Hospital-treated infection associated with Alzheimer’s disease pathology: underlying mechanisms Authors: Tatiana Barichello, Giselli Scaini, Mohd Tayyab, Henrique Ritter Dal-Pizzol, Fabricia Petronilho, Felipe Dal-Pizzol Published: 2026 ABSTRACT ImportanceGrowing epidemiological and mechanistic evidence indicates that infections substantially increase the risk of Alzheimer’s disease (AD) and related dementias. Systemic immune activation and pathogen persistence may act as upstream triggers accelerating neurodegenerative cascades.ObjectiveTo synthesize recent evidence on infection-driven mechanisms contributing to AD pathology, emphasizing how systemic and central immune activation influence amyloid-beta (Aβ) aggregation, tau pathology, blood-brain barrier (BBB) dysfunction, and neuroinflammation.Design, setting, and participantsThis expert review integrates data from large epidemiological cohorts, neuropathological analyses, and mechanistic studies in both human and animal models to address innate and adaptive immune mechanisms linking infection to AD pathogenesis.ExposuresSystemic and central nervous system (CNS) infections, including sepsis, pneumonia, viral, and chronic bacterial infections, as well as hospital-treated infections, that trigger inflammatory and immune signaling cascades impacting the brain’s structural and molecular integrity.ResultsInfections activate the innate immune system through toll-like receptors and inflammasomes (NOD-like receptor family pyrin domain-containing 3 (NLRP3), linear ubiquitin assembly complex (LUBAC)), driving chronic neuroinflammation, pyroptosis, and the release of apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) speck that cross-seeds Aβ and tau aggregation. Additionally, cytokine-induced upregulation of interferon-induced transmembrane protein 3 (IFITM3) enhances γ-secretase activity, thereby increasing Aβ production. Therefore, peripheral inflammation compromises BBB integrity, while extracellular vesicles propagate inflammatory cargoes across the neurovascular unit. Consequently, the expansion of cytotoxic CD8⁺ T cells and the alteration of Th1/Th17 profiles are linked to adaptive immunity and neurodegeneration. Moreover, epidemiologic data show reduced AD risk following herpes zoster, Tdap/Td (tetanus, diphtheria, and acellular pertussis (Tdap) and tetanus and diphtheria (Td) vaccines), and pneumococcal vaccination, supporting immune modulation as a preventive strategy.Conclusions and relevanceInfection-driven immune activation represents a key modifiable pathway in AD pathogenesis, highlighting novel diagnostic and therapeutic targets focused on the inflammasome. Introduction The potential causal association between infections and the development of dementia, particularly Alzheimer’s disease (AD), has been the subject of substantial scientific investigation and debate [1–4]. Growing epidemiological evidence supports a significant association between infections and increased risk of AD and related dementias (ADRD), suggesting a potential contributory role in their pathogenesis and progression [1–3, 5–11]. In a Danish nationwide cohort study of 1,493,896 individuals followed over 40 years (51% were women), hospital-diagnosed infections were associated with a 1.49-fold increased risk of developing dementia. Notably, infections of the urinary tract (Incidence rate ratio (IRR): 2.31; 1.30 after 5 years), nervous system (IRR: 2.41 within 5 years; 1.21 after 5 years), sepsis (IRR: 1.80; 1.29 after 5 years), and respiratory tract (IRR: 1.71; 1.22 after 5 years) were all associated with an elevated risk of developing dementia. The associations were strongest within the first five years following infection, indicating a temporal relationship between infections and increased risk of dementia, with a similar effect for men and women [2]. In an extensive multicohort study of more than 700,000 individuals from Finnish and UK cohorts, hospital-treated infections were significantly associated with an increased risk of incident dementia, with a hazard ratio (HR) of 1.48 (95% CI, 1.37–1.60) in the Finnish cohort and 2.60 (95% CI, 2.38–2.83) in the UK cohort. Hospital-treated infections were associated with a significantly increased risk of AD, showing an HR of 1.80 (95% CI: 1.53–2.13) compared to uninfected individuals. This association persisted and slightly attenuated the relationship after excluding participants with HIV and adjustment for lifestyle and metabolic factors (alcohol use, smoking, BMI, hypertension, diabetes, and APOE genotype), adjusted hazard ratio (aHR 1.70); 95% CI: 1.42–2.04; however, no significant sex interaction was reported [3]. In a nationwide Swedish cohort study conducted from 1970 to 2016, 12,275,551 individuals of both sexes born in Sweden to Swedish-born parents were followed using national health registers. A history of hospital-treated infection was associated with a 16% increased odds of developing AD (OR, 1.16; 95% CI, 1.15–1.18) and a 4% increased odds of Parkinson’s disease (PD) (OR, 1.04; 95% CI, 1.02–1.06), while no significant association was observed for amyotrophic lateral sclerosis (ALS). When stratified by age at diagnosis, associations were strongest among individuals diagnosed before age 60, with an OR of 1.93 (95% CI, 1.89–1.98) for early-onset AD and 1.29 (95% CI, 1.22–1.36) for early-onset PD. In contrast, no significant associations were found for cases diagnosed at age 60 years or older [8]. Another study analyzed a nationwide cohort of 1,742,406 individuals born in New Zealand between 1929 and 1968, followed prospectively from 1989 to 2019 using population-based registry data. During the 30-year observation period, 465,330 individuals (26.7%) were diagnosed with a hospital-treated infection, and 37,212 (2.1%) developed dementia. Hospital-treated infections were associated with a significantly increased risk of subsequent dementia across all age groups and both sexes. The subdistribution hazard ratios (sHRs) were 2.93 (95% CI, 2.68–3.20) for all-cause dementia, 2.28 (95% CI, 2.11–2.48) for AD, and 4.39 (95% CI, 3.83–5.04) for other dementias, including vascular dementia. Notably, sex-stratified analyses revealed that men in younger cohorts had a higher relative risk of dementia following infection (HR, 4.12). In contrast, women in older cohorts exhibited greater relative risk than men (HR, 4.54) [12]. Additionally, a nationwide population-based case-control study using data from the Austrian National Health Insurance Association, including 58,208 individuals, reported increased dementia risk across several common hospital-treated infections, with odds ratios ranging from 1.16 to 1.26 (sepsis: OR 1.17; influenza: OR 1.26; urinary tract infections: OR 1.23; intestinal infections: OR 1.16) [13]. Moreover, a community-based study in Uganda, including 434 individuals, reported that prior central nervous system infections were strongly associated with increased dementia prevalence among older adults (OR: 4.50; 95% CI: 2.76–7.23) [14]. A Taiwan population-based longitudinal cohort included 20,466 patients, from 2001 to 2011, who were hospitalized with septicemia with no previous diagnosis of dementia. Survivors of septicemia demonstrated a significantly elevated risk of developing dementia compared to matched controls. The aHR for all-cause dementia was 1.94 (95% CI, 1.72–2.19) for females and 2.26 (95% CI, 2.00–2.56) for males, whereas the risk for AD was an aHR of 1.15 (95% CI, 0.83–1.60) [15]. Similarly, a population-based study using German health claims data, including 161,567 individuals, reported a markedly increased risk of dementia following sepsis, with odds ratios of 3.14 for patients with sepsis treated outside the intensive care unit and 2.22 for those requiring intensive care treatment [16]. In conclusion, septicemia was associated with an increased risk of dementia, particularly among younger patients and those with more severe infections [15]. Further information on the epidemiological studies evaluating hospital-treated infections and their association with ADRD risk is summarized in Fig. 1A and B and Supplementary Table 1. In AD [17–19], extracellular accumulation of amyloid-beta (Aβ) and intracellular aggregates of the hyperphosphorylated microtubule-binding protein tau (P-Tau) start decades before the onset of cognitive decline [19], suggesting that immune challenges, such as systemic infections, may act as upstream events that influence this pathological cascade [17–19]. In the following sections, this review examines how infection-induced immune activation can lead to long-lasting alterations in brain homeostasis. We will discuss the roles of peripheral and central immune cells, including microglia and astrocytes, in mediating neuroinflammatory cascades. The discussion then turns to how infections compromise the integrity of the blood-brain barrier (BBB), allowing peripheral immune mediators and pathogens to access the CNS. Finally, the review considers how these immune disruptions intersect with the classical pathological features of AD, providing a hypothetical mechanistic framework for understanding how infections may initiate, accelerate, or exacerbate neurodegeneration.Fig. 1Global Epidemiological Evidence Linking Hospital-Diagnosed Infections to Dementia.A Represents the associations of hospital-diagnosed infections and the risk of dementia in the world. B Infection-associated risk estimates for dementia and AD across epidemiological studies. Forest-style visualization summarizing odds ratios (ORs), hazard ratios (HRs), or relative risks (RRs) reported in epidemiological studies evaluating the association between infections and subsequent dementia or AD. Each point represents the effect estimate reported in an individual study, and horizontal lines indicate the corresponding 95% confidence intervals. Colors indicate the type of infection, while symbol size reflects study size on a logarithmic scale. Refer to the Supplementary Table 1. Global Epidemiological Evidence Linking Hospital-Diagnosed Infections to Dementia. A Represents the associations of hospital-diagnosed infections and the risk of dementia in the world. B Infection-associated risk estimates for dementia and AD across epidemiological studies. Forest-style visualization summarizing odds ratios (ORs), hazard ratios (HRs), or relative risks (RRs) reported in epidemiological studies evaluating the association between infections and subsequent dementia or AD. Each point represents the effect estimate reported in an individual study, and horizontal lines indicate the corresponding 95% confidence intervals. Colors indicate the type of infection, while symbol size reflects study size on a logarithmic scale. Refer to the Supplementary Table 1. Do infections have a role in AD pathology? Several lines of evidence support the hypothesis that infectious agents may contribute to AD pathogenesis, including direct detection of pathogens or microbial products in the brain, epidemiological associations between infection and dementia risk, and experimental studies showing that infectious or inflammatory stimuli can amplify amyloid and tau pathology. In the 1990s, Itzhaki and her colleagues demonstrated the presence of Herpes simplex virus type (HSV-1) viral DNA in the brains of older adults and patients with AD [20]. They proposed that the reactivation of latent HSV-1 within the brain, triggered by factors such as stress, peripheral infection, or recurrent injury [21], may lead to increased production of amyloid and AD-like tau pathology as part of a cellular defense response to protect against brain infection [22, 23]. Recently, a study showed that persistent herpes infection is associated with a relative risk (RR) of 2.1 (95% CI, 1.40-3.14) for developing dementia. Moreover, the estimated lifetime cumulative risk of developing dementia following hospitalization for any viral infection was 1.48 (95% CI, 1.15-1.91) [24]. In addition to HSV-1, other members of the Herpesviridae family are also linked to AD. A detailed study of human brain tissue after death found that human herpesvirus (HHV)-6A and HHV-7 are more common in the brains of people with AD than in control [25]. Apart from herpes virus infection, emerging evidence suggests that SARS-CoV-2 infection may also influence neuroinflammatory pathways and potentially contribute to long-term neurodegenerative risk [24]. Retrospective evidence shows that SARS-CoV-2 infection leads to more neurological and psychiatric sequelae than other respiratory infections, with prolonged cognitive risk in adults (up to two years) but only transient effects in children (~75 days) [24]. Epidemiological evidence strongly supports a bidirectional relationship between COVID-19 and AD. Pre-existing dementia significantly increases susceptibility to severe COVID-19, with an odds ratio (OR ≈ 3.07; 95% CI 1.71-5.50) for severe infection [26]. Conversely, SARS-CoV-2 infection increases the risk of subsequent cognitive decline, with incident dementia showing a hazard ratio (HR ≈ 1.39) within 6 months post-infection [26]. SARS-CoV-2 infects endothelial cells, leading to coagulopathy, microvascular damage, and BBB disruption during the acute phase of infection [27]. These preclinical observations align with human data demonstrating BBB disruption in long COVID with cognitive impairment, as shown by dynamic contrast-enhanced magnetic resonance imaging (MRI) and molecular evidence of sustained systemic inflammatory and coagulation signatures that promote endothelial activation [28]. SARS-CoV-2 infection has emerged as a critical modifier of AD risk and progression through converging molecular, vascular, and neuroimmune mechanisms.. Studies suggested that a central pathological intersection between COVID-19 and AD is neuroinflammation. SARS-CoV-2 infection induces a systemic cytokine storm (IL-6, IL-1β, TNF-α) [29], which exacerbates microglial activation, oxidative stress, and inflammasome signaling, all of which are core drivers of AD progression [30]. Moreover, SARS-CoV-2 also directly influences classical AD hallmarks. Experimental models and human brain analyses show increased tau phosphorylation, tau oligomerization, and α-synuclein expression, indicating acceleration of neurofibrillary pathology. Furthermore, the viral spike protein (S1) may promote tau propagation via extracellular vesicles, thereby enhancing trans-synaptic spread of pathology [24]. Importantly, viral presence was associated with disruption of host gene regulatory networks involved in amyloid precursor protein (APP) processing, immune signaling, and synaptic function [25, 31]. Network modeling suggested that viral activity may influence molecular pathways central to AD pathogenesis, supporting the hypothesis that latent viral reactivation may modulate host transcriptional programs relevant to neurodegeneration rather than acting solely through direct cytopathic effects [4, 25]. Other pathogens that persist in the brain, including Chlamydia pneumoniae, were also identified in postmortem brains of patients with AD within pericytes, microglia, and astroglia, but not in those of controls [32]. A recent study showed that Pseudomonas aeruginosa, a common hospital-acquired pathogen in ICU settings, acts as a pathogen-associated molecular pattern (PAMP)-driven infectious stimulus that triggers systemic inflammation via lung infection and induces BBB dysfunction, astrocyte activation, and increased tau phosphorylation in the brain, demonstrating that Tau is required for infection-induced neurovascular injury and linking pneumonia to dementia-related pathology [33]. Regarding the oral microbiota, Porphyromonas gingivalis secretes gingipains, proteolytic virulence factors that can induce chronic periodontitis. The ultrastructural localization of P. gingivalis gingipains has been identified in the substantia nigra of individuals with PD [34] and in the brains and cerebrospinal fluid (CSF) of individuals with AD [35]. Taken together, these observations suggest that both neurotropic and non-neurotropic pathogens may contribute to AD-related pathology through distinct but convergent mechanisms, including latent reactivation, peripheral inflammation, vascular dysfunction, BBB disruption, and sustained innate immune activation [36]. A possible explanation is that damage-associated molecular patterns (DAMPs) and PAMPs may reach the brain, triggering the host immune response, and that the amyloid serves a physiological role within the innate immune system, functioning as an antimicrobial peptide (AMP) [33, 37]. Supporting the hypothesis that amyloid acts as an extracellular trap for the microbe or PAMPs, a large-scale U.S. study analyzing health records from over 100 million individuals found a consistent association between varicella-zoster virus (VZV) reactivation and an increased risk of dementia, even after controlling for nearly 400 potential confounders. Recurrent herpes zoster episodes were linked to higher dementia risk compared to single occurrences. In contrast, vaccination against herpes zoster has been associated with a lower risk of dementia. In a large population-based cohort study including more than 280,000 adults aged ≥65 years with up to 7 years of follow-up, individuals who received the live-attenuated herpes zoster vaccine showed a reduced incidence of dementia compared with those vaccinated with the 23-valent pneumococcal polysaccharide vaccine used as a control. Notably, the protective effect of the live-attenuated zoster vaccine diminished over time, paralleling the decline in its efficacy against herpes zoster reactivation [38]. Additionally, a significant study showed strong causal evidence that herpes zoster vaccination reduces the risk of developing dementia. The analysis included 282,541 adults born between 1925 and 1942, who were approximately 79-80 years old at the start of the vaccination program in 2013, and followed participants for seven years. By exploiting a birth-date-based eligibility threshold in Wales, the study minimized confounding and demonstrated a 20% relative reduction in dementia diagnoses over 7 years, particularly among women. These findings suggest that herpes zoster vaccination may confer neuroprotective benefits beyond its antiviral effects, supporting further investigation into the mechanisms linking viral reactivation, immune modulation, and neurodegeneration [39]. In a retrospective cohort analysis of adults ≥65 years drawn from Optum’s Clinformatics® Data Mart, using propensity-score-matched cohorts, the study analyzed up to 519,546 individuals, depending on the vaccine evaluated. Participants were free of dementia at baseline and were followed for up to 8 years. The investigators examined tetanus/diphtheria ± pertussis (Tdap/Td), herpes zoster (HZ), and pneumococcal immunizations. Across vaccine types, prior immunization was consistently associated with a reduced risk of AD. Compared with unvaccinated individuals, those vaccinated had relative risks of 0.70 (95% CI, 0.68–0.72) for Tdap/Td, 0.75 (95% CI, 0.73–0.76) for HZ, and 0.73 (95% CI, 0.71–0.74) for pneumococcal vaccine, corresponding to absolute risk reductions of 2–3 percentage points [40]. These findings support accumulating epidemiologic evidence that routine adult vaccinations are associated with neuroprotective benefits and lower the risk of developing AD, potentially through mechanisms involving systemic immune modulation or reduced infection-related inflammation. However, most available evidence derives from observational or quasi-experimental studies, and randomized trials will be necessary to determine whether vaccination confers direct neuroprotective effects. The innate immune activation and AD pathology The innate immune system detects invading pathogens by recognizing PAMPs, conserved microbial structures found in bacteria, viruses, and fungi. Examples include lipopolysaccharide (LPS) from Gram-negative bacteria, lipoteichoic acids from Gram-positive bacteria, bacterial flagellin, pili, microbial amyloids, and fimbrial proteins involved in microbial adhesion, as well as bacterial DNA rich in unmethylated CpG motifs and viral RNA. Microbial PAMPs activate toll-like receptor (TLR) signaling and inflammasome pathways, thereby linking systemic infections with neuroinflammatory processes observed in AD [41, 42]. In addition to microbial signals, the innate immune system responds to endogenous DAMPs and alarmins released by stressed, injured, or dying cells, which act as danger signals to initiate inflammation, promote tissue repair, and signal to the adaptive immune system [41]. DAMPs comprise a diverse group of endogenous molecules released during AD, including Aβ, P-Tau, high-mobility group box 1 (HMGB1), heat shock proteins (HSPs), and adenosine triphosphate (ATP). Additional DAMPs implicated in disease progression include mitochondrial DNA (mtDNA), S100 proteins, and nuclear histones [19, 43, 44]. PAMPs and DAMPs are recognized by pattern recognition receptors (PRRs) located on the cell surface or within endosomal compartments. Upon ligand binding, TLRs, including TLR1/2, TLR2, TLR4, TLR5, or TLR6/2, initiate intracellular signaling cascades, most notably the activation of the nuclear factor kappa B (NF-κB) pathway. For example, bacterial LPS activates TLR4, flagellin activates TLR5, and microbial lipoproteins or pili-associated components can activate TLR2 heterodimers. These pathways lead to the transcription and release of key inflammatory mediators, including pro-interleukin-1 beta (pro-IL-1β), pro-IL-18, tumor necrosis factor-alpha (TNF-α), IL-6, and other cytokines and chemokines involved in the innate immune response. Within endosomal compartments, microbial nucleic acids are detected by TLR3, TLR7, TLR8, and TLR9, which activate interferon regulatory factors (IRFs), particularly IRF3 and IRF7, leading to the transcription of type I interferons (e.g., IFN-α and IFN-β), key antiviral cytokines that further amplify the immune response [18, 45, 46]. Another critical pattern recognition receptor is the NOD-like receptor family pyrin domain-containing 3 (NLRP3) inflammasome [47–49], a multi-subunit cytosolic protein complex that plays a central role in sterile inflammation and host defense. Inflammasome activation can be triggered by both endogenous danger signals and infectious stimuli, including bacterial toxins, viral RNA, mitochondrial dysfunction, and metabolic stress associated with systemic infection [41]. During the activation phase, the NACHT domain of NLRP3 undergoes oligomerization, facilitating the recruitment of apoptosis-associated speck-like protein containing a caspase activation and recruitment domain (ASC) through pyrin domain (PYD) interactions. ASC then engages pro-caspase-1 via CARD–CARD interactions, leading to proximity-induced auto-activation of caspase-1 [50]. Activated caspase-1 subsequently cleaves pro-IL-1β and pro-IL-18 into their mature, biologically active forms, and also processes gasdermin D (GSDMD). Cleavage of GSDMD generates an N-terminal fragment (GSDMD-NT) that oligomerizes to form membrane pores, resulting in pyroptotic cell death and the release of pro-inflammatory cytokines and DAMPs. The final step of membrane rupture is mediated by nerve injury-induced protein 1 (NINJ1), which facilitates the full release of intracellular contents and amplifies the local inflammatory response [18, 49, 50]. Human postmortem brain tissue from patients with AD presented cleaved GSDMD-positive cells demonstrating pyroptosis activation to AD-related neurodegeneration [51]. Activation of the NLRP3 inflammasome and subsequent ASC speck formation have emerged as key drivers of chronic neuroinflammation in AD [18, 48, 49]. During pyroptosis, extracellular ASC specks are released and function as DAMPs [49, 50], amplifying inflammation by activating surrounding microglia. In addition to their pro-inflammatory effects, ASC specks exhibit a high affinity for Aβ and p-Tau [48], facilitating the aggregation of these proteins and thereby contributing to the formation of amyloid plaques and neurofibrillary tangles [52], hallmark pathological features of AD [53]. Recent studies have identified ASC specks in blood and cerebrospinal fluid (CSF) [47, 52], and their diagnostic value has been enhanced when combined with other disease biomarkers. For example, the serum ratio (p-Tau-AT8 + ASC)/Aβ demonstrated a diagnostic accuracy of 92% for early-stage AD and 95% for late-stage AD. In comparison, (α-synuclein + ASC)/Aβ and ASC/Aβ ratios achieved up to 98% accuracy for PD [52]. These findings support the hypothesis that NLRP3 inflammasome activation is an early, measurable, and mechanistically relevant event in the pathogenesis of neurodegenerative diseases and neuroinflammation, Fig. 2 [47, 50, 54].Fig. 2The innate immune response during infection and AD.In a microglia cell, AD-related damage-associated molecular patterns (DAMPs), including amyloid-β (Aβ), phosphorylated tau (p-Tau), S100B, and HMGB1, or PAMPs, engage pattern recognition receptors (PRRs) such as TLR4, TLR3, TLR7, IL-1R, and RAGE. These interactions lead to downstream signaling via adaptor proteins (MyD88, TRIF, TRAM, and TIRAP), activating transcription factors (IRF and NF-κB), and upregulating pro-IL-1β, pro-IL-18, and TNF-α. A second signal, such as potassium (K⁺) efflux, mitochondrial reactive oxygen species (ROS), lysosomal cathepsin release, or ionophore treatment, triggers oligomerization of NLRP3 with ASC and procaspase-1, forming the NLRP3 inflammasome complex. Activated caspase-1 cleaves pro-IL-1β and pro-IL-18 into their active forms and processes Gasdermin D (GSDMD), releasing its N-terminal fragment (GSDMD-NT) that forms membrane pores. These pores facilitate additional K⁺ efflux and the release of IL-1β, IL-18, and other DAMPs, exacerbating neuroinflammation. NINJ1 is implicated in membrane rupture and pyroptosis. These processes ultimately drive neuroinflammatory responses and cell death, contributing to AD pathology. The innate immune response during infection and AD. In a microglia cell, AD-related damage-associated molecular patterns (DAMPs), including amyloid-β (Aβ), phosphorylated tau (p-Tau), S100B, and HMGB1, or PAMPs, engage pattern recognition receptors (PRRs) such as TLR4, TLR3, TLR7, IL-1R, and RAGE. These interactions lead to downstream signaling via adaptor proteins (MyD88, TRIF, TRAM, and TIRAP), activating transcription factors (IRF and NF-κB), and upregulating pro-IL-1β, pro-IL-18, and TNF-α. A second signal, such as potassium (K⁺) efflux, mitochondrial reactive oxygen species (ROS), lysosomal cathepsin release, or ionophore treatment, triggers oligomerization of NLRP3 with ASC and procaspase-1, forming the NLRP3 inflammasome complex. Activated caspase-1 cleaves pro-IL-1β and pro-IL-18 into their active forms and processes Gasdermin D (GSDMD), releasing its N-terminal fragment (GSDMD-NT) that forms membrane pores. These pores facilitate additional K⁺ efflux and the release of IL-1β, IL-18, and other DAMPs, exacerbating neuroinflammation. NINJ1 is implicated in membrane rupture and pyroptosis. These processes ultimately drive neuroinflammatory responses and cell death, contributing to AD pathology. Evaluating a different innate immune pathway associated with AD pathology, this study identified interferon-induced transmembrane protein 3 (IFITM3), classically induced during viral infection and innate immune activation, as a γ-secretase modulator that links inflammation to Aβ production [55, 56]. Inflammatory cytokines induce IFITM3 in neurons and astrocytes, where it binds γ-secretase and enhances its activity, increasing Aβ generation. IFITM3 expression increased with aging and AD-related genes in mouse models, while IFITM3 knockout reduces γ-secretase activity and amyloid plaque formation in 5XFAD mice. In patients with late-onset AD, IFITM3 was elevated and strongly correlated with γ-secretase activity. These findings reveal a mechanism by which neuroinflammation drives Aβ pathology via IFITM3-dependent modulation of γ-secretase [55]. A similar shift toward the amyloidogenic processing of APP has also been described in experimental models of sepsis, further supporting the link between systemic inflammation and neurodegenerative pathways [57]. Recent advances in animal models of delirium have improved the ability to study how systemic inflammatory insults, including sepsis and infection, induce acute brain dysfunction and potentially accelerate long-term neurodegenerative processes [58]. During sepsis, increased activity of both β- and γ-secretases has been observed, particularly in later stages of the disease, promoting amyloidogenic APP cleavage and Aβ production. This increase in Aβ generation may contribute to neuroinflammation and the cognitive deficits observed after sepsis. Pharmacological inhibition of these secretases has shown neuroprotective effects: inhibition of either β- or γ-secretase improved cognitive performance 10 days after sepsis induction, while β-secretase inhibition sustained cognitive benefits for up to 30 days. Moreover, β-secretase inhibition reduced brain levels of Aβ and the pro-inflammatory cytokine IL-1β. Together, these findings suggest that inflammatory states such as sepsis may activate amyloidogenic pathways similar to those observed in AD, potentially through mechanisms linking innate immune activation, secretase activity, and Aβ accumulation. The CNS immune response and its connection with AD pathology Skull bone marrow, CNS barriers, and the resident immune cells Peripheral infections can influence brain immune responses through multiple anatomical and cellular interfaces connecting the systemic immune system with the CNS. These include skull bone marrow niches, meningeal immune compartments, the choroid plexus, and vascular barriers, which collectively regulate immune surveillance and neuroinflammatory signaling [59]. The skull and vertebral bone marrow serve as reservoirs of both myeloid and lymphoid cells that contribute to immune surveillance of the meninges and CNS parenchyma [60–62]. The skull bone marrow can be stimulated when CSF directly accesses it through specialized dural channels, allowing bidirectional exchange of molecular signals and immune cues to flow between the CNS and bone marrow niches [63]. This bidirectional communication enables the CNS to influence hematopoietic activity and immune cell behavior within the skull marrow [60, 61, 63]. Recently, an experimental bacterial meningitis model demonstrated that CSF exits into the skull marrow and promotes cranial hematopoiesis, indicating a functional role in shaping local immune responses during meningitis [64]. In another study, neutrophils derived from the skull marrow migrated more efficiently into adjacent brain tissue than those from the tibia, facilitating the vascular connections between the skull bone marrow and the brain surface for the migration of neutrophils [62]. In transgenic mouse models of AD, such as the 5xFAD model, the skull bone marrow exhibits a distinct molecular and cellular profile compared to wild-type controls. Single-cell RNA sequencing (scRNA-seq) and bulk RNA analyses revealed a significant upregulation of pro-inflammatory genes, particularly in neutrophils and myeloid cells, including Nr4a1, Cxcl2, Il1b, and Ptgs2, which are associated with immune cell activation, migration, and tissue remodeling. Notably, the skull bone marrow of 5xFAD mice showed a convergence with meningeal immune cell signatures, suggesting enhanced immune communication across the skull-meninges-brain axis in AD pathology. In the same study, skull bone marrow from patients with AD showed a distinct proteomic profile and evidence of neuroinflammation. Notably, translocator protein (TSPO)-PET imaging revealed increased inflammatory signals in the skull, which correlated with advanced tau pathology and reduced CSF Aβ42 levels. These findings suggest that skull bone marrow mirrors central neuroinflammatory processes, particularly in the later stages of AD, and may serve as a peripheral window into brain pathology [65]. In addition to the skull bone marrow, distinct immunological niches within the brain have been identified, including the meninges [66], choroid plexus (CP) [67], and perivascular spaces [68]. These specialized compartments function as significant immunological interfaces, allowing immune cells to monitor and respond to CNS signals while preserving its delicate neuronal architecture [59, 69]. The meninges represent a complex area containing populations of myeloid and lymphoid cells, including monocytes, neutrophils, eosinophils, innate lymphoid cells type 2 (ILC2s), NK cells, dendritic cells, macrophages, mast cells, B cells, plasma cells, CD4+ and CD8+ T cells, and NK-T cells [70–72]. In addition, the CP comprises a single layer of cuboidal ciliary epithelial and mesenchymal cells covering a highly vascularized core stroma composed of connective tissue, pericytes, and blood vessels. Also, many immune cells, including ILCs, B cells, lymphocytes, macrophages, basophils, dendritic cells, monocytes, and neutrophils, naturally populate the CP. The most abundant are macrophages, which exhibit a spatial organization associated with age-dependent changes [73–75]. Together, the skull bone marrow and CNS border niches form a dynamic immunological interface that shapes and reflects neuroinflammatory processes across both health and disease. These observations suggest that peripheral inflammatory signals, including those generated during systemic infection, may influence cranial hematopoiesis and promote the recruitment of immune cells into the CNS [59, 76]. CNS barriers and the migration of peripheral immune cells During infections, circulating cytokines and microbial products can activate endothelial cells, increasing the expression of adhesion molecules and facilitating the recruitment of peripheral immune cells into CNS border regions [70]. The CNS has three major barrier systems that compartmentalize the brain, including the vascular barriers, the CP epithelium, and the meningeal arachnoid barrier. The immune cells that traffic from the periphery to the CNS are rare in the healthy CNS; however, these barriers facilitate neuroimmune interactions in the CNS [59]. Immune cells migrate across the CNS vasculature, a process facilitated by endothelial cell activation by inflammatory mediators, such as cytokines. Upon activation, endothelial cells upregulate adhesion molecules, including intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), which promote leukocyte rolling and diapedesis toward the source of pro-inflammatory chemotactic signals [77]. The CP vasculature, characterized by fenestrations and features of peripheral vessels, permits homeostatic immune cell trafficking. Similar to immune cells crossing the CNS vasculature, those entering the CP do not immediately access the brain parenchyma and accumulate within the CP stroma. Then those cells traverse the CP epithelium, after which they may circulate and access the subarachnoid space, serving as a site of neuroimmune communication between the periphery and the brain [59]. Immune cells can also migrate from the dura mater to the subarachnoid space via the arachnoid barrier. The dural venous sinuses exhibit high homeostatic expression of cell adhesion molecules, including ICAM-1 and VCAM-1, which facilitate the continuous trafficking of patrolling immune cells within the healthy dura mater [66]. Additionally, systemic inflammation can disrupt the BBB through cytokine-mediated endothelial dysfunction and tight junction protein degradation, facilitating the entry of peripheral inflammatory mediators and neurotoxic molecules into the brain, a mechanism that may contribute to both sepsis-associated encephalopathy and the progression of AD. In experimental sepsis models, increased neuroinflammation was observed [78], and increased BBB permeability occurs in parallel with elevated levels of matrix metalloproteinases (MMP)-9 and MMP-2 in brain microvessels of both the cortex and hippocampus [57]. Pharmacological inhibition of MMP-2 and MMP-9 reverses BBB disruption and is associated with reduced brain levels of IL-6 and decreased oxidative damage, supporting a mechanistic role for MMP-mediated BBB breakdown in inflammation-driven neurodegeneration. The involvement of diverse peripheral immune cell populations in the progression of AD has been reported [79–82]. The CD8⁺ T cells accumulate in the cerebrospinal fluid (CSF) of patients with sporadic AD, where their presence inversely correlates with cognitive performance [79]. In preclinical studies, the depletion of natural killer (NK) cells improved cognitive function, and microglial cells demonstrated a homeostatic-like morphology in a triple-transgenic AD mouse model (3xTg-AD) [83]. In 5xFAD and 3xTg-AD transgenic mouse models, neutrophils were observed to extravasate into the CNS, localizing to regions of Aβ plaques where they released neutrophil extracellular traps (NETs) and IL-17 [84]. The role of extracellular vesicles (EVs) in neuroinflammation Extracellular vesicles (EVs) are phospholipid membrane-enclosed structures released into the extracellular space and represent a highly heterogeneous group [85, 86]. They are generally classified according to their biogenesis, including exosomes that are small vesicles (~30 to 150 nm) that originate within late endosomes through inward budding of the endosomal membrane and are secreted upon fusion of multivesicular bodies (MVBs) with the plasma membrane [86]. In contrast, ectosomes are larger vesicles (~50 nm to >1,000 nm) that form by outward budding and shedding of the plasma membrane; this category includes microvesicles, apoptotic bodies, and large oncosomes [87]. EVs can cross the BBB in both directions, facilitating molecular exchange between the CNS and peripheral tissues. Several mechanisms have been proposed to explain EV transport across the BBB, including receptor-mediated transcytosis, adsorptive-mediated transport, and paracellular passage under conditions of barrier dysfunction [85]. Through these mechanisms, EVs originating from peripheral tissues, including immune cells, endothelial cells, platelets, and gut-derived cells, can reach the brain and influence neuroinflammatory signaling. Conversely, EVs released by neurons, astrocytes, microglia, oligodendrocytes, and brain endothelial cells can enter the systemic circulation and serve as accessible biomarkers reflecting ongoing neuropathological processes. This bidirectional trafficking allows EVs to function as long-distance signaling vehicles linking peripheral inflammation, infection, and metabolic alterations with CNS responses [88]. Every immune cell type involved in the inflammatory response can release EVs [86]. During pyroptosis, inflammasome activation triggers a substantial release of exosomes. This process has been linked to inflammasome-driven cleavage of the Rab-interacting lysosomal protein (RILP), which, in turn, regulates the selective incorporation of microRNAs (non-coding RNA molecules) into EVs [89]. In sepsis, EVs can transport acute-phase proteins, such as C-reactive protein (CRP), which promotes the release of IL-8 from monocytes, thereby driving neutrophil recruitment [90]. EV surfaces can also bind complement factors and regulators, and antibody-bearing EVs may act as decoys to protect releasing cells. Thus, EVs modulate innate immunity in a context-dependent manner, exerting both pro and anti-inflammatory effects by transferring proteins, lipids, nucleic acids, and danger signals. Additionally, EVs-mediated communication is not restricted to host cells, many microorganisms, including Gram-negative and Gram-positive bacteria, release vesicle-like structures known as bacterial outer membrane vesicles (OMVs) or membrane vesicles, which contain a variety of microbial components such as LPS, bacterial DNA and RNA, enzymes, toxins, and virulence factors [91]. Recently, gut microbiota extracellular vesicles (GMEVs) have also been implicated in the microbiota‐gut brain communication carriers [92]. These microbial vesicles can enter the systemic circulation and interact with host immune cells, activating innate immune pathways through pattern recognition receptors such as TLR2 and TLR4. OMVs and GMEVs have been shown to trigger inflammatory signaling, promote cytokine production, and modulate host immune responses even in the absence of direct bacterial invasion. Emerging evidence suggests that microbial vesicles derived from the gut microbiota or from infectious pathogens may reach distant organs and influence inflammatory signaling in the CNS [92, 93]. Through these mechanisms, microbial EVs may represent an additional pathway by which peripheral infections and gut dysbiosis contribute to systemic inflammation, neuroimmune activation, and potentially neurodegenerative processes associated with AD. EVs carry disease-associated proteins implicated in several neurodegenerative disorders, including Prion disease, AD, PD, and ALS [94, 95]. In AD experimental models and postmortem human brain tissue, numerous proteins, lipids, and regulatory RNAs have been detected in brain-derived EVs [88, 96]. These vesicular cargos have been investigated as potential biomarkers for prediction, diagnosis, prognosis, disease monitoring, and assessment of therapeutic response [97]. In AD, EVs released by neurons, astrocytes, microglia, and oligodendrocytes carry hallmark pathological proteins (Aβ, Tau), synaptic (neurogranin, neuromodulin, SNAP25, synaptophysin) and mitochondrial markers (reduced levels of electron transport chain complexes but increased mitochondrial RNAs, reflecting mitochondrial dysfunction), inflammatory mediators, complement proteins, and dysregulated RNAs. These cargoes not only serve as biomarkers but also actively contribute to the propagation of pathology and neuroinflammation [86, 98] and represent a novel therapeutic avenue, Fig. 3. Together, these findings highlight that EV-mediated communication is a conserved biological mechanism operating across both host and microbial systems, with potential implications for neuroinflammation and the progression of neurodegenerative disease.Fig. 3Neuroinflammatory cascades linking infection and AD pathology.Systemic infections elevate circulating cytokines, chemokines, and matrix metalloproteinases (MMPs), leading to blood-brain barrier (BBB) disruption through loss of tight junction proteins (claudin, occludin, ZO-1). Peripheral immune mediators and cells infiltrate the CNS, amplifying inflammatory signalling. In the brain parenchyma, astrocytes exhibit reduced neurotransmitter uptake and impaired potassium buffering, releasing reactive oxygen species (ROS) and cytokines. Activated microglia secrete cytokines, complement proteins (C3, C1q), and inflammasome components (NLRP3, ASC, LUBAC), fueling chronic neuroinflammation. Neurons under stress release extracellular vesicles (EVs) carrying amyloid-β (Aβ), phosphorylated tau (p-Tau), neurofilament light chain (NfL), and neurotoxic proteins. ASC specks act as nucleation sites, accelerating aggregation of Aβ and tau into plaques and tangles. Together, these processes create a feed-forward loop of neuroinflammation, synaptic dysfunction, and neuronal injury, ultimately contributing to AD progression. Neuroinflammatory cascades linking infection and AD pathology. Systemic infections elevate circulating cytokines, chemokines, and matrix metalloproteinases (MMPs), leading to blood-brain barrier (BBB) disruption through loss of tight junction proteins (claudin, occludin, ZO-1). Peripheral immune mediators and cells infiltrate the CNS, amplifying inflammatory signalling. In the brain parenchyma, astrocytes exhibit reduced neurotransmitter uptake and impaired potassium buffering, releasing reactive oxygen species (ROS) and cytokines. Activated microglia secrete cytokines, complement proteins (C3, C1q), and inflammasome components (NLRP3, ASC, LUBAC), fueling chronic neuroinflammation. Neurons under stress release extracellular vesicles (EVs) carrying amyloid-β (Aβ), phosphorylated tau (p-Tau), neurofilament light chain (NfL), and neurotoxic proteins. ASC specks act as nucleation sites, accelerating aggregation of Aβ and tau into plaques and tangles. Together, these processes create a feed-forward loop of neuroinflammation, synaptic dysfunction, and neuronal injury, ultimately contributing to AD progression. The overlap of innate and adaptive immune responses from infection and AD The involvement of the innate and adaptive immune response hypothesis in AD pathogenesis [99] has been supported by genome-wide association studies (GWAS), with 25% of the identified potential AD genetic risk factors being highly or exclusively expressed by microglia and/or linked to immune-related functions [100, 101]. Within the CNS, microglia serve as the resident immune sentinel cell, continuously monitoring the neural environment and engaging with neurons, astrocytes, oligodendrocytes, and infiltrating immune cells. In physiological conditions, they regulate synaptic remodeling, mediate repair processes, maintain tissue homeostasis, clear cellular debris, and support the function of other glial cells [102]. Additionally, microglia can become primed following systemic inflammatory challenges and respond rapidly and strongly in models of neurodegenerative disease [103]. Single-cell and single-nucleus transcriptomic analyses of isolated human microglia have identified several small subclusters, termed microglial states, that deviate from the homeostatic phenotype, more closely associated with Cluster 1-2, and with AD microglia phenotypes exhibiting distinct patterns of marker, including gene upregulation associated with proliferation (Cluster 9), anti-inflammatory (Cluster 5-6), interferon response (Cluster 4), and antigen presentation (Cluster 7, CD74^high, overlaps with the disease-associated microglia (DAM) signature described initially in mice) [104, 105]. Microglia are therefore viewed as a dynamic ecosystem of cellular states that shift in response to brain conditions such as aging and neurodegeneration. Transcriptomic enrichment analyses further suggest that cluster 4 is associated with multiple sclerosis, clusters 5 and 6 with several neurological disorders, including neurovascular disease, encephalitis, neoplastic diseases, and AD, whereas cluster 7, enriched for DAM genes, showed strong associations with inflammatory demyelination, ischemia, and AD [104]. Microglia and astrocytes are the key cellular mediators of neuroinflammation in human brains with AD pathology [17], as well as during environmental insults, including tissue injury and microbial infection [106]. Reactive astrocytes are a population of astrocytes that undergo structural, molecular, and functional changes in response to CNS injury, infection, or disease [107]. When exposed to Aβ, astrocytes release higher amounts of cytokines and enhance microglial uptake of Aβ. However, astrocytes harboring the APOE4 risk allele for AD exhibit diminished priming capacity and are less effective at supporting microglial phagocytosis [106]. Transcriptomic and proteomic profiling of microglia and astrocytes has confirmed the expression of the linear ubiquitin assembly complex (LUBAC). The LUBAC complex consists of three essential subunits: HOIP (RNF31), the catalytic subunit responsible for M1-linked ubiquitin chains; HOIL-1L (RBCK1), which provides additional ubiquitin-binding and ligase activity; and SHARPIN, which stabilizes the complex and enhances its function. This E3 ligase complex serves as a critical regulator of NF-κB signaling [108] downstream of TLR and TNF-α activation, promoting the expression of pro-inflammatory mediators, including IL-1β, TNF, and IL-6. Beyond canonical NF-κB regulation, LUBAC has also been implicated in the priming and activation of the NLRP3 inflammasome [17] and is essential for the initiation and progression of Aβ pathology in a murine model [47]. Furthermore, LUBAC activity may contribute to tau pathology through Aβ-driven microglial activation. In addition to its role in inflammatory signaling, LUBAC has been associated with autophagic processes, in which linear ubiquitin chains decorate TDP-43-positive cytoplasmic inclusions in neurons, and are proposed to facilitate their clearance through selective autophagy [17, 109, 110]. In a large GWAS meta-analysis of over 111,000 AD cases and 677,000 controls, gene prioritization identified SHARPIN and RBCK1, two subunits of the LUBAC, as genes associated with an increased risk of AD. Pathway enrichment analyses highlighted immune and TNF-signaling pathways, implicating dysregulated linear ubiquitination in AD susceptibility [111], supporting a central role for immune and inflammatory pathways in disease development. Genetic evidence also suggests that adaptive immune mechanisms are involved in combating Tau and Aβ42 [112]. Additionally, using single-cell TCR sequencing, the authors of this study found clonally expanded CD8+ effector memory T cells reexpressing CD45RA (TEMRA cells, which present a cytotoxic and pro-inflammatory profile) in the CSF of patients with AD and mild cognitive impairment (MCI) [79]. In another study, extravascular CD3 + T cells in the brains of patients with AD correlate with Tau but not with amyloid pathology [113]. In this study, the data show that subjects at the asymptomatic, preclinical stage of AD can mount a CD4+ helper T cell response to the Aβ peptide and present CD8 + T cells reexpressing CD45RA in the CSF, along with a less immunosuppressive genetic signature of peripheral regulatory T cells. Conversely, in MCI, increased frequencies of TEMRA/CD8+ effector cells were observed in the periphery, characterized by pro-inflammatory gene expression [114, 115]. T helper (Th) cell subsets exhibit distinct cytokine profiles. Th1 cells mainly produce IFN-γ and TNF-α, driving cell-mediated immunity. Th2 cells secrete IL-4, IL-5, and IL-13, which promote humoral and allergic responses. Th17 cells are characterized by the production of IL-17A, IL-17F, and IL-22, which mediate protection against extracellular bacteria and fungi, while also contributing to the development of autoimmunity. Additional subsets, such as regulatory T cells (Tregs) and T follicular helper cells (Tfh), further diversify the functional landscape of CD4⁺ T cell responses. Using single-cell mass cytometry, this study shows that the abundance of Th1 cells in the peripheral blood of patients with AD negatively correlated with plasma neurofilament light chain (NfL) levels, a biomarker of axonal injury and neurodegeneration. Additionally, analysis of public scRNA-seq datasets showed that, unlike peripheral blood, Th1 cells were enriched in the CSF of AD patients compared with healthy controls. These cells exhibited an enhanced pro-inflammatory transcriptional profile [116], with a more pro-inflammatory gene expression profile (enhanced IFN-γ signaling) and increased infiltration of peripheral immune cells into the CNS [17]. Another study demonstrated that plasma NfL levels were a more robust marker of AD in men [117]. Microbiota-gut brain axis and AD pathology Microbiota-gut-brain communication pathways and microbial metabolites Growing evidence indicates that alterations in the gut microbiota may contribute to neuroinflammatory processes and neurodegeneration via the microbiota-gut-brain axis [118, 119]. This bidirectional communication system links intestinal microbial communities with the central nervous system through immune, endocrine, metabolic, and neural pathways, highlighting neuronal connections and chemical messaging [120]. The autonomic nervous system (ANS) is linked to the immune system through the sympathetic and parasympathetic divisions and plays a pivotal involuntary role in regulating physiological functions and maintaining homeostasis. It facilitates bidirectional communication within the microbiota-gut-brain axis, influencing crucial gastrointestinal functions and responding to environmental stimuli through feedback loops. The vagus nerve (cranial nerve X), along with pelvic afferents, is a major component of this communication network, providing a direct link between the gut and the brain [121]. Approximately 80–90% of vagal fibers are afferent (gut-to-brain), while 10–20% are efferent (brain-to-gut), and the vagal afferent fibers originate in the gut wall and transmit sensory signals to the brainstem, particularly the nucleus tractus solitarius (NTS) in the medulla [120, 121]. Activation of vagal afferent pathways can occur through multiple signaling mechanisms, including microbial metabolites such as short-chain fatty acids (SCFAs), inflammatory mediators such as cytokines (IL-1β and TNF-α), bacterial components including lipopolysaccharide (LPS, accounting for 75% of the external surface of Gram negative bacteria; lipopeptides, lipoteichoic acid, etc.), and enteroendocrine signaling mediated by serotonin (5-HT) released from enterochromaffin cells [122]. Microbiota-host interactions within the gut also play an important role in systemic and neural communication by promoting the release of cytokines, chemokines, neurotransmitters, neuropeptides, endocrine signals, and microbial metabolites. These molecules can enter the circulation via the blood and lymphatic systems or influence neural signaling via vagal and spinal afferent pathways, enabling continuous communication between the gut and the brain [123]. Importantly, dysbiosis-related inflammatory signaling may also increase intestinal barrier permeability and compromise the integrity of the BBB, facilitating the translocation of microbial products and inflammatory mediators into the systemic circulation and potentially into the brain, where they may contribute to neuroinflammation and neurodegenerative processes [118, 124]. SCFAs, primarily acetate (2 carbons), propionate (3 carbons), and butyrate (4 carbons), are produced by commensal anaerobic bacteria in the colon through the fermentation of dietary fibers and resistant starches [118, 123, 125]. These metabolites interact with host receptors and modulate epithelial cytoskeletal organization and tight junction protein expression, thereby reducing intestinal permeability and limiting the translocation of bacterial endotoxins such as LPS into the systemic circulation [126]. SCFAs also exert distal effects on the BBB, where they increase claudin-5 and occludin expression and reduce BBB permeability [127, 128]. In addition to SCFAs, gut microbiota also regulate the production and transformation of other biologically active metabolites that influence neuroimmune signaling, including bile acids and tryptophan-derived metabolites. In addition, the liver contributes to maintaining intestinal microbial homeostasis by secreting bile acids into the intestinal lumen via the biliary tract [122]. These bile acids not only facilitate lipid digestion but also exert antimicrobial effects that help control bacterial growth in the gut. For example, bile acids can activate the farnesoid X receptor (FXR) in intestinal epithelial cells, inducing the production of antimicrobial peptides such as angiogenin-1 and RNase family member 4, which limit bacterial overgrowth and help preserve intestinal barrier integrity [121]. Furthermore, microbiota-derived metabolites, SCFAs, tryptophan metabolite indole, indole-3-aldehyde, indole-3-propionic acid, metabolites generated through the kynurenine pathway, and some secondary bile acids (for example, lithocholic acid), which act as physiological ligands for aryl hydrocarbon receptor (AHR), a ligand-activated transcription factor involved in immune regulation in the gut and in the central nervous system [121]. In the CNS, AHR signaling has been shown to modulate glial cell function by suppressing pro-inflammatory gene expression in astrocytes and microglia, thereby reducing cytokine production, including IL-1β, IL-6, and TNF-α. Through these mechanisms, microbiota-derived SCFAs and AHR-mediated pathways contribute to microglial maturation, astrocyte regulation, and the maintenance of neuroimmune homeostasis [129]. Disruption of these microbiota-dependent regulatory mechanisms, gut dysbiosis has been shown to reduce SCFA production in several neuroinflammatory [125] and neurodegenerative conditions, including AD [119], sepsis [130], and PD [131]. This reduction in SCFAs may impair intestinal barrier integrity and BBB function, thereby disrupting brain homeostasis and promoting neuroinflammation, Fig. 3 [118]. Translocation of bacterial components beyond the gastrointestinal tract drives systemic and neuroinflammation Disruption of intestinal barrier integrity during gut dysbiosis can facilitate the translocation of microbial products from the oral cavity and gastrointestinal tract into the systemic circulation. These bacterial components can trigger peripheral immune activation and promote systemic inflammation, which may, in turn, influence central nervous system immune responses. Increasing evidence suggests that microbially derived molecules from the gut microbiota may represent an important mechanistic link between peripheral infections, systemic inflammation, and neurodegenerative processes [118]. Elevated circulating levels of endotoxin, particularly LPS, have been reported in the plasma of patients with sporadic amyotrophic lateral sclerosis (sALS) and AD compared with healthy controls [96]. Consistent with the concept that severe systemic inflammation is accompanied by increased exposure to bacterial products, elevated LPS-related inflammatory signatures have also been reported in critical illness, including sepsis [132, 133]. In parallel, PS-binding protein (LBP), an acute-phase reactant that mediates immune responses triggered by LPS and has been used as a blood marker for endotoxin exposure, has been associated with PD risk in prospective cohort analyses [96]. Postmortem analyses have demonstrated that LPS was present at higher levels in AD brains than in age-matched controls and colocalizes with amyloid plaques, perivascular amyloid deposits, neurons, and oligodendrocytes [134]. Escherichia coli K99 pili protein and LPS levels were greater in postmortem AD brains compared to control brains. LPS colocalized with Aβ1-40/42 in amyloid plaques and around vessels in AD brain [135]. These findings suggest that bacterial products may accumulate in the brain and interact with key pathological structures associated with neurodegeneration [134]. Additionally, postmortem analyses have repeatedly identified LPS from periodontal bacteria P. gingivalis and E. coli, mostly present in the gut, in AD brains [135, 136], with LPS accumulating in neocortical and hippocampal neurons [137]. A recent experimental study demonstrates that low concentrations of P. gingivalis LPS or Aβ alone do not directly induce neuronal synaptic alterations. However, microglia respond sensitively to these stimuli, becoming activated and producing TNF-α through TLR-2 signaling. While microglial reactivity is part of the brain’s innate immune defense against pathogens, sustained or excessive cytokine release can promote chronic neuroinflammation, contributing to synaptic dysfunction and disease progression [138]. In a preclinical study, LPS induced Aβ and tauopathy in rat brain [134], and in postmortem brain samples, LPS promotes amyloidogenic processing of the amyloid precursor protein, increases Aβ accumulation, and contributes to oligodendrocyte injury and myelin disruption [134, 135]. Mechanistically, LPS can activate innate immune signaling via TLR4 and CD14 expressed on microglia and other immune cells, triggering NF-κB-mediated transcription of pro-inflammatory cytokines [135]. Together, these findings suggest that the translocation of bacterial endotoxins such as LPS beyond the gastrointestinal tract may link peripheral infections with systemic inflammation and neurodegenerative processes. The detection of endotoxins in AD brains and their association with key pathological features support the hypothesis that microbial products and endotoxin-driven immune activation may contribute to AD pathology. Hypothesis linking microbial amyloids and cross-seeding mechanisms in AD pathology Since their initial discovery in 2002, bacterial amyloids have been identified in numerous bacterial species, including E. coli (CsgA and CsgB; curli fibers), Salmonella typhimurium (curli), P. aeruginosa (FapC), Klebsiella pneumoniae (microcin E492), Mycobacterium (Type IV pili, CarD), Clostridium (Rho), Bacillus subtilis (TasA), and Staphylococcus aureus (phenol-soluble modulins, PSM) [139, 140]. These functional amyloids are structural components of bacterial biofilms and play important roles in microbial adhesion and host-microbe interactions. In the gastrointestinal tract, microbial amyloids can interact with the host immune system by activating innate immune pathways, and experimental studies have shown that bacterial amyloids can stimulate macrophages and other immune cells through PRRs, including MHC molecules, TLR2, and TLR4, leading to activation of NF-κB signaling and the release of pro-inflammatory cytokines [140, 141]. This inflammatory response may also promote the generation of reactive oxygen species and contribute to systemic immune activation [142]. In a preclinical model of systemic infection triggered by laparotomy and gut polymicrobial infection, the data did not detect Aβ in the gut; however, Aβ deposition was exacerbated 130 days after surgery in the brains of APP/PS1 mice [46]. Conclusion and future directions Collectively, converging epidemiological, mechanistic, and translational evidence supports a conceptual framework in which infections act as upstream triggers of neuroinflammatory cascades that may accelerate AD pathology. Although infections alone are unlikely to initiate AD, repeated immune challenges across the lifespan may lower the threshold for neurodegeneration in genetically, metabolically, or environmentally susceptible individuals. Heneka and colleagues (2015) highlighted that AD is not exclusively a neuronal disorder but rather involves profound immune and inflammatory mechanisms within the brain, demonstrating that microglia and astrocytes play central roles in disease initiation, progression, and chronic neurodegeneration [48, 143]. Later, they emphasized that the inflammasome cascade is a unifying pathogenic mechanism across diseases such as AD, PD, multiple sclerosis (MS), ALS, stroke, and sepsis [18]. In this context, glial cell reactivity has also been linked to gut microbiota dysbiosis, suggesting that pro-inflammatory signals originating in the intestine may modulate the brain’s immune response and amplify neuroinflammation. Thus, although the gut microbiota is often described quantitatively using diversity indices or relative abundance measures, it represents far more than numbers, as such metrics fail to fully capture its complex functional and biological relevance in bidirectional communication with the brain [144, 145]. The gut-brain axis plays a pivotal role in neurodegenerative disorders, where infection- or dysbiosis-induced immune activation contributes to neuropathology. In AD, systemic inflammation, sepsis, microbial dysbiosis, and altered gut permeability amplify Aβ and tau pathology by disrupting the BBB and chronic microglial activation. The infection and neurodegenerative diseases share overlapping pathways involving NLRP3 inflammasome activation, ASC speck propagation, and peripheral-to-central immune signaling, ultimately leading to barrier dysfunction, mitochondrial stress, and progressive neuronal injury. Microbial metabolites such as SCFAs typically maintain gut and BBB integrity by upregulating tight junction proteins (claudin-5, occludin, ZO-1) and supporting microglial homeostasis. However, inflammation, infection, or dysbiosis reduces SCFA production, exacerbating barrier leakage and systemic inflammation, thus perpetuating a feed-forward cycle of neuroinflammation. Collectively, these mechanisms underscore how peripheral immune dysregulation translates into central pathology, linking infection, gut dysbiosis, and neurodegeneration. Future studies should aim to define microbial and viral signatures that accelerate neurodegenerative diseases through longitudinal microbiome and metabolomic profiling. The contribution of gut-derived EVs to the transport of inflammatory and/or pathogenic cargo into the brain warrants further investigation. Additionally, therapeutic efforts should focus on inhibiting inflammasome activation, restoring the SCFA-producing microbiota, and protecting the integrity of the intestinal and BBB. Integrating immune imaging and multi-omic analyses [146, 147] may enable monitoring of infection-driven neuroinflammation and inform the development of personalized interventions. Finally, expanding preventive immunomodulatory strategies, including routine vaccination, may represent a pragmatic and scalable approach to reducing infection-associated neurodegenerative risk. Converging epidemiological data suggest that immunization against common pathogens is consistently associated with lower dementia incidence across diverse populations. However, it is important to acknowledge that most available evidence derives from retrospective and observational cohort studies, including claims-based analyses, and therefore does not establish causality. Prospective and quasi-experimental studies will be essential to clarify the extent to which vaccination confers direct neuroprotective effects versus reflecting broader health behaviors or systemic resilience. Supplementary information Supplemmentary Table 1 Supplemmentary Table 1