AD, an irreversible, progressive neurodegenerative disorder, contributes to 60%–80% of dementia cases among the elderly (Alzheimer’s Association, 2015). It is a complex multifactorial disease that has a strong genetic component with more than 50 risk loci identified (Silva et al., 2019). Mutations in the genes APP, PSEN1, and PSEN2 encoding for amyloid precursor protein, presenilin 1 and presenilin 2, respectively, account for most of the early-onset AD, while mutated apolipoprotein E (APOE) gene is frequently associated with late-onset AD (Silva et al., 2019). Such genetic defects were initially linked to amyloidosis—a well-known histopathological hallmark of AD, the removal of which has been attempted as a treatment; unfortunately, such clinical trials have failed (Oxford et al., 2020). While this appears to preclude amyloidosis as a primary causal factor for AD, it does not diminish the importance of genetic factors in AD, and continued efforts to give biological meaning to genetic information may facilitate the identification of aetiological agent(s) and/or key immunopathological factor(s), which ultimately are necessary for therapeutic discovery.

AMD is clinically defined as progressive loss of central vision resulting from neuronal and non-neuronal degeneration in the macula, the central part of the retina that is responsible for the finest spatial, temporal, and spectral acuity. The mainstay of management is secondary preventative treatments: in particular, there is a widespread use of anti-vascular endothelial growth factor agents (anti-VEGF) to manage disease progression for neovascular “wet” AMD, but there is no currently approved therapy for the atrophic “dry” form of AMD. There are multiple risk factors for AMD, including advanced age, genetic polymorphisms, light iris color, smoking, and a high-fat diet (Lambert et al., 2016). The first two are the two strongest risk factors, with genetic polymorphisms of CFH and age-related maculopathy susceptibility 2 (ARMS2) accounting for more than 50% of the heritability of AMD (DeAngelis et al., 2017). Although both AD and AMD have a strong genetic component, none of the AD-associated genes are linked to AMD pathology, and vice versa (Kaarniranta et al., 2011). Given the complex multifactorial nature of AMD, the precise aetiological sequence remains elusive.

The concept of an infectious aetiology of AMD largely stems from direct evidence of a serological association between microbes and AMD, indirect evidence that links microbe-associated diseases to AMD, or evidence that implicates microbe-mediated inflammatory responses to AMD pathogenesis. Three case-control studies separately established a significant serological association between wet and/or dry AMD and C. pneumoniae infection (Ishida et al., 2003; Kalayoglu et al., 2003; Shen et al., 2009), an infectious pathogen that has emerged as a risk factor for common non-infectious diseases, including AD and cardiovascular disease. AMD patients with high antibody titers of C. pneumoniae were also found to have a 2- to 3-fold increased risk of disease progression (Robman et al., 2005). Another case-control study, however, found a significant association between CMV infection and both forms of AMD, but not infection with C. pneumoniae and H. pylori (Miller et al., 2004), while four other studies reported no significant association between AMD and C. pneumoniae or Mycoplasma pneumoniae (Klein et al., 2005; Haas et al., 2009; Turgut et al., 2010; Khandhadia et al., 2012). The inconsistent findings might have been attributable to variations of types and stages of AMD cases included in the studies (Chen et al., 2014), or underlying variations in other factors contributing to the aetiology of AMD.

While activation of the complement system is central to controlling microbial infection, it also is a well-recognised player in AMD pathogenesis. This is demonstrated by the identification of complement signalling protein constituents in drusen—a pathological hallmark of AMD, the accumulation of which disrupts retinal homeostasis supported by retinal pigment epithelium (RPE; Anderson et al., 2010; Kawa et al., 2014; Weber et al., 2014; McHarg et al., 2015). Second, genetic polymorphisms of complement pathway-inhibiting genes contribute to an increased risk of AMD (Lambert et al., 2016). Of particular interest, the membrane cofactor protein CD46, which was found to be downregulated in the RPE of early geographic AMD, is an HHV-6A-specific receptor (Vogt et al., 2011). In multiple sclerosis, HHV-6A infection in astrocytes leads to CD46 downregulation, resulting in hyperactivation of the complement system that is damaging to the local tissue (Pinter et al., 2000). It, therefore, has been speculated that HHV-6A infection may trigger the pathological events that eventually lead to AMD development and progression (Fierz, 2017).

During BM, the first line of innate immune defense involves activation of the host classical complement pathway to label the invading pathogen for eradication by immune cells, with subsequent inhibition of the alternative complement pathway by complement receptor 1, CFH, and complement protease complement factor I to prevent an excessive, tissue-damaging immune response. Activation of the classical complement pathway is induced by C-reactive protein (CRP), upregulation of which is common to BM (Prasad et al., 2005) and both AD (O’Bryant et al., 2010) and AMD (Molins et al., 2018). Furthermore, genetic polymorphism of the non-coding CFH gene has been shown to associate with reduced CFH level in the cerebrospinal fluid and increased mortality in both clinical and experimental BM (Kasanmoentalib et al., 2019). This major allele (G) rs6677604 has also been described as a risk factor for AMD (Ansari et al., 2013). While there is no genetic linkage of this allele to AD, another risk allele, rs1061170, is common to both AD and AMD (Zhang et al., 2016), and molecular analyses of drusen (AMD) and senile plaques (AD) reveal the presence of common complement components, including C3, C5, C6-9, and factors B, H, and I (Sivak, 2013), implying the occurrence of a common inflammatory response in both diseases, which also shares similarities to complement activation during BM (Molins et al., 2018).

Together with complement system activation during BM, the toll-like receptors (TLRs) and NOD-like receptors (NLRs) expressed on or within CNS antigen-presenting cells may become activated by bacterial ligands, triggering the production of various cytokines and chemokines (the “mediators”) that facilitate recruitment of immune cells (the “effectors”) to the central infectious loci to eradicate the bacteria (Mook-Kanamori et al., 2011). At high bacterial infectious doses in the CNS, the “effectors” that also carry the “sensors,” for example, neutrophils expressing TLR2, relentlessly elicit an immune response, tipping the balance of pro-and anti-inflammatory responses towards the former, resulting in an exaggerated immune response and ensuing cytokine storm (Mook-Kanamori et al., 2011, 2012). Similarly, in AD, over-activated microglia contribute to heightened production of inflammatory cytokines, triggering a positive feedback loop; genetic polymorphisms of several cytokines, such as interleukin (IL)-1α, IL-1β, IL-6, and tumor necrosis factor (TNF), modify the risk of AD development and progression in certain populations (Su et al., 2016). In AMD, parainflammation is proposed to be dysregulated when an inciting insult pushes the response beyond a threshold that can be coped with by the normal host autonomous response for cellular repair during aging. This results in excessive activation of resident and recruited immune cells and a subsequent cytokine storm (Chen and Xu, 2015). Since certain immune cells release reactive oxygen species as part of the pathogen-killing process, the excessive production of cytokines and associated infiltration of activated immune cells into affected tissues contributes to heightened oxidative stress. This is postulated to be a synapse/neuron-damaging phenomenon and is observed in both BM and AD/AMD (Bonda et al., 2010; Barichello et al., 2013; Chen and Xu, 2015).

During infection, the complement system is activated to combat intruding pathogens. This represents the first line of defense by host innate immunity during BM (Prasad et al., 2005). Similarly, the complement system is known to play a key role in AD (McGeer and McGeer, 2002; O’Bryant et al., 2010) and AMD (Molins et al., 2018) disease pathogenesis. However, there remains no clear answer about the trigger(s) of complement activation in these diseases. The first theory of AD/AMD aetiology posits the accumulation of endogenous waste products, i.e., amyloid (AD)/drusen (AMD), as the potential trigger. Alternatively, it has been suggested that the presence of infectious agents activates the complement system. BM and AD/AMD share other common inflammatory events (Table 1) that end with pathological inflammation and oxidative stress, resulting in local tissue damage and loss of function. Activated microglia, macrophages, monocytes, and T cells are some of the common cellular immune mediators responsible for disease pathogenesis in BM and AD/AMD, while upregulation of IL-1β and CC chemokine ligand 2 has also been demonstrated in all three diseases (Kaarniranta and Salminen, 2009; Coutinho et al., 2013; Chen and Xu, 2015; Domingues et al., 2017). Furthermore, researchers in the different disease disciplines (BM, AD, or AMD) have separately looked at the role of the proinflammatory cytokine, interferon-gamma (IFN-γ), in disease pathogenesis. It has been proposed that targeting IFN-γ might have therapeutic potential in all three conditions. In our studies of experimental pneumococcal meningitis, we identified an important role of IFN-γ in mediating host immune responses that link to enduring neurological impairments in mice that survived following antibiotic treatment (Too et al., 2014). We found that IFN-γ-deficient mice treated with the antibiotic ceftriaxone survived pneumococcal meningitis with reduced cognitive and behavioral disorders compared to their wild-type counterparts and that the nexus between the toll-like receptors (TLRs) 2 and 4, IFN-γ and the enzyme indoleamine 2,3-dioxygenase-1 contributed, at least in part, to the neurological sequelae resulting from pneumococcal meningitis (Too et al., 2014; Too and Mitchell, 2021). Similarly, neutralising IFN-γ in a transgenic mouse model of AD was found to ameliorate behavioral deficits and amyloid plaque burden (Browne et al., 2013). In the context of AMD, analysis of serum IFN-γ in AMD patients has given mixed results (Afarid et al., 2019; Litwinska et al., 2019). Interestingly, though, in human RPE-derived ARPE-19 cells, recent evidence suggests that this cytokine induces expression of BRAF-activated non-coding RNA (BANCR), a regulatory transcript involved in immunopathological processes (Kutty et al., 2018); inhibition of IFN-γ has therefore been postulated as a potential therapeutic option in AMD.

Despite there being no direct serological evidence that links causative pathogens of BM to AD/AMD, a decreased cerebrospinal fluid level of Aβ42—a biological phenomenon of AD—was found in patients with acute purulent BM (Sjogren et al., 2001). Moreover, the disease-causing immunopathological pathways seen during the acute phase of BM partially resemble the active disease stage of AD/AMD when host inflammatory responses tip towards detrimental effects (Table 1). While correlations between BM and age-related diseases have rarely been investigated, it is worth mentioning that the lifelong cognitive and behavioral disorders among young BM survivors may predispose them to reduced cognitive reserve during later life, which increases the risk of AD (Stern, 2012). Furthermore, in the context of BM, although antibiotic treatment substantially reduces the mortality rate, survivors often exhibit neurological disorders upon recovery. The process of bacterial clearance by antibiotics, and bacterial autolysis, release immunoactive bacterial components that excessively amplify the host immune response and thereby cause permanent damage to host tissue (Brouwer et al., 2010). In this regard, effective treatment for BM can conceivably be achieved by optimally suppressing key pathological immune processes without compromising bacterial clearance. The search for an adjunctive treatment that mitigates the augmented inflammatory response, to be administered alongside a non-bacteriolytic antibiotic, may be a promising therapeutic direction (Bewersdorf et al., 2018).

In light of these data, two fresh insights can be taken from BM research. First, an immunomodulatory approach may not work to ameliorate AD/AMD if microbial agents remain in the host system (e.g., in the case of dysbiotic microbiota), since it may affect the host’s innate immunity to combat pathogens. Second, broad-spectrum antibiotic treatment may not be feasible if the source of inflammation is unknown. For instance, attempting treatment with a bacteriolytic antibiotic in AD/AMD patients who may have been infected with pathogens that are prone to release immunoactive components when lysed, might well contribute to an undesirable outcome. Moreover, the administration of broad-spectrum antibiotics may unfavorably alter the oral/gut microbiome. Several clinical trials have tested the efficacy of antibiotics in preventing or ameliorating cognitive impairments in AD. The findings remain inconsistent. For instance, antibiotic treatment with doxycycline and rifampin of patients with mild to moderate AD, whose medical history of infections was unclear, demonstrated protection against cognitive decline in a 2004 clinical trial (Loeb et al., 2004), but not in a 2013 trial (Molloy et al., 2013). Not identifying the source of inflammation may have contributed to the inconsistent findings seen in these clinical trials (Loeb et al., 2004; Molloy et al., 2013). An immunomodulatory therapy as an adjunct to antibiotic treatment may be useful to alleviate inflammation-driven catastrophic events without compromising host-pathogen clearance.

Our understanding of any parainfectious aetiology of age-related neurodegenerative diseases remains in its infancy. The mechanisms of AD and AMD are multifactorial, as a consequence of the range of phenotypes and the stereotypical nature of pathological responses to disease. These conditions may each represent a final common pathway to a variety of disease-causing processes which are clinically grouped as single entities. To advance AD and AMD research, we propose that several issues will need to be addressed: