Increasing scientific evidence consistently confirms the significance of immunity in the pathogenesis of AD. These findings increasingly suggest a combined involvement of both innate and adaptive immunity. TNF exhibits a continual, low-level expression in the typical adult brain [53,54,55,56]. TNF-α is secreted by many cell types, albeit cells of the monocytic lineage, including microglia and are the primary synthesizers [57]. TNF-α modulates and influences many neuronal activities in the CNS, which include glutamatergic transmission [58,59,60] and gamma-aminobutyric acid (GABA) transmission [58,61,62], with an overall outcome of these alterations of an augmentation of excitatory synaptic transmission compared to inhibitory transmission in hippocampal slices [58] and potentially in other neuronal circuits [56]. The strengthening of excitatory transmission is further amplified by the release of glutamate from astrocytes influenced by TNF-α [56,63].

TNF-α is involved in a variety of activities within the CNS. It influences sleep [64] directly acting on neurons in key sleep regulatory areas of the brain, such as the hypothalamic preoptic area and basal forebrain, impacting sleep architecture [66]. Intracerebroventricular injection of TNF-α induces a significant increase in slow-wave sleep, as demonstrated by polysomnography, emphasizing its role in promoting specific sleep stages [67]. In mice, the administration of TNF-α to the peritoneum leads to significant elevations in non-rapid eye movement sleep duration while causing reductions in slow wave activity [68].

The roles of TNF-α receptors (TNFR) in CNS physiology are different. TNFR1 exhibits neuroprotective effects in the CNS by safeguarding cells against necrosis [69]. Despite the anti-apoptotic attributes of TNFR1 signaling, it only partially elucidates its neuroprotective role in acute CNS injuries. Notably, TNFR1 fails to shield neurons from insults like retinal ischemia–reperfusion injury and oligodendrocytes from immune-mediated damage in conditions such as multiple sclerosis (MS) and experimental autoimmune encephalomyelitis (EAE) [56]. Studies exploring caspase inhibition as a potential remedy for TNF toxicity yielded surprising results. Pan-caspase inhibition exacerbated TNF toxicity in mice, intensifying oxidative stress and mitochondrial damage [70]. TNFR1-mediated cell death involves a switch between caspase-induced apoptosis and necrosis, with key signaling molecules such as receptor-interacting protein 1 (RIP1) and RIP3 identified in TNF-induced programmed necrosis or necroptosis [71,72]. This reveals TNFR1 signaling and the Fas-associated death domain protein (FADD)/caspase 8 apoptotic platform as an additional protective layer, enabling cells to undergo apoptosis instead of inflammation-inducing necroptotic cell death. The neuroprotection mediated by TNFR1 in the CNS may operate on two levels: gene induction and de novo production of survival molecules and, in the absence of adequate survival signals, caspase 8-dependent apoptosis [56].

TNFR2 also plays a crucial role in neuroprotection and repair processes in the CNS. TNFR2 is expressed in regulatory T cells, endothelial cells [73], oligodendrocyte lineage cells [74], and specific neuron populations [50] and is able to activate pro-survival signaling through TRAF2 recruitment [75] and PI3K/NF-κB pathway activation [76,77]. TNFR2 activation is associated with Akt/protein kinase B activation, crucial for neuroprotection against excitotoxicity and ischemia–reperfusion injury [56]. Actually, TNFR2 and TNFR1 cooperate to enhance neuroprotection, with additive effects on cell survival through NF-κB activation. Indeed, TNFR2 reinforces TNFR1-mediated cell survival and apoptotic signaling, maximizing neuroprotection [77,78]. Finally, TNFR2 may also promote oligodendrocyte precursor cell proliferation and remyelination, contributing to repair mechanisms [78,79].

However, TNF-α represents one of the pivotal pro-inflammatory cytokines and triggers the production of IL-1, IL-6, and IL-8, fostering chronic inflammation unless balanced by anti-inflammatory cytokines like IL-10 [80]. It induces APP and BACE1 expression in mouse astrocytes, activating γ-secretase in HEK cells, and releasing Aβ peptides [24,25,26]. Persistent brain inflammation forms a self-amplifying cycle, sustaining high TNF-α levels, potentially stimulating Aβ synthesis, causing neuronal loss, and hindering microglia Aβ phagocytosis [27]. The role of TNF-α in tau hyperphosphorylation is less understood, but few interesting data on mice suggest a possible connection [81,82].

Historically, microglia activation and gliosis were perceived as secondary to neurodegeneration in AD. However, recent genetic investigations in late-onset AD indicate the involvement of microglia- and astrocyte-related pathways together with the activation of multiple immune pathways in the very early stages of the disease [6,10,11,83,84,85,86,87,88]. Particularly, the innate immune receptor triggering receptor expressed on myeloid cells 2 (TREM2), expressed on microglia and myeloid cells, has garnered attention [89,90]. Rare TREM2 mutations linked to AD propose that TREM2 deficiency plays a role in AD susceptibility [91]. Functioning as a negative regulator, TREM2 modulates the release of TNF-α and other inflammatory cytokines via the Toll-receptor pathway [92]. Impairment of TREM2 function in monocytes or macrophages may contribute to systemic TNF-α production, potentially serving as a treatable risk factor for AD [91,93].

The TNF-α gene promoter region contains multiple single nucleotide polymorphisms (SNPs), with the G308A mutation being of interest for AD [94]. This mutation elevates TNF-α mRNA and protein expression and a possible pathogenetic mechanism has been proposed. However, conflicting meta-analyses reveal regional variations, emphasizing the need for further investigation into the correlation between TNF-α G308A and AD risk [95].

Altogether, these insights underscore the intricate interplay of immune responses in AD pathogenesis and highlight the role of TNF-α. In the following paragraph, we will review the main evidence in human and animal studies and finally, we will discuss the possible future directions.

The primary evidence for this association derives from an initial study performed on the CSF of mild cognitive impairment (MCI) patients which demonstrated higher TNFα and tau protein levels and lower TGFß and Aß levels in MCI patients compared to controls. Interestingly, MCI patients progressing to AD showed elevated CSF TNFα [96]. Elevated serum TNF-α concentrations in AD patients compared to healthy individuals and individuals with MCI were subsequently demonstrated [15,17,97]. A 6-month study exploring the association between TNF-α levels and cognitive performance in 300 subjects with varying AD severity, found that acute inflammation correlated with a twofold cognitive decline increase, and high-baseline TNF-α levels quadrupled the decline [98]. Kim and colleagues performed an analysis of serum cytokine levels in individuals affected by AD, in those with MCI, and in a cohort of healthy controls to ascertain the connections between these cytokine levels and the neuropsychological parameters. Noteworthy associations between the levels of TNF-α and interleukin-6 (IL-6) and the cognitive performance, assessed using the Mini-Mental State Examination (MMSE) score were described [16]. Another interesting study explored the correlation between CSF TNF-α and functional connectivity in 64 older adults and found that elevated TNF-α levels correlated with reduced connectivity in decision-making, inhibitory control, and memory regions, with APOE4 status moderating this effect [99]. The time courses of levels of multiple plasma and CSF cytokines in patients with AD and age-matched control subjects were assessed in a study by Llano and colleagues, by measuring cytokine levels 7 times over 24 h in plasma and CSF using a lumbar catheter. The authors found that CSF levels of IL-1β, IL-2, IL-10, IL-12p70, granulocyte-macrophage colony-stimulating factor, interferon-γ, and TNF-α diverged over time, with higher levels in AD subjects compared to controls, with no difference in cytokine trajectories observed in plasma [100]. Another noteworthy investigation revealed that TNFR displayed more robust connections with total tau and p-tau in comparison to Aβ1-42, spanning healthy controls, MCI, and AD subjects. Within the longitudinal cohort, individuals with MCI who exhibited heightened levels of CSF TNFR1 and diminished levels of TNFR2 had an increased susceptibility to advancing to AD [101].

Moreover, extensive research has suggested that specific TNF-α gene polymorphisms contribute to an increased risk of AD. In a meta-analysis aiming to define the association of common TNF-α gene polymorphisms with the risk of AD, Di Bona et al. selected 17 studies and evaluated them with a model-free method approach, to comprehensively analyze the results of these case-control genetic association studies. Notably, their research indicates a correlation between the −850 C > T polymorphism and the susceptibility to AD [102]. Additionally, Yang et al. performed a case-control study in the Southern Chinese population, involving the use of polymerase chain reaction-sequence specific primers (PCR-SSP) to assess TNF-α genotypes and alleles in 112 sporadic AD patients and 121 controls. Additionally, they quantified serum TNF-alpha levels through radioimmunoassay. They found significantly higher levels of serum TNF-α in sporadic AD patients and that both −308 A/G polymorphism and elevation of serum level of TNF-α were both associated with an increased risk of AD [103].

Observational investigations conducted on individuals with systemic inflammatory disorders treated with TNF-α inhibitors offer additional substantiation for the implication of TNF-α in the pathogenesis of AD. A large retrospective case-control analysis, encompassing 56 million adult patients afflicted with inflammatory conditions, revealed a diminished risk of AD development with the administration of TNF-α inhibitors [93]. A previous nested case-control study revealed a higher AD prevalence among RA patients compared to those without rheumatoid arthritis, with a further significant increase in risk in those who were affected by chronic conditions including coronary artery disease, diabetes, and peripheral vascular disease. Notably, exposure to anti-TNF agents, particularly etanercept, was associated with a reduced risk of AD in these patients [104]. Elderly rheumatoid arthritis patients showed improved cognitive performance with subcutaneous anti-TNF-α therapy using drugs like etanercept and adalimumab [105]. In contrast, a double-blind study in mild and moderate AD patients treated with subcutaneous etanercept did not show significant changes in cognitive function, behavior, and global functions though there was a positive trend in the anti-TNF-α treatment group [106]. The limited effectiveness of subcutaneous anti-TNF-α therapies in AD patients may be attributed to the large molecular weight of anti-TNF-α monoclonal antibodies because this makes the passage through the blood–brain barrier impossible under physiological conditions [82].

Interesting clinical studies collectively suggested that perispinal administration of etanercept may lead to significant and sustained improvements in AD and primary progressive aphasia, showcasing its potential as a therapeutic intervention. In a prospective, single-center, open-label, pilot study, involving 15 patients with AD and primary progressive aphasia, perispinal etanercept infusion (25–50 mg weekly for six months) demonstrated significant improvement in cognition [107]. The benefits of this treatment were sustained in patients who continued for over two years [108]. Furthermore, another case report describing an 81-year-old patient who was given 25 mg of perispinal etanercept by posterior cervical interspinous injection outlined swift cognitive enhancement, starting within minutes. The rapid cognitive improvement observed in this patient has been hypothesized to be linked to the mitigation of the effects of excess TNF-α on gliotransmission or other synaptic mechanisms in AD [109,110]. Due to the limitations of large molecules like etanercept in crossing the blood–brain barrier with traditional systemic administration, the perispinal direct drug delivery system likely played a role in the treatment’s efficacy [110].

Studies exploring the efficacy of anti-TNF-α drugs on the progression of AD are summarized in Table 1. To the best of our knowledge, there is currently no ongoing study that can adequately elucidate the potential role of these drugs in AD. Further research and larger-scale studies are warranted to establish the efficacy of TNF-α inhibition in AD conclusively.