A plethora of medical conditions are associated with alterations in the gut microbiome, invariably involving decreases in the short-chain fatty acid, butyrate [42]. As noted, butyrate is primarily induced during fasting, including when asleep [40]. Pineal melatonin effects include the suppression of gut permeability and associated decrease in gut dysbiosis, concurrently elevating butyrate levels [43]. Melatonin is also very highly produced in the gut, especially after feeding where is seems to increase the swarming of gut bacteria in the presence of food [44, 45]. The gut microbiome and the timing of food intake are integral aspects of the circadian rhythm.

The benefits of butyrate are mediated via a number of processes, including: (1) activation of the G-protein coupled receptors (GPR), GPR-41, -43 and -109 [46]; (2) via histone deacetylase inhibition (HDACi) thereby allowing butyrate to epigenetically regulate systemic and CNS cells [47]; (3) by optimizing mitochondrial function, involving the upregulation of sirtuin-3 and the deacetylation and disinhibition of PDC, thereby increasing ATP from OXPHOS and the TCA cycle, coupled to decreased mitochondrial oxidant production [48, 49]. As the upregulation of the mitochondrial melatonergic pathway is intimately linked to mitochondrial optimization, as shown in intestinal epithelial cells [50], butyrate upregulates the mitochondrial melatonergic pathway [50]. Factors influencing the tryptophan-melatonin pathway in a given cell will therefore have consequence for how the gut microbiome regulates that cell.

Stress/GR activation and proinflammatory cytokines can suppress pineal melatonin [51], whilst increasing gut permeability/dysbiosis, thereby lowering butyrate levels [52, 53]. Such data highlights the dynamic two-way interactions of systemic processes and CNS processes over the circadian rhythm (see Fig. 1). Variations in pineal melatonin, local tryptophan-melatonin pathway regulation and gut microbiome-derived butyrate also interact to modulate the consequence of HPA axis activation, including in the course of the cortisol awakening response (CAR).

The HPA axis has been extensively investigated following its induction by stress, including in Alzheimer’s disease pathoetiology, where heightened levels of circulating cortisol are evident, as shown in a recent meta-analysis [54]. However, although of unknown physiological relevance, morning CAR does not seem significantly altered in Alzheimer’s disease patients [54]. CAR drives a large surge in cortisol release starting just before the end of sleep and lasting for the first 30 minutes following awakening. The role of CAR is generally placed in the vague context of ‘preparing the body for the coming day’, with hope that its relevance may be better clarified by more rigorous methodology [55]. The effects of CAR, and indeed stress driven HPA axis activation, is considerably complicated by the diverse ways that glucocorticoid receptor (GR) can influence cell processes. The GR is the main mediator of HPA axis and CAR effects, with the GR being predominantly expressed in the cytoplasm in a complex with heat shock protein (hsp)90 and p23. Upon activation, the GR is transported to the nucleus where it induces genes containing the glucocorticoid receptor element (GRE) in their promotor.

The availability of pineal and/or local melatonin is relevant to CAR effects, with melatonin preventing GR nuclear translocation [56], including possibly via the upregulation of bcl2-associated athanogene (BAG)1 [41]. Butyrate, via its capacity as a HDACi, also suppresses GR nuclear translocation, involving increased acetylation of the GR and hsp90 [57, 58]. Consequently, variations in pineal and local melatonin as well as butyrate availability over the course of sleep not only regulate inflammation, antioxidant status and mitochondrial function, but also have direct impacts on GR nuclear effects during CAR. This may be of some importance given the dramatic effects that cortisol can have on the function of all immune and glial cells, in contrast to melatonin and butyrate (see Table 1 in [41]), and therefore on patterned immune/glia responses for the coming day. This would indicate that although CAR level/slope may not be significantly different in Alzheimer’s disease [54], the dramatic decrease in pineal melatonin and gut microbiome-derived butyrate will significantly regulate CAR consequences in Alzheimer’s disease. As the different cells in a given local microenvironment, including the tumor microenvironment [59], may have their tryptophan-melatonin pathway differentially regulated, GR activation can have distinct effects in the cells of a given microenvironment. This alters the nature of the homeostatic interactions that occur in the course of CAR ‘preparing the body for the coming day’. Such homeostatic alterations are proposed to contribute to the pathoetiology of ‘autoimmune’/‘immune-mediated’ disorders [33], which recent data indicates to include Alzheimer’s disease [60].

Notably, this is complicated by the diverse effects of GR activation. The GR has a number of genomic and non-genomic effects, including via the induction of intracellular signaling pathways following plasma membrane GR activation [61]. The GR may also interact with other transcription factors in the nucleus to modulate diverse genes with no GRE in their promoter. GR effects, both in CAR and stress/HPA axis activation, may also be further complicated by factors upregulating BAG-1, as recently proposed for melatonin [62]. BAG-1 can not only prevent nuclear translocation but can also chaperone the GR to mitochondria [63], where it can be translocated over the inner and outer mitochondrial membranes (by mitochondrial import inner membrane translocase subunit (TIM) 23 and mitochondrial import outer receptor subunit (TOM) 20) allowing the GR to interact with PDC and hsp60, with consequences for mitochondrial metabolism, including the regulation of the mitochondrial melatonergic pathway [41].

Importantly, the consequences of CAR/GR, butyrate and pineal/local melatonin interactions are crucially determined by how they interact to regulate mitochondrial function. This involves both the dampening of residual inflammation and mitochondrial oxidant production, as well as how CAR may prime systemic and CNS cells for the coming day. It is widely recognized that suboptimal mitochondrial function is an integral aspect of dementia, both systemically [64] and in brain cells [65, 66], with mitochondrial dysfunction evident in many of dementia’s currently conceptualized ‘comorbidities’, such as obesity [67], T2DM [68] and depression [69]. Mitochondrial function may be especially relevant in reactive cells, namely immune and glial cells, including astrocytes, which have a powerful role in the regulation of neuronal activity and survival [69].

There is a growing appreciation of the importance of astrocytes in the regulation of a wide array of neuropsychiatric and neurodegenerative conditions that have been classically conceptualized as ‘neuronal’ disorders, including depression [69], Parkinson’s disease [70], amyotrophic lateral sclerosis (ALS) [14], and multiple sclerosis [71]. The role of astrocytes in such conditions is attributed to a wide array of processes including by astrocyte AhR nuclear translocator (ARNT)-like 1 and BAG3 determining levels of $\alpha$-synuclein and hyperphosphorylated tau, implicating the circadian rhythm and BAG regulation in the modulation of neurodegenerative processes [3]. It is proposed here that dysregulation of the astrocyte tryptophan-melatonin pathway is an early change in the pathoetiology of Alzheimer’s disease.

Astrocyte Mitochondrial Tryptophan-Melatonin Pathway

TLR2/4 can be activated by numerous ligands, including LPS, hsp60, hsp90, amyloid-$\beta$, $\alpha$-synuclein, fibrinogen and high-mobility group box (HMGB)1. Astrocyte TLR2/4 signaling increases NF-kB and YY1, thereby increasing BACE1 and amyloid-$\beta$ production [72, 73]. This is parsimonious with amyloid-$\beta$ as an endogenous antimicrobial [9]. As both NF-kB and YY1 upregulate the melatonergic pathway, as shown in other cell types [35, 36], the concurrent/sequential release of melatonin following TLR2/4 activation will have autocrine and paracrine effects that dampen and time-limit inflammatory responses and raised oxidants. As cytokines and oxidants can induce further TLR2/4 ligands as well as the nucleotide-binding domain, leucine-rich–containing family, pyrin domain–containing-3 (NLRP3) inflammasome [74], the autocrine and intracrine effects of astrocyte melatonin will prevent prolonged signaling via TLR2/4. Melatonin suppresses the NF-kB and reactive oxygen species (ROS)-driven NLRP3 inflammasome [74], which may be partly mediated over the circadian rhythm by pineal melatonin upregulating the $\alpha$7 nicotinic acetylcholine receptor ($\alpha$7nAChR) [75], which suppresses immune inflammation and is a significant treatment target in Alzheimer’s disease [76]. As to whether local paracrine melatonin release from astrocytes regulates local $\alpha$7nAChR levels will be important to determine. Such data highlights the potentially important role of pineal and local astrocyte melatonin in the regulation of an Alzheimer’s disease pathophysiological factor (NLRP3) [77] and protective factor ($\alpha$7nAChR) [78]. As stress/GR activation can regulate $\alpha$7nAChR [79], the interactions of pineal melatonin’s circadian induction of the $\alpha$7nAChR [80] with the wider night-time changes in morning CAR modulation will be important to determine. Notably, the $\alpha$7nAChR is also expressed on the mitochondrial outer membrane where its binding by agonists suppresses Ca${}^{2+}$ influx and decreases cytochrome c release [81], suggesting that it may linked to optimizing mitochondrial resilience under challenge [82] as well as wider cellular and mitochondrial plasticity.

Astrocyte melatonin release has been long established [83], being regulated by the main Alzheimer’s disease susceptibility gene, apolipoprotein (Apo)E4 [84]. ApoE4 is the major susceptibility gene for Alzheimer’s disease, with carriers of two ApoE4 alleles having an 8-15-fold increase in Alzheimer’s disease susceptibility [85]. ApoE is predominantly expressed in astrocytes in the brain, where it regulates lipid metabolism, with ApoE4 increasing the unsaturated fatty acid chains on triglycerides [86]. Whether the ApoE4 upregulation of astrocyte mitochondrial melatonergic pathway is relevant to neuronal loss in dementia will be important to determine, including whether there is any role for a heightened dependence of astrocyte melatonin in ApoE4 carriers. ApoE4 effects include the downregulation of astrocyte monoamine oxidase (MAO)-A and MAO-B, thereby increasing the availability of serotonin as a precursor for the melatonergic pathway [85]. This is one means by which astrocyte ApoE4 can increase astrocyte melatonin production [85], thereby contributing to a role for enhanced astrocyte melatonin in the maintenance of homeostatic interactions of astrocytes with neurons and other brain cells. The suppressed capacity to induce astrocyte melatonin over aging may therefore have more significant impacts on ApoE4 carriers, thereby indicating a more significant role for the greater drop in local melatonin dampening of local inflammation in ApoE4 carriers.

As noted, heightened amyloid-$\beta$ levels are evident in other classical neurodegenerative conditions, including glioblastoma [10], Parkinson’s disease with Lewy bodies [13], tauopathies [87] and ALS [14]. The pathophysiologies of all of these conditions implicate a significant astrocyte role [14, 88, 89, 90], including via regulation by ApoE alleles [91, 92, 93, 94]. Clearly, it will be important to clarify the nature of astrocyte ApoE in the regulation of the astrocyte tryptophan-melatonin pathway over aging in the pathoetiology of dementia and associated neurodegenerative conditions.

The loss of serotonergic neurons is intimately linked to the suppression of the astrocyte tryptophan-melatonin pathway in dementia, with serotonergic neuronal loss being a relatively early event in Alzheimer’s disease [95]. Decreased serotonin levels are also evident in Alzheimer’s disease platelets, with platelet serotonin inversely correlating with cerebrospinal fluid (CSF) hyperphosphorylated-tau and amyloid-$\beta$ [96]. Such systemic suppression of serotonin availability may be linked to decreased gut microbiome-derived butyrate, which downregulates MAO-B, thereby increasing serotonin availability [97]. Butyrate’s suppression of MAO-B, thereby enhancing serotonin availability may be another route whereby butyrate upregulates the melatonergic pathway [50] in astrocytes and other cells as well as modulating the capacity of pineal melatonin at night to also upregulate the astrocyte melatonergic pathway. The regulation of serotonin availability may therefore be another aspect of gut-pineal interactions in the modulation of morning CAR activation of the GR. Inflammation driven indoleamine 2,3-dioxygenase (IDO) and GR induced tryptophan 2,3-dioxygenase (TDO), by converting tryptophan to kynurenine decrease tryptophan availability and thereby will also suppress serotonin availability.

Whether the role of autocrine melatonin in switching from an inflammatory M1-like phenotype in macrophages [35] and microglia [80] to a quiescent, pro-phagocytic M2-like phenotype is paralleled in astrocytes will be important to determine. This would seem not unlikely, given the role of exogenous melatonin in dampening astrocyte reactivity [98, 99]. The maintenance of astrocyte reactivity is an important driver of microglia reactivity, which is also dampened by autocrine and paracrine melatonin [100]. Such data highlights the importance of maintained astrocyte reactivity in regulation of the wider CNS inflammatory activity in dementia, as well as the role of astrocyte TLR2-4/NF-kB/YY1/BACE1/amyloid-$\beta$ in driving the maintained amyloid-$\beta$ production that further contributes to inflammation and neuronal loss in the course of dementia. However, although previous data is highly suggestive of a role of such processes, clearly the astrocyte mitochondrial melatonergic pathway requires investigation. The proposed sequential release of melatonin following TLR2/TLR4/NF-kB/YY1/BACE1/amyloid-$\beta$ pathway in astrocytes (as well as microglia and neurons) will dampen the maintained heightened inflammation that is widely thought to underpin neuronal loss in dementia and across other neurogenerative conditions, as indicated by the effects of exogenous melatonin across diverse preclinical models [101, 102, 103]. Such studies also indicate that in the course of dampening inflammatory signaling melatonin will have autocrine and paracrine effects that optimize mitochondrial function, thereby acting on core process that underpin the complex and dynamic flurry of cell fluxes that are typical of the chaos of ‘endpoint’ disorders, such as the current classification of later stage Alzheimer’s disease. The beneficial effects of pineal melatonin and gut microbiome-derived butyrate may then be intimately linked to their night-time capacity to induce the astrocyte mitochondrial melatonergic pathway, thereby optimizing astrocytes in their regulation of neuronal activity and survival. This would suggest that dementia, like cancer [26], may be importantly determined by night-time, sleep linked processes, upon which the morning CAR activation of the GR will act to regulate local microenvironment homeostasis. See Fig. 2.

Fig. 2.

Shows how changes over aging suppress pineal melatonin and gut microbiome-derived butyrate as well as platelet and neuronal serotonin (yellow shade), which all contribute to suppress astrocyte (gold shade) melatonin and NAS availability. This contributes to astrocyte mitochondrial dysfunction, with enhanced astrocyte reactivity increasing microglia reactivity, thereby contributing to maintained, raised levels of HMGB1 and LPS activation of TLR2/4 and increased NLRP3 inflammasome induced IL-1$\beta$ and IL-18, further contributing to the inflammatory milieu that enhance BACE1/amyloid-$\beta$ and associated increase in hyperphosphorylated tau. The emergent plaques and tangles in the ‘end-point chaos’ of later stage Alzheimer’s disease may therefore be powerfully determined by factors, including systemic, that regulate the availability of the astrocyte tryptophan-melatonin pathway. Abbreviations: 5-HT, serotonin; acetyl-CoA, acetyl-coenzyme A; AANAT, aralkylamine N-acetyltransferase; AhR, aryl hydrocarbon receptor; BACE1, $\beta$-site amyloid precursor protein-cleaving enzyme 1; CRH, corticotrophin-releasing hormone; GR, glucocorticoid receptor; HMGB1, high-mobility group box; Hyperphos, hyperphosphorylated; LPS, lipopolysaccharide; MAO-B, monoamine oxidase-B; Mito, mitochondrial; NAS, N-acetylserotonin; NF-kB, nuclear factor kappa-light-chain-enhancer of activated B cells; NLRP3, nucleotide-binding domain, leucine-rich–containing family, pyrin domain–containing-3; PDC, pyruvate dehydrogenase complex; TLR, toll-like receptor; YY1, yin yang 1.

Raised aryl hydrocarbon receptor (AhR) levels and activation are closely linked to aging [104], including via disruption to the circadian rhythm [105]. As highlighted in Fig. 1, AhR activation induces cytochrome P450 (CYP)1A2 and CYP1B1, thereby O-demethylating melatonin to NAS as well as increasing melatonin hydroxylation [106, 107]. The AhR is typically held in a cytoplasmic complex with hsp90 and when activated by an AhR ligand translocates to the nucleus where it forms a dimer with the AhR nuclear translocator (ARNT), thereafter inducing genes with the xenobiotic response element (XRE) in their promotor. The AhR can also be expressed on the mitochondrial membrane where it regulates Ca${}^{2+}$ influx via the voltage dependent anion channel (VDAC)1 [108] and interacts with the mitochondria-located translocator protein kDa18 (TSPO) [109]. The diverse and sometimes contrasting effects of the AhR may be partly determined by site of translocation as well as by other alternatively sited mitochondria receptors, such as the $\alpha$7nAChR, TrkB and GR [110].

The AhR has numerous endogenous and exogenous ligands, including kynurenine derived from pro-inflammatory cytokine induction of indoleamine 2,3-dioxygenase (IDO) and cortisol/GR induction of tryptophan 2,3-dioxygenase (TDO). The array of AhR ligands, site of expression, and the AhR regulation of the mitochondrial melatonergic pathway allows AhR activation to have a complexity of consequences. Such plasticity of AhR activation considerably complicates its role in aging [111], whilst also highlighting the adaptability of the AhR in the regulation of core cellular processes as determined over the course of evolution. As the conversion of tryptophan to kynurenine by IDO and TDO also increases other kynurenine pathway products, including neuroregulatory kynurenic acid and quinolinic acid, AhR activation by kynurenine will be coupled to concomitant AhR-independent effects of kynurenine pathway products on neuronal activity and interarea communication across the brain. Given that 60% of brain kynurenine is derived from the periphery [112], systemic processes that drive large kynurenine increases, such as from white adipocyte (WAT) in obesity [113], can drive significant changes in the CNS relevant to dementia pathophysiology, including depriving tryptophan for the tryptophan-melatonin pathway, and kynurenic acid/quinolinic acid neuroregulation as well as AhR activation.

By O-demethylating melatonin to NAS, thereby increasing TrkB activation, the AhR is a significant contributor to Alzheimer’s disease pathophysiology. As noted, NAS is a brain-derived neurotrophic factor (BDNF) mimic via its activation of the BDNF receptor, TrkB [114]. TrkB activation of the TrkB-FL is a treatment target for Alzheimer’s disease, given its trophic and metabolic benefits [115]. However, in the presence of amyloid-$\beta$ the truncated TrkB-T1 is markedly increased and significantly contributes to neuronal loss, as shown in Alzheimer’s disease preclinical models [116, 117]. By increasing the NAS/melatonin ratio and associated TrkB-T1 activation, the AhR and its ligands can contribute to Alzheimer’s disease pathophysiology by a number of mechanisms. The loss of pineal melatonin’s suppression of pro-inflammatory cytokines/IDO/kynurenine and cortisol/GR/TDO/kynurenine will contribute to heightened AhR activation, thereby increasing the NAS/melatonin ratio and heightened TrkB-T1 activation-linked neuronal loss, especially in the presence of raised amyloid-$\beta$ levels. The suppression of gut microbiome butyrate will also contribute to this via decreased HDACi, thereby increasing GR nuclear translocation and TDO induction. As noted, the marked suppression of pineal melatonin and gut derived butyrate will enhance GR nuclear translocation, thereby modulating not only the HPA axis stress response, but also the consequence of CAR as it ‘prepares the body for the coming day’. Such processes allow the AhR to be intimately linked to the circadian rhythm in the regulation of Alzheimer’s disease pathophysiology. See Fig. 1.

Interestingly, stress/GR activation of astrocytes induces MER proto-oncogene, tyrosine kinase (MERTK), thereby driving astrocytes to phagocytose excitatory synapses on to GABAergic neurons, which is mediated by the GR acting as a nuclear transcription factor [118]. MERTK is one of the TAM receptors (Tyro3, Axl, and Mertk), which are linked to immune regulation, cell differentiation and apoptotic cell/debris clearance [119], with MERTK alleles associated with Alzheimer’s disease risk, especially in females [120]. As to whether stress/GR-driven increase in astrocyte MERTK, by driving astrocyte phagocytosis of excitatory synapses onto GABAergic neurons, contributes to the increased glutamatergic excitotoxicity in Alzheimer’s disease [121] requires investigation. As heightened glutamatergic N-methyl-D-aspartate receptor (NMDAr) activation, including by amyloid-$\beta$, contributes to the raised TrkB-T1 levels in Alzheimer’s disease [122], the loss of melatonin and butyrate suppression of GR nuclear translocation will enhance GR-induced MERTK and GABAergic suppression and heightened glutamatergic NMDAr activation, thereby increasing TrkB-T1 levels. This indicates a role for circadian and systemic factors in the modulation of substantial changes in how astrocytes regulate neuronal interactions and activation. Notably, the GR can also increase BACE1 and amyloid-$\beta$ via plasma membrane GR activation [123] as well as by regulating presenilin (PSEN)1 assembly on the endoplasmic reticulum (ER), thereby inducing amyloid-$\beta$ accumulation on the ER mitochondrial associated membrane (MAM) [124]. The roles of melatonin and butyrate in the modulation of these other routes of BACE1 and amyloid-$\beta$ induction by stress/CAR activation of the GR have still to be determined.

Whether the loss of night-time pineal melatonin and gut microbiome-derived butyrate increases the morning CAR activation of astrocyte GR to not only increase BACE1 and amyloid-$\beta$ but also contribute to MERTK induction and heightened TrkB-T1 levels, especially in the presence of amyloid-$\beta$, will be interesting to determine. When this is coupled to an AhR-driven increase in the NAS/melatonin ratio, thereby enhancing NAS activation of TrkB-T1 (on the plasma and/or mitochondrial membranes), neuronal loss will be potentiated [116, 117]. Overall, the circadian loss of night-time melatonin and gut microbiome-derived butyrate will also be relevant to later stages of neuronal loss in dementia, where heightened amyloid-$\beta$ and TrkB-T1 levels are evident, as well as in the pathoetiology of dementia.

However, it is important to note the widespread mitochondrial dysfunction in dementia cannot be simply understood as a consequence of various fluxes and challenges, but rather mitochondrial function is a dynamic core aspect of intercellular homeostatic interactions, with the suppression of the mitochondrial melatonergic pathway a target for other cells in a given microenvironment when the products of a challenged cell drive a prolonged maladaptive dyshomeostasis. Such intercellular dyshomeostasis drives the pathoetiology of ‘autoimmune’/‘immune-mediated’ disorders [12, 33], where the regulation of the mitochondrial melatonergic pathway in a ‘targeted’ cell (such as in pancreatic $\beta$-cells in T1DM [12]) is dynamically shaped by other cells in the local microenvironment. This is most evident in the tumor microenvironment, where cancers release kynurenine to activate the AhR to shape the metabolic function of other tumor microenvironment cells in the course of generating ‘immune tolerance’ and shaping intercellular fluxes [59, 125].

The dynamic intercellular fluxes within a given microenvironment modulate processes that shape mitochondrial function, including ROS production, thereby driving ROS-dependent microRNAs (miRNAs) [126] that shape patterned gene expression in an ever-changing dynamic over the circadian rhythm. Whether such systemic and microenvironment determined mitochondrial ROS-driven miRNAs underpin the increased TrkB-T1, especially in the presence of amyloid-$\beta$, will be important to determine over the course of dementia progression. How this is coordinated with melatonergic pathway suppressing miRNAs, such as miR-7, miR-375 and miR-451 [127, 128, 129], may be of particular importance, given the use of streptozotocin (a known melatonergic pathway inhibitor [130]) in preclinical models of dementia [131] and autoimmune disorders [132]. Such suppression of mitochondrial melatonin would overlap neuronal loss in Alzheimer’s disease with that of Parkinson’s disease, where ‘autoimmune’/‘immune-mediated’ processes involving CD8${}^{+}$ t cell chemoattraction may be driven by decreased melatonin regulation of PINK1/parkin-mediated mitophagy [133] and associated upregulation of major histocompatibility complex (MHC)-1 [33], driving neuronal [134, 135] and possibly microglia [136] destruction. This is supported by data showing the presence of heightened CD8${}^{+}$ t cells levels in the Alzheimer’s disease brain in proximity to amyloid plaques [137, 138], which is strengthened further by data showing CD8${}^{+}$ t cells to drive neuronal loss in Alzheimer’s disease models [134].

Such data indicates that the suppression of the mitochondrial melatonergic pathway is a significant driver of dysregulated mitophagy via the loss of melatonin’s regulation of PINK1/parkin/mitophagy [133], arising from systemic (pineal melatonin, gut butyrate, CAR/GR, WAT kynurenine) changing the dynamic interactions in a given microenvironment where prolonged inflammation and microbial/viral-type signaling (TLR2/4/9) drives a dyshomeostasis that cannot be dealt with by the immediately responding immune cells (astrocytes and microglia), leading to the chemoattraction (via MHC-1) of CD8${}^{+}$ t cells that drives cell destruction (predominantly neurons but also glia). This has parallels to the course of acute viral infection, as evident in the COVID-19 pandemic [139], where the immune response is initially determined by ‘first responders’ (such as neutrophils, macrophages and mast cells) but is taken over by the immune system ‘second wave’ (CD8${}^{+}$ t cells and natural killer cells). Neuronal loss in Alzheimer’s disease therefore has significant ‘autoimmune’/‘immune-mediated’ overlaps with neuronal loss in the Parkinson’s disease substantia nigra pars compacta [33, 140].

Overall, the suppressed capacity to maintain the mitochondrial melatonergic pathway in astrocytes may be an important initiator of dementia pathoetiology with the subsequent protracted pathophysiology involving alterations in dynamic mitochondrial homeostatic interactions in a given microenvironment. Systemic and local microenvironment processes interact to suppress the mitochondrial melatonergic pathway, thereby initiating the chemoattraction of CD8${}^{+}$ t cells and the ‘autoimmune’/‘immune-mediated’ destruction of neurons that underpin the progressive cognitive loss in dementia. The above suggests changes in the nature of Alzheimer’s disease pathophysiology over time, including in the nature of immune involvement, rather than a static progression of the same pathophysiological processes. This may be more akin to concepts of ‘staging’ in neuropsychiatric conditions, such as bipolar disorder [141], likely in association with distinct treatments at different physiologically determined ‘stages’ [142]. Such ever-changing processes ultimately culminate in the ‘end-point chaos’ of the currently defined late-stage Alzheimer’s disease. See Fig. 3.

Fig. 3.

Shows how gut dysbiosis, gut permeability, pro-inflammatory cytokines, GR, AhR, NAS/melatonin ratio, and decreased pineal melatonin act primarily on astrocytes to decrease astrocyte melatonin release, thereby suppressing the anti-inflammatory and mitochondria optimizing effects of autocrine and paracrine melatonin. Consequently, astrocyte and neuronal BACE1 and amyloid-$\beta$ continue to be produced in association with raised tau hyperphosphorylation, thereby driving the endpoint ‘chaos’ of plaques and tangles that classically define Alzheimer’s disease. The decreased astrocyte mitochondrial melatonin at night is at least partly mediated via a decrease in pineal melatonin and gut microbiome-derived butyrate, although there may be more aging-associated factors that also suppress the astrocyte tryptophan-melatonin pathway (top right-hand, gold shade). The loss of pineal and astrocyte melatonin suppresses lactate dehydrogenase (LDH) and therefore lactate production by astrocytes, thereby depriving energy for neuronal mitochondria and decreasing neuronal mitochondria melatonergic pathway induction. The suppression of astrocyte melatonin maintains the NF-kB induction of YY1, thereby decreasing EAAT2, which, like the increased System X${}_c$${}^{-}$ and MERTK, enhances glutamatergic excitotoxicity that raises neuronal apoptotic susceptibility and contributes to spreading apoptotic susceptibility (gold shade). Gut derived LPS, as well as other TLR2/4/9 ligands (grey shade) induce NF-kB and YY1 to increase BACE1 and amyloid-$\beta$, which in the absence of melatonin leads to prolonged amyloid-$\beta$ production that exacerbates neuronal mitochondrial function. Decreased pineal melatonin and gut butyrate will heighten the GR-TDO, pro-inflammatory cytokine-IDO and white adipocyte (WAT) derived kynurenine to activate the AhR. By increasing NAS and suppressing melatonin, the AhR will contribute to neuronal susceptibility via oxidative stress induced TrkB-T1, which NAS activates. Oxidative stress and neuronal mitochondrial dysfunction also increase pro-BDNF and its ligand, p75${}^{\text{NTR}}$, which also contributes to neuronal vulnerability as well as dysregulating patterned immune responses. The neuronal mitochondrial dysfunction and suppressed melatonin decrease PINK1/parkin/mitophagy, and increase MHC-1, which chemoattracts CD8${}^{+}$ t cells that drive neuronal destruction (yellow shade). Such immune-mediated processes will be contributed to, if not initiated by, alterations in the homeostatic interactions in the neuronal microenvironment, which will be subject to epigenetic and genetic risk factors (yellow shade). Abbreviations: AhR, aryl hydrocarbon receptor; BACE1, $\beta$-site amyloid precursor protein-cleaving enzyme; BDNF, brain-derived neurotrophic factor; EAAT, excitatory amino acid transporter; GR, glucocorticoid receptor; hsp, heat shock protein; HSV1, herpes simplex virus; IDO, indoleamine 2,3-dioxygenase; LDH, lactate dehydrogenase; LPS, lipopolysaccharide; MERTK, MER proto-oncogene, tyrosine kinase; MHC-1, major histocompatibility complex-class 1; Mito, mitochondria; NAS, N-acetylserotonin; NF-kB, nuclear factor kappa-light-chain-enhancer of activated B cells; PINK1, PTEN-induced kinase 1; ROS, reactive oxygen species; TDO, tryptophan 2,3-dioxygenase; TLR, toll-like receptor; TrkB-T1, tyrosine kinase receptor B-truncated; WAT, white adipocyte; YY1, yin yang 1.

Whether local paracrine melatonin release from astrocytes regulates local $\alpha$7nAChR levels, thereby impacting on immune/glia inflammation and associated consequences on cognition will be important to determine.

Whether the suppressed capacity to induce astrocyte melatonin over aging has heightened consequences for ApoE4 carriers will be important to clarify.

There is a growing appreciation of the role of platelets, and their regulation by gut microbiome-circadian interactions, in the pathophysiology of an array of diverse medical conditions, including Alzheimer’s disease, amyotrophic lateral sclerosis, and cancer [169]. The role of gut microbiome-circadian interactions in the modulation of Alzheimer’s disease pathophysiology will be important to determine, including via the regulation of platelet serotonin as a precursor for the mitochondrial melatonergic pathway.

The importance of the gut microbiome in Alzheimer’s disease is indicated by clinical and preclinical data indicating that the increased risk of Alzheimer’s disease in the partners of Alzheimer’s disease patients may be mediated cohabitation linked similarities in their gut microbiomes [170].

Whether the role of autocrine melatonin in switching from an inflammatory M1-like phenotype in macrophages [35] and microglia [80] to a quiescent, pro-phagocytic M2-like phenotype is paralleled in astrocytes will be important to determine.

A number of processes may contribute to the increased glutamate/GABA ratio in Alzheimer’s disease, including MERTK phagocytosis of excitatory inputs to GABA neurons, YY1 suppression of EAAT2, and increased System X${}_C$${}^{-}$. It will be important to determine the relative influences of such processes, including heightened glutamatergic activity induction of pro-BDNF, TrkB-T1 and the p75${}^{\text{NTR}}$ in driving neuronal loss and Alzheimer’s disease pathophysiology.

Whether the loss of night-time pineal melatonin and gut microbiome-derived butyrate increases the morning CAR activation of astrocyte GR to not only increase BACE1 and amyloid-$\beta$ but also contribute to MERTK induction and heightened TrkB-T1 levels, perhaps especially in the presence of amyloid-$\beta$, will be important to determine.

The conversion of glucose to pyruvate in astrocytes favors the conversion of pyruvate to lactate, with relatively little pyruvate utilized by the astrocyte mitochondrial PDC, as evident in neurons under challenge, during neuronal activity and in conditions of hypoxia [151]. Whether this is dependent upon the availability of the astrocyte mitochondrial melatonergic pathway, including in the course of aging when the astrocyte mitochondrial melatonergic pathway may be suppressed will be important to determine.

Whether a suppressed astrocyte mitochondrial melatonergic pathway underpins System X${}_C$${}^{-}$ upregulation in a quest to regulate astrocyte antioxidant status by GSH provision will be important to determine. Whether the loss of astrocyte (and/or pineal) melatonin decreases astrocyte lactate dehydrogenase, thereby decreasing the conversion of pyruvate to lactate will be important to determine. As astrocyte lactate is converted to pyruvate in neurons, thereby enhancing neuronal mitochondrial function, acetyl-CoA and neuronal mitochondrial melatonergic pathway induction, the suppression of astrocyte melatonin may have dramatic consequences for neuronal function, as well as the loss of astrocyte and neuronal melatonin increasing levels of hyperphosphorylated tau. This is important for future research to determine.

The above systemic conceptualization of Alzheimer’s disease and its emphasis on astrocytes and the astrocyte melatonergic pathway as a crucial hub has a number of treatment implications that will be better clarified when the astrocyte melatonergic pathway is more extensively investigated. However, a number of treatment targets are readily apparent.

(1) The utilization of melatonin and monitoring of the gut microbiome to optimize butyrate production is likely to suppress Alzheimer’s disease pathoetiology for those genetically at risk and in the early stages of mild cognitive impairment. The utility of melatonin and butyrate is likely to wane as dementia progresses, although this has still to be systematically investigated.

(2) A number of factors may act to suppress the astrocyte tryptophan-melatonin pathway, including suppressed 14-3-3 isoforms, serotonin and acetyl-CoA. Whether the adjunctive use of tryptophan supplements with melatonin and butyrate affords any added protection will be important to determine, given their safety profiles and ready availability.

(3) A number of AhR inhibitors show utility in Alzheimer’s disease preclinical models, including green tea’s epigallocatechin gallate [171], curcumin [172] and resveratrol [173], with efficacy often being modelled on diverse physiological processes, such as the induction of sirtuin-1 [174], although all dampen inflammatory processes in astrocytes [172, 175, 176], whilst also inhibiting and regulating the AhR [177, 178, 179]. As well as helping to preserve melatonin levels via AhR inhibition, these nutriceuticals can increase the tryptophan-melatonin pathway via other mechanisms, including via the inhibition of MAO, thereby increasing serotonin availability [180, 181, 182]. Increasing serotonin availability from dorsal raphe neurons as well as platelets is likely to have utility under conditions when the AhR is relatively suppressed [169].

(4) Although requiring more technical development, the utilization of mesenchymal stem cell exosomes that target the astrocyte and/or pinealocyte tryptophan-melatonin pathway would allow a more precise treatment focus on key hubs in Alzheimer’s disease pathophysiology.

(5) It is important to mention that social processes and physical contact can have important physiological consequences that can modulate Alzheimer’s disease pathophysiology, as highlighted by the acceleration of dementia by loneliness, with effects at least partly mediated via the regulation of the HPA axis [183].

(6) Some of the physiological consequences of social interaction may be mediated via the upregulation of hypothalamic and amygdala oxytocin and the oxytocin suppression of stress/dysphoria associated CRH via hypothalamic and amygdala astrocytes [163, 164]. Recent work indicates that oxytocin intranasal administration has a number of benefits, including cognitive in Alzheimer’s disease patients [166].