Microglial NF-κB Signaling: Distinguishing Amyloid Clearance, Inflammation, and Tau Propagation

The role of NF-κB in microglia cannot be reduced to a simple balance between beneficial amyloid clearance and harmful inflammation. Microglia perform multiple functions in AD, including migration toward pathological deposits, uptake and degradation of Aβ and tau, cytokine production, synaptic regulation, trophic support, and host defence. These functions are related but distinct. In particular, increased uptake of Aβ or tau does not necessarily indicate complete intracellular degradation, and reduced plaque or aggregate burden does not by itself demonstrate reduced neurotoxicity. Across experimental studies, several microglial surveillance and cargo-uptake functions remain intact when sustained IKK–NF-κB inflammatory signaling is suppressed, whereas persistent pathway activity is more consistently associated with inflammatory neuronal injury and, in tauopathy models, propagation of seeding-competent tau.

Studies targeting PARP1 and IKKβ illustrate this distinction particularly well. In hAPP-J20 mice and complementary microglial models, PARP1 deficiency reduced Aβ-induced NF-κB activation, nitric oxide and TNF-α production, microglial neurotoxicity, and cognitive abnormalities without impairing Aβ phagocytosis; trophic factors including TGF-β and VEGF were also increased (Kauppinen et al., 2011b). Similarly, conditional deletion of Ikbkb in the myeloid lineage of TgCRND8 mice reduced TNF-α and IL-1β, cerebral Aβ accumulation, and inflammatory activation while preserving synaptic proteins and improving memory performance (Liu et al., 2014). Microglial recruitment to plaques remained intact, and IKKβ-deficient macrophages showed enhanced uptake of oligomeric Aβ42. Together, these findings indicate that chronic canonical NF-κB signaling is not required for microglial recognition or internalization of Aβ. They do not, however, establish that all subsequent steps of lysosomal degradation remain unaffected.

The importance of separating phagocytosis from inflammatory state is further demonstrated by Nfkb1 deletion. Rolova et al. found that loss of p50 increased microglial Aβ uptake and modestly reduced amyloid burden, but simultaneously increased inflammatory gene expression and leukocyte infiltration (Rolova et al., 2014). This apparently conflicting phenotype is consistent with the dual regulatory role of p50, which can participate in transcriptionally active p50/RelA complexes but can also form p50/p50 homodimers that restrain inflammatory transcription. Thus, greater Aβ uptake can coexist with a more inflammatory microglial state, showing that neither phagocytosis nor plaque burden alone is sufficient to define whether a microglial response is beneficial.

Several studies further indicate that the inflammatory environment can amplify and prolong microglial NF-κB signaling. Bonaiuto et al. showed that interferon-γ markedly enhanced RelA/p50 activation produced by relatively weak Aβ25–35 stimulation in microglia and human monocytes (Bonaiuto et al., 1997b). Likewise, Aβ25–35 and IL-1β produced a stronger and more sustained NF-κB response when combined than when applied individually (Samuelsson et al., 2005). These findings suggest that prior inflammatory activation can lower the threshold for subsequent Aβ-induced NF-κB signaling and potentially promote a persistent inflammatory state.

A direct functional connection between microglial NF-κB and neuronal injury was subsequently provided by Chen et al. In Aβ-stimulated microglia–neuron systems, constitutive inhibition of microglial NF-κB using a nondegradable IκBα super-repressor markedly reduced Aβ-dependent neuronal toxicity. Aβ also increased acetylation of RelA/p65 at lysine 310, a post-translational modification that enhances the transcriptional competence of RelA. Increasing sirtuin 1 (SIRT1), which deacetylates RelA, or treating the cells with the SIRT1-activating compound resveratrol reduced NF-κB signaling and protected neighboring neurons (Chen et al., 2005b). The convergence of direct pathway blockade and regulation of RelA acetylation strengthens the causal interpretation. In this system, microglial NF-κB did not merely accompany neuronal injury; suppression of the microglial pathway reduced that injury. The experiment therefore identifies an important non-cell-autonomous mechanism in which Aβ activates RelA-dependent transcription in microglia, and the resulting microglial signals subsequently damage neurons.

Not all microglial NF-κB activity is initiated directly by Aβ. Receptor-proximal regulatory mechanisms can determine how readily microglia enter an inflammatory state. Low-density lipoprotein receptor-related protein 1 (LRP1), which participates in ligand uptake as well as intracellular signaling, appears to function as one such restraint. In primary mouse microglia, Lrp1 knockdown or pharmacological blockade with receptor-associated protein increased c-Jun N-terminal kinase (JNK) and NF-κB signaling and exaggerated inflammatory cytokine production, particularly following lipopolysaccharide stimulation. Conversely, NF-κB inhibition reduced cytokine production and partially restored LRP1 expression suppressed by inflammatory stimulation (Yang et al., 2016). These findings suggest that LRP1 normally restrains microglial inflammatory responsiveness and that loss of this restraint can facilitate NF-κB activation. Because LRP1 also participates in ligand-clearance biology, its dysfunction provides a plausible point at which impaired handling of extracellular material and excessive inflammation could converge. Nevertheless, the experiment did not directly test Aβ or tau pathology in an AD model, so it identifies a regulatory mechanism relevant to AD rather than demonstrating that LRP1–NF-κB signaling drives disease progression.

MicroRNA regulation provides another mechanism through which transient inflammatory activation may become more persistent. Zhao et al. identified miR−25802 as a microRNA increased in the plasma of individuals with AD and in the hippocampus of APP/PS1 and 5xFAD mice. Using gain- and loss-of-function experiments in EOC20 microglial cells and AD mouse models, they showed that miR−25802 targets Krüppel-like factor 4 (KLF4), a transcription factor associated in this context with a less inflammatory microglial state. Increasing miR−25802 reduced KLF4 and enhanced NF-κB-associated inflammatory signaling, whereas restoring KLF4 attenuated microglial inflammation and improved pathological and cognitive outcomes (Zhao et al., 2024b). The rescue experiment is important because it orders the pathway more convincingly than parallel expression changes alone: the findings support a miR−25802 → KLF4 suppression → NF-κB activation sequence. Nevertheless, the mechanistic work relied partly on an immortalized microglial line and familial amyloid models, so whether the same regulatory circuit dominates in aged sporadic human AD remains unknown.

A related microRNA study identified a mechanism capable of stabilizing IKKβ itself. Wang et al. found that miR−155–5p suppresses S-phase kinase-associated protein 2 (SKP2), thereby reducing ubiquitination and degradation of IKKβ. As a result, IKKβ protein becomes more stable and remains available for continued signaling. In APP/PS1 mice and Aβ1–42-treated N2a cells, inhibition of miR−155–5p or knockdown of IKKβ reduced Aβ deposition and improved pathological and cognitive outcomes, whereas loss of SKP2 produced the opposite phenotype (Wang et al., 2022a). This study provides a plausible mechanism by which inflammatory kinase activity can become persistent: rather than merely increasing acute IKKβ activation, miR−155–5p reduces the cellular machinery responsible for IKKβ turnover. However, this experiment was not microglia-specific; the cellular work used N2a neuronal cells and the in-vivo manipulations were performed in the hippocampus. It should therefore be regarded as supporting evidence for an AD-relevant miR−155–SKP2–IKKβ mechanism rather than definitive evidence that the same sequence operates specifically within microglia.

More recent work has identified a receptor–adaptor mechanism that does include microglial inflammatory signaling. Liu et al. integrated human AD transcriptomic datasets with Aβ1–42-treated SH-SY5Y cells, LPS-stimulated HMC3 microglia, and APP/PS1 mice to examine signaling lymphocytic activation molecule family member 8 (SLAMF8) and nerve injury-induced protein 2 (NINJ2). SLAMF8 was increased in the disease models, and its overexpression enhanced TLR4/NF-κB signaling, pro-inflammatory cytokine production, oxidative stress, and AD-like abnormalities. NINJ2 interacted functionally with SLAMF8, and NINJ2 loss prevented SLAMF8-induced p65 activation and markedly attenuated the inflammatory and oxidative phenotype (Liu et al., 2025). This pathway-ordering experiment supports a SLAMF8–NINJ2–TLR4/NF-κB axis more strongly than simple pathway colocalization. However, because different experimental stimuli were applied to neuronal and immortalized microglial lines, the relative contribution of this mechanism within specific cell populations of the aged sporadic AD brain remains unresolved.

The strongest evidence for a detrimental microglial NF-κB program comes from studies of tau, where the relevant outcome is not simply aggregate uptake but the release and intercellular propagation of seeding-competent tau. Wang et al. showed that primary mouse microglia exposed for 24 h to full-length tau fibrils at 2 µg/mL or K18/PL tau fibrils at 2.5 µg/mL strongly activated an NF-κB reporter, with the response approaching that produced by 50 ng/mL lipopolysaccharide. To examine what microglia subsequently did with pathological tau, the investigators exposed them to AD-derived tau at 1 µg/mL after two hours of pretreatment with 0.1 µM TPCA−1, an IKKβ inhibitor. IKKβ inhibition reduced the release of tau capable of seeding new aggregates and improved chaperone-mediated autophagy (Wang et al., 2022a). Thus, inhibition did not merely reduce an inflammatory marker; it changed the biological fate of tau processed by microglia.

The in-vivo component of the same study substantially strengthened this conclusion. Tamoxifen-inducible Cx3cr1-CreERT2 was used to either delete or constitutively activate microglial Ikbkb in PS19 tauopathy mice. In the tau-seeding experiments, three-month-old animals received unilateral hippocampal K18/PL tau inoculation—2 µL at 2.5 µg/µL in the loss-of-function experiments and 2 µL at 0.2 µg/µL in the gain-of-function experiments—and were examined one month later. Chronic outcomes were also assessed in 8–10-month-old PS19 mice using histopathology, Morris water-maze testing, bulk RNA sequencing, and single-nucleus RNA sequencing. Microglial IKKβ/NF-κB inactivation reduced tau seeding and anatomical spread, improved autophagic function, shifted disease-associated microglial transcription toward a less pathological state, and improved cognitive performance. Constitutive microglial IKKβ activation produced the opposite result and intensified tau propagation (Wang et al., 2022a). The bidirectional, microglia-specific genetic design provides unusually strong evidence that IKKβ–NF-κB activity is not merely associated with tau pathology but actively facilitates its intercellular propagation in this model.

An apparently paradoxical result from these experiments further illustrates why pathological endpoints must be considered separately. Chronic microglial NF-κB inhibition increased the amount of tau remaining within neuronal inclusions even though extracellular seeding, anatomical propagation, microgliosis, and cognitive impairment were reduced. This finding initially appears to indicate worsening pathology, but it is compatible with reduced processing and extracellular transfer of seeding-competent tau: more pathological tau remained within affected neurons, while less was propagated to new cells. The experiment therefore separates four outcomes that are often treated as interchangeable—intracellular tau burden, extracellular seeding activity, anatomical spread, and neurotoxicity. It does not demonstrate that long-term intracellular tau retention is harmless, but it strongly cautions against judging therapeutic benefit solely by the amount of intracellular aggregate detected.

The therapeutic relevance of the NF-κB inflammatory axis is also supported by a broader intervention study targeting NF-κB together with the NOD-like receptor protein 3 (NLRP3) inflammasome. Wahl et al. administered a brain-targeted nanoligomer cocktail directed against both NF-κB and NLRP3 for four weeks to 19-month-old wild-type mice and young rTg4510 tauopathy mice. Treatment reduced brain inflammatory cytokines and glial activation, improved cognitive performance, and favorably altered tau-related pathology and transcriptomic signatures associated with inflammation and neuronal health (Wahl et al., 2024). These findings support the therapeutic relevance of the broader NF-κB–NLRP3 inflammatory network in aging and tauopathy. However, because NF-κB and NLRP3 were targeted simultaneously and the intervention was not restricted specifically to microglia, the study cannot determine how much of the benefit resulted from NF-κB inhibition alone or from a particular microglial NF-κB program.

Taken together, the microglial evidence provides little support for the idea that sustained NF-κB activation is necessary for beneficial clearance functions. PARP1 deficiency and myeloid IKKβ deletion show that inflammatory and neurotoxic signaling can be reduced while microglial recruitment and Aβ uptake remain preserved, whereas p50 deficiency demonstrates that enhanced Aβ uptake can coexist with greater inflammation. Studies of inflammatory priming, RelA acetylation, LRP1, microRNAs, and receptor–adaptor pathways further explain how inducible NF-κB activity can become amplified or persistently maintained. Most importantly, microglia-specific gain- and loss-of-function experiments demonstrate that sustained IKKβ–NF-κB signaling can alter tau processing and actively promote the release and propagation of seeding-competent tau.

The apparent duality of microglial NF-κB therefore arises largely because NF-κB regulates only part of a much broader microglial functional repertoire. Suppressing pathological NF-κB signaling is not equivalent to suppressing microglial function as a whole. Microglia remain essential for surveillance, host defence, debris recognition, appropriate phagocytosis, lysosomal processing, trophic support, and synaptic homeostasis. A more plausible therapeutic strategy would therefore be to restrain persistent IKK–RelA-dependent inflammatory and tau-propagating programs while preserving physiological microglial functions, including recruitment, appropriate cargo uptake, and effective intracellular degradation.