Beyond isolated signalling cascades, AD neuroinflammation also arises from complex crosstalk between innate immune pathways.

NSAIDs were among the earliest pharmacological interventions evaluated for AD, owing to epidemiological studies that suggested reduced AD incidence among long-term NSAID users. These agents inhibit cyclo-oxygenase (COX) enzymes, particularly COX-2, which is up-regulated in AD brains and contributes to prostaglandin-mediated neuroinflammation 30. However, clinical trials have yielded mixed outcomes. Although NSAIDs such as naproxen and celecoxib demonstrated modest preventive potential in preclinical AD models, their efficacy in symptomatic AD has been limited, possibly because treatment commenced at later disease stages 31. In clinical studies, NSAIDs likewise appear to confer greater benefit in preventing the transition from cognitive impairment to AD than treating established disease, and additional research is needed 31. A Phase I trial of salsalate is ongoing, and preliminary data indicate CNS anti-inflammatory effects. Phase II studies evaluating the leukotriene-receptor antagonist montelukast and the p38 MAP-kinase inhibitor neflamapimod (VX-745) are also in progress 32. Overall, NSAIDs may hold more promise as preventive agents than as treatments, yet definitive disease-modifying effects have not been demonstrated.

In the Alzheimer’s Disease Anti-inflammatory Prevention Trial (ADAPT; n = 2,528; naproxen 220 mg twice daily or celecoxib 200 mg twice daily for 2–3 years), no cognitive benefit was detected 33. A smaller randomised controlled trial (n = 351) that evaluated rofecoxib 25 mg/day or naproxen 220 mg twice daily in patients with mild-to-moderate AD for 12 months likewise failed to attenuate decline on the Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) or the Clinical Dementia Rating–Sum of Boxes (CDR-SB) 34. This discrepancy probably reflects the later disease stage and shorter treatment exposure in the latter study.

Modulating microglial hyperactivation represents a promising therapeutic strategy for neuroinflammatory disorders. Inhibitors of the colony-stimulating factor-1 receptor (CSF1R), such as PLX3397, effectively suppress microglial proliferation and attenuate the release of pro-inflammatory cytokines. Preclinical studies demonstrate that CSF1R blockade mitigates Aβ deposition and preserves synaptic integrity 35. Similarly, minocycline, an inhibitor of microglial activation, reduces lipopolysaccharide (LPS)-induced neuroinflammation and associated cognitive deficits in rodent models 36. Beyond CSF1R signaling, TREM2 has emerged as a pivotal regulator of microglial phenotypes. Augmenting TREM2 activity enhances phagocytosis while suppressing pro-inflammatory signaling, underscoring its potential as a disease-modifying target in neurodegenerative conditions 17,18,20,35,37.

In APP/PS1 mice, PLX3397-mediated CSF1R inhibition markedly reduces microgliosis and Aβ pathology 35; however, clinical translation remains limited. Conversely, minocycline improves cognitive performance in LPS-challenged rats 36, yet human trials have yielded inconsistent results, probably owing to delayed treatment initiation and insufficient central nervous system penetration.

Curcumin, a polyphenolic compound derived from turmeric, inhibits NF-κB and MAPK signalling, reduces astrocytic and microglial activation, and suppresses IL-1β and TNF-α expression. It also interferes with Aβ aggregation and tau hyperphosphorylation 38,39,40. In aged experimental models, chronic curcumin administration significantly decreased lipid peroxidation and lipofuscin accumulation while simultaneously enhancing superoxide dismutase and Na⁺/K⁺-ATPase activities 40.

Resveratrol, a stilbene-type polyphenol abundant in grapes and red wine, exerts potent antioxidant and anti-apoptotic effects. It activates sirtuin-1 (SIRT1), which deacetylates transcription factors such as p53, the FOXO family, and NF-κB, thereby attenuating pro-apoptotic signalling 41. SIRT1 activation also suppresses NF-κB transcriptional activity and reduces pro-inflammatory cytokine expression. Consistent with these mechanisms, resveratrol has shown promise in pre-clinical models and early-phase clinical studies by modulating inflammatory and amyloid biomarkers 42.

Long-chain omega-3 polyunsaturated fatty acids, particularly docosahexaenoic acid (DHA), exert anti-inflammatory actions by competing with arachidonic acid and serving as precursors for specialised pro-resolving mediators (resolvins). These mediators dampen pro-inflammatory cytokine release and glial activation, and supplementation has conferred cognitive benefits in early-stage AD 43.

Nevertheless, clinical translation of these anti-inflammatory agents has been disappointing. Although curcumin and resveratrol display robust antioxidant and anti-inflammatory activities in vitro and in animal models, randomised controlled trials (RCTs) have yielded predominantly negative or inconclusive results, owing in part to limited BBB permeability, poor oral bioavailability, and rapid systemic metabolism 42,44,45. Similarly, RCTs of NSAIDs such as naproxen and celecoxib were terminated early or showed no cognitive benefit, highlighting the difficulty of repurposing peripheral anti-inflammatories for central nervous system indications 33,34. Contributing factors include sub-optimal dosing, brief treatment duration, and heterogeneous participant populations, which collectively hinder detection of modest therapeutic effects. These findings underscore the need for improved delivery platforms (e.g., nanoparticles, intranasal formulations), rigorous trial design, and combination strategies to overcome the limitations of single-agent therapy.

In a 52-week, phase-II resveratrol trial (n = 119), cerebrospinal fluid Aβ levels and regional brain volumes were altered without parallel cognitive improvement 41,42. A 24-week curcumin RCT (n ≈ 34) likewise failed to improve ADAS-Cog scores 40. By contrast, omega-3 supplementation in individuals with mild cognitive impairment or early AD (n = 46) produced modest cognitive gains 43, reinforcing the importance of disease stage and treatment duration in determining therapeutic response.

However, emerging evidence suggests that therapeutic benefit in AD may be greatest when interventions target key nodes of pathway crosstalk, where multiple signaling axes converge to perpetuate glial dysfunction46. One such node is the TREM2–DAP12–SYK pathway, which regulates lipid sensing, phagocytosis, and microglial state transitions47. TREM2 agonists and monoclonal antibodies are under investigation as strategies to enhance microglial clearance of amyloid deposits and cellular debris, thereby restoring homeostatic surveillance48. By promoting lipid metabolism, cholesterol efflux, and debris processing, TREM2 activation not only attenuates Aβ accumulation but also indirectly limits NLRP3-inflammasome priming and downstream IL-1β secretion49. Consequently, microglia shift from a chronically activated, M1-like pro-inflammatory phenotype to a more homeostatic or reparative state49. Early-phase biologics such as AL002, a humanized agonistic anti-TREM2 antibody, have already entered clinical evaluation and have demonstrated favorable target engagement and safety profiles in first-in-human trials49. These findings underscore the therapeutic promise of TREM2-directed interventions for modulating amyloid pathology and recalibrating microglial immune tone, thereby addressing both innate immune dysregulation and neuroinflammatory amplification characteristic of AD48,49. Preclinical activation of TREM2 with antibodies such as AL002 reduces Aβ plaque burden and restores microglial homeostasis48. By contrast, early-phase human studies have thus far focused primarily on safety and target engagement rather than clinical efficacy49, discrepancies that are largely attributable to differences in trial phase and selected endpoints (biomarker versus clinical outcomes).

The NLRP3 inflammasome, a multiprotein complex that is activated within microglia by Aβ, mitochondrial stress signals and defective lipid sensing, serves as a pivotal trigger for caspase-1 activation and the subsequent maturation of IL-1β and IL-18, thereby fueling chronic neuroinflammation in AD. Small-molecule inhibitors, such as MCC950, selectively impede NLRP3 assembly, thereby preventing caspase-1 activation and downstream cytokine processing. Preclinical studies have demonstrated that MCC950 mitigates microgliosis, decreases Aβ deposition and enhances cognitive performance in transgenic AD models 50,51. Importantly, NLRP3 inhibition represents a convergent node that intersects with TREM2 dysfunction, aberrant TLR–NF-κB signaling, and mitochondrial stress-derived danger signals, including ROS and mitochondrial DNA (mtDNA) release 52,53. By interrupting this feed-forward loop, NLRP3 blockade not only dampens microglial and astrocytic overactivation but also ameliorates synaptic dysfunction and cognitive decline. MCC950 and next-generation inflammasome inhibitors are progressing through translational pipelines, highlighting their potential to re-establish a more homeostatic neuroimmune milieu in AD 50,51,52,53. Consistent with these findings, MCC950 robustly suppressed IL-1β signaling, diminished Aβ burden and improved cognitive outcomes in multiple transgenic mouse models of AD 50,51. Nevertheless, clinical development remains in its infancy, and the apparent translational gap between animal and human data largely reflects differences in trial maturity and a current emphasis on safety assessment rather than an absence of therapeutic efficacy 52,53.

The JAK/STAT signaling cascade is a central mediator of glial proliferation and cytokine-driven neurotoxicity in neurodegenerative contexts. In preclinical models, pharmacological inhibition of JAK2/STAT3 has been shown to attenuate neuroinflammation, reduce gliosis, and improve learning and memory outcomes 54. Aberrant IFN-I signaling perpetuates glial activation through persistent STAT1/3 phosphorylation, amplifying cytokine release and neurotoxic feedback loops. Small-molecule JAK2/STAT3 inhibitors, including ruxolitinib and baricitinib, have demonstrated therapeutic promise in experimental neuroinflammatory models. For example, ruxolitinib reduces microglial proliferation and suppresses TNF-α, IL-1β, and IL-6 expression after spinal cord injury 55 and attenuates inflammatory gene expression in post-infectious inflammatory syndromes 56. Similarly, baricitinib reverses neurocognitive deficits and dampens neuroinflammatory cytokine release in models of HIV-associated neurocognitive disorder 57. By suppressing cytokine transcription and disrupting the synergy between IFN-I–driven STAT1/3 signaling and NLRP3 inflammasome activation, JAK/STAT blockade offers a compelling therapeutic avenue either as a monotherapy or in combination therapy to restore immune balance and limit glial-mediated neurodegeneration.

JAK inhibitors, such as ruxolitinib, have demonstrated potent glial anti-inflammatory and neuroprotective effects in preclinical models of AD, including suppression of microglial activation and downstream cytokine signaling 55. However, human investigations to date have been limited primarily to early-phase or biomarker-focused studies, with no conclusive evidence of cognitive benefit. This discrepancy largely stems from differences in study design and endpoints; preclinical studies typically assess neuroinflammatory and molecular outcomes, whereas clinical trials emphasize cognitive and functional measures and are constrained by limited treatment duration and subtherapeutic central exposure during early human testing.

The APOE4 allele is the strongest, most consistently validated genetic risk factor for late-onset AD, conferring a three- to twelve-fold increase in risk relative to APOE3 or APOE2 carriers 58. Beyond its canonical role in cholesterol and lipid transport, APOE4 profoundly influences microglial transcriptional programs, driving them toward pro-inflammatory, disease-associated phenotypes 59. This maladaptive re-programming disrupts lipid metabolism, impairs Aβ clearance, and acts synergistically with TREM2 dysfunction to amplify neuroinflammatory cascades 59. To counteract these effects, several APOE-directed therapeutic strategies are currently under development. APOE mimetic peptides, such as CS-6253, restore lipid efflux and normalize cholesterol homeostasis, thereby indirectly attenuating microglial activation and synaptic toxicity 58. Gene-therapy approaches aim to convert APOE4 to a more protective, APOE2-like profile, thereby recalibrating the balance between lipid metabolism and immune regulation 60. In parallel, antisense oligonucleotides (ASOs) are being investigated to selectively lower APOE4 expression, thereby reducing its pathogenic gain-of-function effects and mitigating downstream inflammatory signalling 61. Collectively, these interventions target the metabolic–immune interface, positioning APOE modulation as a strategy not only to enhance amyloid clearance but also to indirectly regulate TREM2-dependent phagocytosis and NF-κB-driven cytokine release, thus dampening the chronic neuroinflammation that characterises APOE4-associated AD 58,59,60,61. APOE mimetics and antisense oligonucleotides have demonstrated promising effects in preclinical models, including modulation of lipid metabolism, reduction of amyloid deposition, and attenuation of microglial activation 58. These agents are designed to restore homeostatic APOE function, enhance lipid transport, and mitigate neuroinflammation—key processes implicated in AD pathogenesis. However, the ongoing phase 1 clinical trials are focused primarily on evaluating safety, tolerability, and target engagement rather than cognitive outcomes. Consequently, the current evidence of efficacy is confined to biomarker improvements, reflecting the early stage of clinical translation and the need for longer-term trials that assess cognitive and functional endpoints.

Toll-like receptors (TLRs), particularly TLR2 and TLR4, act as innate immune sensors that recognize aggregated Aβ, oxidized lipids, and mitochondrial debris, thereby driving persistent activation of the NF-κB pathway in AD 62. Chronic engagement of these receptors results in transcriptional up-regulation of pro-inflammatory cytokines such as IL-1β, TNF-α, and IL-6, which sustain microglial activation, astrocytic reactivity, and neuronal injury 63,64.

Pharmacological modulation of TLR signalling represents a promising upstream strategy for attenuating innate immune overactivation. Resatorvid (TAK-242), a selective TLR4 inhibitor, disrupts receptor–adaptor interactions, limits NF-κB nuclear translocation, and diminishes downstream cytokine production, yielding neuroprotective effects in preclinical AD models 65.

Importantly, TLR modulation may be particularly relevant in APOE4 carriers, who exhibit impaired lipid metabolism and reduced amyloid clearance 66. In this genetic context, excess lipid debris and amyloid aggregates serve as persistent TLR ligands, further amplifying microglial activation and neurotoxic feedback loops 66. Natural compounds with TLR-modulating properties, such as curcumin and resveratrol, may also provide adjunctive benefits, although their poor bioavailability hampers clinical translation 67. Collectively, these findings identify TLR–NF-κB inhibition as a strategic intervention point that bridges amyloid pathology, genetic risk, and innate immune dysregulation, positioning TLR modulators as candidates for genotype-informed therapies in AD 62,63,64,65,66,67.

TLR4 inhibition with TAK-242 has been shown to markedly suppress Aβ-induced cytokine release and neuroinflammatory signalling in both cellular and rodent models 65, highlighting TLR4 as a critical upstream mediator linking innate immune activation to microglial-driven neurotoxicity in AD. Nevertheless, despite compelling preclinical evidence, no pivotal clinical trials have yet been conducted in AD populations. The preclinical–clinical gap likely reflects the complex and pleiotropic nature of TLR4 signalling, which participates in both protective and pathological immune responses, as well as the predominance of late-symptomatic enrolment in human studies, when neuroinflammatory cascades are already entrenched and less amenable to modulation.

In summary, neuroinflammation in AD arises from complex crosstalk among metabolic, immune, and glial pathways, and therapeutic efforts are shifting toward interventions at these convergent hubs. By targeting mechanisms such as TREM2 signalling, NLRP3 inflammasome activation, JAK/STAT pathways, and APOE-driven immune dysregulation, emerging strategies may restore immune balance and attenuate neuronal injury. Future progress will depend on improved drug delivery, rigorous clinical trial design, and combination approaches that transcend single-pathway paradigms to achieve meaningful disease modification.

Early immunotherapeutic efforts in AD primarily targeted the removal of Aβ plaques. However, recent advances have shifted towards directly modulating neuroinflammatory pathways. Persistent activation of microglia and astrocytes plays a critical role in synaptic dysfunction and neuronal loss. Emerging immunotherapeutic approaches aim to reprogram microglia from a pro-inflammatory (M1-like) phenotype to an anti-inflammatory (M2-like) phenotype 9. Additionally, targeting immune checkpoints, such as the PD-1/PD-L1 axis, may help re-establish immune homeostasis within the central nervous system 68. Another promising avenue involves modulation of TREM2 signaling, a key regulator of microglial function that is involved in phagocytosis, cytokine regulation, and lipid metabolism 69. Collectively, these strategies seek to reduce neuroinflammation, enhance glial cell function, facilitate clearance of cellular debris, and support neuronal survival.

Advancements in genetic engineering have enabled the development of precise molecular tools to manipulate inflammatory signaling in the brain. CRISPR/Cas9 genome editing is being explored to silence genes that drive chronic inflammation, such as genes encoding components of the NLRP3 inflammasome or pro-inflammatory cytokines, including IL-1β and TNF-α 70. Antisense oligonucleotides (ASOs) and RNA interference (RNAi) technologies have been employed to reduce the expression of specific targets implicated in neurodegeneration, including tau, TREM2, and inflammatory cytokines 71,72. Although these strategies confer high specificity and are reversible, efficient delivery—particularly across the BBB—remains a major obstacle. Promising preclinical results suggest that gene-editing and RNA-based therapeutics may provide a powerful means of fine-tuning immune responses and mitigating neuroinflammatory damage in Alzheimer’s disease (AD).

Stem cell–based therapies provide a multipronged strategy for restoring immune homeostasis, augmenting neuroprotection, and promoting tissue repair. Mesenchymal stem cells (MSCs), isolated from bone marrow, adipose tissue, and other adult sources, exert potent immunomodulatory actions by secreting anti-inflammatory cytokines (e.g., IL-10, TGF-β), driving microglial polarization toward an anti-inflammatory phenotype, suppressing astrocytic activation, and attenuating oxidative stress 76,77. Neural stem cells (NSCs) not only possess the capacity to replace lost neurons but also release trophic and immunoregulatory mediators that foster neurogenesis and mitigate inflammatory processes 78. In preclinical models, stem-cell therapy has been shown to decrease glial activation, improve cognitive performance, and remodel the neuroinflammatory milieu 79. Although clinical experience remains limited, early-phase trials indicate that stem cell–based interventions may act as disease-modifying treatments capable of addressing the intricate interplay between immune dysregulation and neurodegeneration in Alzheimer’s disease (AD).

Next-generation modalities, including CRISPR/Cas9 gene editing, immune-checkpoint inhibition (anti-PD-1/PD-L1), and additional stem-cell platforms, have garnered substantial interest; however, all remain at low technology-readiness levels (TRLs) and thus warrant cautious interpretation. CRISPR/Cas9 applications in AD are currently confined to preclinical proof-of-concept studies (TRL 2–3) that target APOE4 alleles or amyloidogenic drivers 80. Immune-checkpoint modulation with anti-PD-1 antibodies has demonstrated enhanced immune surveillance and amyloid clearance in murine models, yet clinical development is still at TRL 3–4, with no large-scale efficacy data in patients 81. Stem-cell therapies, particularly MSC infusions, have advanced further, with several phase-I/II trials (TRL 4–5) confirming feasibility and safety (e.g., NCT02600130, NCT02833792); definitive efficacy signals, however, are not yet available 82. Collectively, these interventions remain experimental, and despite their promise, their readiness for routine clinical application in AD is currently limited.

To facilitate critical appraisal, we systematically assessed each therapeutic category with respect to both mechanistic plausibility and clinical maturity. Table 1 collates representative agents and platforms, grades them by development phase (preclinical, phases 1–3), and lists key outcome measures (e.g., biomarker modulation, cognitive endpoints, safety/tolerability). This structured synopsis provides clearer insight into translational readiness and differentiates preliminary observations from approaches undergoing clinical validation.

In summary, these emerging strategies underscore a paradigm shift in AD research, moving beyond classical amyloid- and tau-centric targets toward a comprehensive focus on neuroinflammation. Although challenges regarding delivery, safety, and sustained efficacy persist, these innovative approaches offer realistic prospects for more effective, individualized interventions that directly address the inflammatory dimension of AD pathogenesis.

Across therapeutic classes, the overall strength of evidence remains moderate to low, reflecting the predominance of early-phase or preclinical studies. NSAIDs and omega-3 fatty acids are the most clinically advanced candidates; however, large prevention and symptomatic trials have produced inconsistent cognitive outcomes, providing at most moderate evidence. Natural compounds such as curcumin and resveratrol yield biochemical and biomarker improvements but lack adequately powered RCTs, resulting in low-quality evidence because of small sample sizes, surrogate end points, and formulation heterogeneity. Microglia-targeted agents (CSF1R, TREM2, NLRP3 and TLR modulators) and JAK/STAT inhibitors demonstrate strong mechanistic plausibility in animal models; however, they are supported only by preclinical or biomarker-level findings, classifying them as low-level evidence for clinical efficacy. Gene- and RNA-based strategies, as well as stem-cell therapies, remain exploratory; they are constrained by safety uncertainties, delivery challenges and insufficient long-term follow-up.

Substantial heterogeneity in study design, disease stage and outcome measures impedes cross-study comparison. Most human studies enrol small, heterogeneous cohorts, frequently at late symptomatic stages when neuroinflammatory cascades are entrenched and less amenable to reversal. Publication bias towards positive or mechanistically interesting results further distorts perceived efficacy, whereas negative or null findings remain under-reported. Translational gaps between animal and human studies persist, driven by inter-species differences in immune architecture, dosing regimens and treatment duration. Collectively, these factors limit generalisability and underscore the need for standardised outcome frameworks, longitudinal biomarker integration and adequately powered multicentre trials. Future emphasis on early-stage, precision-stratified interventions and transparent data reporting will be essential for strengthening the evidentiary base for neuroinflammation-targeted therapies in Alzheimer disease.