift the microglial phenotype from protective clearance to senescence-associated chronic inflammation (77).

Modulating microglial function by targeting inhibitory surface receptors provides a critical strategy complementary to promoting activation, aimed specifically at engaging clearance mechanisms by lifting the physiological “brake” imposed by CD33. CD33, a sialic acid-binding immunoglobulin-like lectin (Siglec-3), is encoded by a key AD susceptibility gene and is predominantly expressed on microglia and other myeloid cells (78). Its inhibitory function stems from its intracellular Immunoreceptor Tyrosine-based Inhibition Motif (ITIM) domain, which, upon phosphorylation, recruits phosphatases that suppress core microglial activities such as phagocytosis and proliferation (79). This mechanism is highly relevant to AD pathology, as elevated CD33 expression correlates strongly with impaired Aβ clearance and increased plaque burden (80). Conversely, genetic evidence supports the benefit of counteracting this brake: individuals carrying protective CD33 alleles exhibit reduced functional expression and a significantly lower risk of developing AD (80).

Consequently, inhibition of CD33 represents a compelling therapeutic avenue to restore microglial phagocytic capacity. Preclinical models validate this approach, demonstrating that CD33 knockout significantly enhances microglial phagocytosis of Aβ aggregates, resulting in reduced amyloid pathology and improved cognitive function in 5xFAD mice (78). Critically, emerging evidence suggests that the protective effect of relieving the CD33 brake is contingent upon the presence of functional TREM2 signaling (78). This establishes that TREM2 acts downstream of CD33 in modulating microglial pathology, indicating that the therapeutic benefit of CD33 inhibition can only be fully realized when the TREM2-dependent clearance machinery is intact. Therapeutic candidates pursuing this strategy include small-molecule inhibitors, such as the sialic acid mimetic P22, and antagonistic mAbs like AL003 and Lintuzumab (81). While trials for agents like AL003 have highlighted the challenges of clinical translation, the strategy of blocking CD33 activity to unleash microglial clearance remains a priority in drug discovery.

The TAM receptor family—comprising Tyro3, Axl, and MerTK—orchestrates a critical, yet distinct, axis of microglial regulation focused on efferocytosis and the clearance of membrane debris. Unlike the constitutively expressed receptors involved in general surveillance, the TAM family exhibits a distinct expression shift during AD progression: MerTK is predominant in homeostatic microglia and mediates the immunologically “silent” clearance of apoptotic cells, whereas Axl is normally quiescent but undergoes dramatic upregulation in plaque-associated microglia (82). These receptors function through a unique bridging mechanism, binding to ligands such as growth arrest-specific 6 (Gas6) and Protein S (Pros1), which in turn recognize “eat-me” signals like exposed phosphatidylserine (PtdSer) on the surface of apoptotic neurons or amyloid aggregates (21).

The contribution of TAM receptors to AD pathology is characterized by a complex functional dichotomy that challenges straightforward therapeutic intervention. Mechanistically, TAM signaling is indispensable for the recognition and engulfment of amyloid plaques. Studies employing Axl/MerTK double-knockout models have revealed that the loss of this signaling pathway results in a failure of microglia to engage with plaques, leading to the accumulation of uncompacted, “fluffy” amyloid deposits and severe dystrophic neurites (21). This evidence suggests that TAM-driven phagocytosis is essential for plaque compaction—a process that creates a physical barrier to contain neurotoxicity. However, this protective function presents a paradox: the very machinery that contains toxicity also promotes the consolidation of dense-core plaques. While these dense plaques are generally considered less neurotoxic than diffuse oligomers, they remain a pathological hallmark, raising the question of whether therapeutic modulation should aim to inhibit plaque formation or enhance compaction.

Emerging evidence provides a compelling consensus to this question: therapeutic interventions should aim to preserve or augment, rather than suppress, TAM-mediated plaque compaction. The initial hypothesis—suggesting that inhibiting upregulated Axl could beneficially mitigate plaque-associated inflammation—has been substantially challenged by recent in vivo models. Given that the abrogation of TAM signaling actively exacerbates neuritic dystrophy and neurotoxicity, the therapeutic paradigm is currently pivoting from broad receptor antagonism toward precision enhancement and targeted modulation.

One promising strategy involves stabilizing the full-length receptors on the microglial surface. During disease progression, metalloproteases such as ADAM10 and ADAM17 cleave Axl and MerTK into soluble forms (e.g., sAxl), which act as decoy receptors that sequester endogenous ligands like Gas6, thereby interrupting the protective signaling cascade (83). Inhibiting these metalloproteases could theoretically restore microglial phagocytic capacity. Furthermore, emerging preclinical evidence highlights the potential of small molecules to therapeutically engage this system. For instance, Ganoderic acid A has been shown to relieve Aβ burden by enhancing Axl-mediated autophagy, and Jujuboside A promotes Aβ clearance via the Axl/PPARγ axis (84, 85). Ultimately, both interventions ameliorate cognitive deficits in animal models.

Perhaps the most innovative translational approach utilizes the TAM system to decouple Aβ clearance from neurotoxic inflammation. Recent engineered biologics have successfully fused single-chain anti-Aβ antibodies to the TAM-receptor-binding domain of Gas6. This novel chimeric strategy selectively directs Aβ to TAM receptors, inducing robust microglial phagocytosis without triggering the classical inflammatory response or risking cerebral amyloid angiopathy (86). Consequently, future therapeutic strategies targeting the TAM family must move beyond simplistic agonism or antagonism. Success will rely on sophisticated interventions that harness MerTK and Axl to facilitate the immunologically “silent” clearance of pathological fibrils, thereby restoring neuroimmune homeostasis without collateral inflammatory damage.

The CX3CL1-CX3CR1 axis serves as a fundamental homeostatic checkpoint for neuron-microglia communication, distinguished by its high-affinity, one-to-one fidelity in the CNS (87). In the healthy brain, neuronal CX3CL1 (Fractalkine) binds to microglial CX3CR1, functioning as a constitutive “calming signal” that inhibits pro-inflammatory pathways, such as NF-κB, and maintains microglia in a surveillance state (88). Emerging mechanistic insights have revealed that this interaction is bidirectional: beyond the “forward signaling” that regulates microglial phagocytosis and oxidative stress via Nrf2, a “back-signaling” mechanism exists where the intracellular domain (ICD) of CX3CL1 translocates to the neuronal nucleus, promoting cell survival and neurogenesis through TGF-β/Smad pathways (89, 90).

In the context of Alzheimer’s pathology, this axis presents a complex “Janus-faced” paradox. Genetic deficiency or disruption of CX3CR1 signaling acts as a “release of the brake,” which facilitates the phagocytic clearance of fibrillar amyloid plaques; however, this disinhibition inadvertently increases the accumulation of neurotoxic soluble oligomers and exacerbates synaptic loss (91, 92). Crucially, the integrity of this axis is unequivocally protective against tau pathology. Disruption of signaling amplifies a p38 MAPK-IL-1β inflammatory loop that drives GSK3β-mediated tau hyperphosphorylation, while simultaneously impairing the microglial capacity to internalize and degrade extracellular tau seeds (53, 93, 94).

Given these divergent roles, therapeutic modulation requires a precise approach that avoids the pitfalls of broad antagonism. While CX3CR1 antagonists may reduce peripheral inflammation, they risk accelerating tau propagation and cognitive decline within the CNS (95). Consequently, the most promising strategies involve agonism to restore this neuroprotective “handshake.” Interventions utilizing viral vectors to deliver soluble Fractalkine (sFKN) or its C-terminal fragment have demonstrated efficacy in reducing tau pathology and preventing neurodegeneration without compromising amyloid clearance mechanisms (89, 96, 97). Future development aims to identify small-molecule agonists or gene therapies that can selectively boost these signaling pathways to decouple neuroinflammation from neurodegeneration.

The NLRP3 inflammasome acts as a critical convergence point for the multiple pathological insults characterizing the AD brain, functioning not merely as a responder to pathology but as an active generator of it. Structurally, this cytosolic complex consists of the sensor molecule NLRP3, the adaptor protein ASC, and the effector pro-caspase-1. Its activation in microglia is tightly regulated by a two-step mechanism comprising priming and activation. In the AD context, the initial priming signal triggers NF-κB-mediated transcriptional upregulation of NLRP3 and pro-IL-1β via Aβ interaction with PRRs (98, 99). While this prepares the cell, the catalytic trigger arises from metabolic and physical stress. Specifically, the “frustrated phagocytosis” of fibrillar Aβ destabilizes lysosomes, releasing cathepsin B, while P2X7R activation by extracellular ATP (eATP) drives potassium efflux. Simultaneously, mitochondrial dysfunction generates reactive oxygen species (mtROS) and releases oxidized mitochondrial DNA (mtDNA), which are sensed by the NACHT domain of NLRP3, initiating rapid complex assembly (100, 101).

Once assembled, the inflammasome drives pathology through unique downstream mechanisms, most notably via the formation of ASC specks. Released into the extracellular space following pyroptosis, these specks exhibit prion-like bioactivity. Driven by specific electrostatic interactions and molecular mimicry, the ASC pyrin domain (PYD) binds Aβ, significantly accelerating its aggregation and spreading plaque deposition in a cross-seeding manner (102, 103). Furthermore, NLRP3 activation creates a feed-forward loop with tau pathology. The release of IL-1β enhances the activity of tau kinases (e.g., GSK-3β) while suppressing phosphatases. Recent evidence further indicates that tau monomers can activate NLRP3, which reciprocally promotes tau acetylation and aggregation, thereby driving neurofibrillary tangle formation and hippocampal atrophy (104).

The terminal output of this cascade is pyroptosis, a lytic form of cell death mediated by Gasdermin D (GSDMD). Active caspase-1 cleaves GSDMD, liberating its N-terminal domain to form large non-selective pores in the plasma membrane. These pores serve as the necessary conduit for the release of cytokines IL-1β and IL-18, which lack secretory signal peptides, fueling a cytokine storm (105, 106). Beyond cytokine release, GSDMD pore formation compromises the integrity of microglia and, as recent high-resolution studies suggest, surrounding neurons. The resulting loss of synaptic integrity due to GSDMD cleavage directly links the inflammatory machinery of the inflammasome to the neurodegeneration and cognitive decline observed in AD (107).

Targeting the NLRP3 inflammasome offers a precision medicine approach that addresses both the seeding of proteinopathies and the neurotoxic inflammatory response while sparing other innate immune pathways essential for host defense. The prototype for this strategy, MCC950 (CRID3), specifically binds to the Walker B motif of the NLRP3 NACHT domain, locking the protein in an inactive conformation (108, 109). While MCC950 successfully reversed cognitive deficits and altered microglial phenotypes in APP/PS1 mice, its development was halted due to hepatotoxicity and limited BBB penetrance (110, 111). Nevertheless, it paved the way for safer candidates like OLT1177 (Dapansutrile) and CY-09. OLT1177 inhibits NLRP3 ATPase activity without affecting transcriptional priming and has demonstrated efficacy in restoring synaptic plasticity in AD models (112). Similarly, CY-09 targets the ATP-binding motif to block assembly and restore cerebral glucose metabolism, while the repurposed anti-allergic drug Tranilast offers a safe, immediate option for inhibiting NLRP3 oligomerization (113).

The field is now advancing towards “next-generation” inhibitors explicitly optimized for CNS pharmacokinetics and potency. Leading this wave is NT-0796, which utilizes an innovative ester prodrug mechanism. Being lipophilic, NT-0796 efficiently crosses the BBB and cell membranes, where intracellular esterases convert it into a less permeable active acid species, effectively “locking” the drug at the target site. In a recent Phase 1b/2a trial for Parkinson’s disease, NT-0796 not only reduced inflammatory cytokines in Cerebrospinal Fluid (CSF) but also significantly lowered levels of Neurofilament Light Chain (NfL), providing the first clinical proof-of-concept that oral NLRP3 inhibition can attenuate neuronal damage in humans (114).

In parallel, other industry candidates are leveraging structural biology to enhance selectivity and potency. VENT-02, discovered using a proprietary structural platform, has demonstrated a “best-in-class” potency profile with full target engagement and excellent brain penetrance, currently progressing through Phase 2 trials. Additionally, Roche’s Selnoflast is utilizing translocator protein (TSPO) positron emission tomography (PET) imaging in clinical trials to directly visualize its effect on microglial activation in the living brain (110). These advancements signal a shift from preclinical validation to clinical translation, where potent, brain-penetrant NLRP3 inhibitors are poised to modify the disease trajectory by breaking the cycle of neuroinflammation and neurodegeneration.

The P2X7R acts as a critical sentinel within the microglial neuroimmune response, serving as the primary link between extracellular damage signals and the explosive activation of the NLRP3 inflammasome. Unlike other purinergic receptors that function at physiological ATP levels, P2X7R has a high activation threshold, responding specifically to the millimolar concentrations of eATP found in the “purinergic halo” surrounding necrotic cells and amyloid plaques.1 Upon sustained activation by this pathological stimulus, the receptor undergoes a structural dilation to form a non-selective “macropore,” which facilitates massive potassium (K⁺) efflux and calcium (Ca²⁺) influx (115, 116). This ionic deregulation is the indispensable second signal required for NLRP3 inflammasome assembly, effectively positioning P2X7R as the “upstream trigger” that drives the maturation and release of IL-1β and IL-18 (117).

In the context of AD pathology, P2X7R function shifts from homeostatic maintenance to neurodegenerative toxicity. Histological studies confirm that the receptor is aggressively upregulated in the AD brain, particularly in activated microglia clustering around senile plaques, where expression levels correlate with amyloid load and disease severity (118, 119). While basal P2X7R activity acts as a “scavenger” facilitating cytoskeletal reorganization for phagocytosis, the chronic hyperactivation observed in AD creates a functional paradox. The formation of the transmembrane macropore induces pyroptosis—a highly inflammatory form of cell death—and disrupts the metabolic machinery required for engulfment (117). Consequently, overstimulated microglia become effectively paralyzed; they are unable to clear amyloid debris while simultaneously propagating a feed-forward inflammatory loop through lysosomal destabilization and cytokine release (115, 120).

Therapeutic strategies targeting P2X7R aim to sever this connection between damage sensing and the inflammatory cascade without compromising the receptor’s basal physiological roles. Early interventions using the dye-derivative Brilliant Blue G (BBG) successfully reduced neuronal loss and intracellular Aβ accumulation in mouse models, though its clinical translation was hindered by toxicity and species-specific potency issues (121, 122). The field has since advanced to second-generation, BBB-penetrant antagonists such as the JNJ series (e.g., JNJ-54175446), which have demonstrated robust target engagement and safety in human clinical trials (123). By blocking the receptor’s pore-forming state, these small molecules offer a multifaceted advantage over downstream cytokine inhibitors: they not only halt the release of IL-1β and IL-18 but also impede the exosomal spreading of tau pathology and reduce oxidative stress, potentially functioning as a “circuit breaker” for chronic neuroinflammation (115, 119).

Microglia in AD brain undergo a profound metabolic reprogramming analogous to the Warburg effect in oncology, characterized by a shift from oxidative phosphorylation (OXPHOS) to aerobic glycolysis even under normoxic conditions (124). While initially adaptive for rapid cytoskeletal remodeling, this glycolytic surge becomes maladaptive in the chronic amyloid environment, locking microglia into a pro-inflammatory state driven by the HIF-1α and mTOR signaling pathways (124). Recent epigenetic findings highlight a vicious cycle wherein excess lactate generated by this glycolytic flux promotes histone lactylation (H4K12la) at the promoters of glycolytic genes, effectively cementing this metabolic dysfunction and hindering the cells’ ability to clear pathology (125). Consequently, targeting the enzymatic machinery governing this switch offers a novel avenue to reprogram microglia from a neurotoxic phenotype back to a neuroprotective, phagocytic state.

Among the metabolic regulators, Hexokinase 2 (HK2) has emerged as a critical therapeutic checkpoint due to its pivotal role in gating glucose entry and its specific upregulation in plaque-associated microglia (126). Crucially, therapeutic intervention requires a nuanced approach, often described as a “Goldilocks” effect; complete ablation of HK2 can precipitate metabolic collapse and exacerbate inflammasome activation due to the loss of mitochondrial stability (127). Conversely, partial inhibition—mimicked pharmacologically by agents such as Lonidamine—induces a beneficial metabolic flexibility (126). By constraining glycolytic flux without starving the cell, partial HK2 inhibition compels microglia to utilize alternative fuel sources, specifically driving a shift toward fatty acid oxidation (FAO) and lipid metabolism (128). This metabolic flexibility not only sustains cellular energetics but also enhances the clearance of Aβ via the upregulation of lipid sensors like TREM2, while simultaneously dampening cytokine secretion (126).

Beyond the initial glycolytic entry point, downstream enzymes such as Pyruvate Kinase M2 (PKM2) also present viable targets for modulating immunometabolism. In AD, PKM2 often exists as a dimer that translocates to the nucleus to drive inflammatory gene expression; stabilizing PKM2 into its tetrameric form with small molecules like TEPP-46 can sequester the enzyme in the cytoplasm, thereby restoring metabolic efficiency via the TCA cycle and blocking nuclear inflammatory signaling (129). Collectively, these strategies underscore a paradigm shift in AD therapy: rather than broadly suppressing the immune system, the objective is to metabolically reprogram microglia to correct the bioenergetic deficits that underlie their functional failure. However, translational efforts must account for biological variables, such as the observed sex dimorphism in metabolic susceptibility, where female microglia exhibit a more pronounced glycolytic shift and may require stratified dosing regimens (130).

Recent single-cell transcriptomic analyses have identified LDAMs as a distinct, maladaptive phenotype that differs fundamentally from the neuroprotective DAM signature. Characterized by the aberrant sequestration of neutral lipids—primarily triacylglycerols—into cytosolic organelles, this state represents a critical metabolic bottleneck linking genetic risk factors, particularly APOE4, to immune failure (39). The biogenesis of LDAMs is driven by a specific biochemical cascade: upon encountering fibrillar Aβ, microglia upregulate ACSL1, an enzyme that diverts fatty acids from mitochondrial oxidation toward esterification (131). This process is consolidated by the enzyme Diacylglycerol O-acyltransferase 2 (DGAT2), which catalyzes the formation of triacylglycerols, and PLIN2, which stabilizes the resulting droplets (132). In APOE4 carriers, this metabolic deviation is exacerbated, creating a “genotype-specific” failure where lipids accumulate rather than being transported or metabolized, effectively locking the cell into a dysfunctional state (131).

The functional consequences of this lipid burden are catastrophic for the local neuro-environment, manifesting as a “double hit” of lost protective capacity and gained neurotoxicity. First, LDAMs exhibit severe phagocytic blockade; the physical crowding of the cytoplasm and the metabolic diversion of energy substrates result in cellular “indigestion,” rendering the microglia unable to clear Aβ aggregates or myelin debris (132, 133). Second, these cells transition into a pro-inflammatory senescence-associated state, releasing high levels of cytokines and Reactive Oxygen Species (ROS) due to mitochondrial defects (39, 42). Crucially, LDAMs actively propagate neurodegeneration through dysregulated lipid crosstalk; evidence suggests that APOE4 LDAMs secrete lipotoxic factors that induce tau hyperphosphorylation in neurons, thereby acting as a vector that translates amyloid pathology into tau-mediated toxicity (131, 134).

Therapeutic strategies in this domain focus on “unclogging” the microglial machinery to restore phagocytic function. The most advanced approach involves the direct inhibition of triglyceride synthesis; small molecule inhibitors of DGAT2 have been shown to clear microglial lipid droplets and significantly reduce amyloid plaque load in preclinical models (132). Precision medicine approaches are also exploring the inhibition of ACSL1 (e.g., via Triacsin C) to prevent the initial metabolic diversion in APOE4 carriers, potentially shielding neurons from lipotoxicity without broadly suppressing lipid metabolism (131). Furthermore, strategies that enhance catabolism, such as Transcription Factor EB (TFEB) agonists that promote “lipophagy” (the autophagic degradation of lipid droplets), offer a complementary route to resolve the intracellular lipid backlog and re-engage the cell’s degradative pathways (135).

The functional integrity of microglia in AD is inextricably linked to the Autophagy-Lysosomal Pathway (ALP), which is required not only for the degradation of phagocytosed Aβ but also for maintaining immunometabolic homeostasis (136, 137). In the AD microenvironment, this system frequently collapses due to the dysregulation of TFEB, the master regulator of the Coordinated Lysosomal Expression and Regulation (CLEAR) network (138). Under chronic neuroinflammation, hyperactive signaling via mTORC1 and GSK3β results in the inhibitory phosphorylation of TFEB, sequestering it in the cytoplasm and preventing the transcription of genes essential for lysosomal acidification and biogenesis (139, 140). Therapeutic strategies employing TFEB agonists, such as celastrol or the curcumin analog C1, aim to bypass this blockade. These agents have shown efficacy in restoring lysosomal pH, upregulating the degradative machinery, and re-establishing the phagocytic clearance of protein aggregates in preclinical models (138, 141).

Parallel to transcriptional modulation, the disaccharide trehalose represents a distinct class of autophagy enhancers that operates via an mTOR-independent mechanism. Trehalose functions as a competitive inhibitor of the SLC2A8 (GLUT8) transporter, inducing a state of cellular pseudo-starvation that activates AMPK and subsequently ULK1 to initiate autophagosome formation (142). This mechanism is particularly critical for resolving the newly identified LDAM phenotype, a dysfunctional state characterized by ACSL1 expression and lipid metabolic failure, often exacerbated in APOE4 carriers (131). By stimulating lipophagy, trehalose facilitates the catabolism of accumulated lipid droplets and dysfunctional mitochondria (143). This restoration of autophagic flux exerts a dual protective effect: it clears intracellular cargo and simultaneously suppresses the activation of the NLRP3 inflammasome by removing endogenous triggers such as mitochondrial ROS and leaking cathepsins (99, 144, 145).

Despite robust preclinical evidence, the clinical translation of autophagy enhancers has historically been hindered by pharmacokinetic barriers, most notably the rapid degradation of oral trehalose by the intestinal enzyme trehalase. To address this, current clinical efforts focus on advanced delivery methods, such as SLS-005 (intravenous trehalose), which is currently being evaluated in a Phase 2/3 trial (NCT05332678↗) (146). This investigational approach utilizes high-dose systemic administration to saturate enzymatic degradation and achieve therapeutic CNS concentrations. If successful, this trial would provide biological proof-of-concept that restoring microglial autophagy can reduce Aβ and tau pathologies while resolving the chronic neuroinflammatory milieu (147).

The CSF1R signaling axis serves as the primary survival switch for microglia, regulating their proliferation and viability through the binding of ligands CSF1 and IL-34 (148). Pharmacological blockade of this receptor, utilizing small-molecule inhibitors such as PLX5622 or the clinically relevant JNJ-40346527, triggers rapid, non-inflammatory apoptosis of the microglial compartment, effectively “scrubbing” the brain of dysregulated or senescent immune populations (149, 150). The therapeutic rationale rests on the concept of “microglial resetting”: upon withdrawal of the inhibitor, the CNS is rapidly repopulated by the clonal expansion of residual surviving cells (151). Crucially, these repopulating microglia often exhibit a renewed, homeostatic transcriptomic signature that resembles a younger phenotype, effectively reversing age-associated inflammatory markers and restoring synaptic plasticity (152, 153). This strategy offers a unique opportunity to eliminate the chronic, maladaptive neuroinflammation characteristic of AD without permanently compromising the brain’s innate immune defense.

However, the application of CSF1R inhibition in AD models reveals a complex, stage-dependent dichotomy, particularly regarding amyloid versus tau pathology. While depletion consistently arrests the trans-synaptic propagation of tau and rescues neurodegeneration in tauopathy models, its impact on amyloidosis is fraught with trade-offs (154, 155). Prophylactic depletion can prevent the initial seeding of plaques, but removing microglia in established amyloid pathology disrupts the protective “plaque compaction” barrier (17). This loss of containment leads to diffuse plaque morphology, exacerbated neuritic dystrophy, and a dangerous redistribution of amyloid to the vasculature, potentially precipitating Cerebral Amyloid Angiopathy (CAA) (46, 47). Consequently, the field is pivoting away from chronic ablation toward “cyclic depletion-repopulation” strategies. This pulsed approach aims to periodically clear senescent and tau-spreading microglia while allowing the repopulated cells to return and maintain essential barrier functions, thereby balancing neuroprotective renewal with plaque containment (149, 156).

The clinical translation of microglial immunomodulators has historically been impeded by the “double barrier” challenge: the highly selective BBB that excludes most macromolecular therapeutics, and the difficulty of selectively targeting the microglial population (10–15% of brain cells) without off-target effects on neurons or astrocytes (157, 158). To overcome these obstacles, advanced nanoparticle systems have been engineered to exploit physiological transport mechanisms, primarily receptor-mediated transcytosis (RMT). While traditional strategies targeted the Transferrin Receptor (TfR), recent innovations utilize ligands such as the T7 peptide, which binds to a distinct pocket on TfR to avoid competition with endogenous transferrin, thereby significantly enhancing BBB penetration (159). Similarly, targeting the LRP1 receptor with high-affinity peptides like Angiopep-2 has enabled liposomal carriers to achieve superior transcytosis rates (160). These physicochemical modifications allow nanocarriers to navigate the neurovascular unit, serving as a prerequisite for delivering therapeutic cargoes to the brain parenchyma.

Upon traversing the BBB, the next critical objective is achieving precise “homing” to activated microglia, often by hijacking their intrinsic phagocytic and scavenger receptors (161). Strategies utilizing ligand-based targeting, such as mannose-coated lipid nanoparticles (MLNPs), have demonstrated remarkable specificity by engaging the CD206 receptor, achieving high co-localization with microglia while sparing neurons (162). Furthermore, biomimetic approaches are being developed to disguise nanoparticles with “eat-me” signals; for instance, coating particles with Gas6-overexpressing membranes allows them to engage TAM receptors (Tyro3, Axl, MerTK), effectively mimicking apoptotic debris to trigger specific engulfment (163). To resolve the “target conflict” between BBB entry and microglial uptake, sequential “sheddable” systems have been designed. These sophisticated constructs, such as the Res@TcMNP/ASO system, feature a pH-sensitive outer shell that facilitates BBB crossing via TfR but sheds within the endosome to reveal a secondary ligand (e.g., MG1) optimized for microglial binding, ensuring multi-stage delivery efficiency (164).

The versatility of these nanoplatforms has facilitated a shift from simple drug delivery to complex cellular reprogramming, targeting the genetic and metabolic drivers of neuroinflammation. Lipid nanoparticles (LNPs) optimized for microglial transfection have successfully delivered siRNAs to silence the master regulator PU.1, dampening inflammatory transcriptional signatures, and delivered antagomirs against miR-17 to restore autophagic flux (162, 165). In the realm of immunometabolism, “smart” nanozymes like the LND@HMPB-T7 system not only inhibit HK2 to switch microglial metabolism from glycolysis back to OXPHOS but also scavenge the resulting reactive ROS, effectively uncoupling phagocytic clearance from cytokine production (159). Additionally, encapsulating inflammasome inhibitors like MCC950 allows for the concentrated blockade of NLRP3 signaling, preventing pyroptosis and enhancing amyloid clearance without systemic immunosuppression (166). Collectively, these nanoparticle-mediated strategies represent a paradigm shift towards precision nanomedicine, offering the potential to arrest the neuroinflammatory cascade by directly remodeling microglial function (Figure 6).

The recent clinical validation of anti-amyloid mAbs, such as lecanemab and donanemab, marks a paradigm shift in AD therapeutics; however, the modest clinical efficacy observed in Phase 3 trials underscores the limitations of targeting proteinopathy in isolation (4, 5). While these agents effectively reduce plaque burden, clinical data suggest that amyloid clearance alone is insufficient to fully arrest the neurodegenerative cascade once it has been initiated (167). A significant challenge remains the “therapeutic gap,” where disease progression persists despite amyloid removal, likely due to the multifaceted nature of AD pathophysiology involving downstream tau propagation and chronic neuroinflammation (167, 168). Furthermore, the administration of passive immunotherapies is frequently complicated by ARIA, a dose-limiting adverse event driven by perivascular inflammation and BBB compromise (169). Consequently, the field is shifting toward combinatorial regimens designed to synergize mechanical plaque clearance with targeted immunomodulation, aiming to enhance efficacy while mitigating vascular toxicity.

To maximize the therapeutic potential of anti-amyloid antibodies, combinatorial strategies focus on optimizing the functional state of microglia, which effectuate plaque phagocytosis. The efficacy of immunotherapy relies on the responsiveness of the host’s endogenous myeloid cells; however, chronic AD pathology is often associated with dysregulated signaling pathways that may impede efficient clearance (124, 168). Strategies targeting the TREM2-CD33 axis propose a dual mechanism: utilizing TREM2 agonists to sustain essential intracellular signaling (PI3K-mTOR) for cell survival and phagocytosis, while concurrently antagonizing the inhibitory CD33 receptor to lower the activation threshold (70, 78). Complementing this signaling modulation, metabolic reprogramming represents another critical avenue. Glucagon-like peptide-1 (GLP-1) receptor agonists have been shown to modulate microglial immunometabolism, potentially shifting cells from a pro-inflammatory glycolytic profile toward OXPHOS (170). This metabolic restoration could provide the bioenergetic support necessary for microglia to sustain the high energy demands of antibody-mediated plaque clearance and subsequent tissue repair.

Beyond enhancing clearance efficiency, combinatorial approaches are essential for decoupling phagocytosis from neurotoxic inflammatory responses. The internalization of amyloid aggregates can trigger the assembly of the NLRP3 inflammasome, leading to the release of pro-inflammatory cytokines (IL-1β) and pyroptosis, which may exacerbate neuronal damage and BBB dysfunction (98). The co-administration of small molecule NLRP3 inhibitors aims to suppress this deleterious cascade without impairing the phagocytic uptake of plaques (171, 172). Additionally, inhibiting the complement cascade (specifically C3) may serve as a protective strategy to prevent aberrant synaptic pruning that can occur during the active phase of plaque removal (173). Finally, engineering advancements such as bispecific antibodies targeting the Transferrin Receptor 1 (TfR1) facilitate RMT (174). By enhancing brain parenchymal delivery, these novel vehicles reduce the requirement for high systemic dosing, thereby minimizing the vascular accumulation associated with ARIA and improving the overall safety profile of combinatorial interventions (Figure 7) (175).