Abstract
Aging societies have high incidence rates of Alzheimer’s disease (AD). AD is diagnosed at later disease stages and has a poor prognosis, and effective drugs and treatments for AD are lacking. The molecular mechanism of AD is not clear, and current research focuses primarily on amyloid-β (Aβ) deposition and tau protein hyperphosphorylation. Aβ deposition is the most frequently hypothesized initiating factor of AD, and Aβ clearance during the pathogenesis of AD may be an optional strategy to suppress AD development. Monocytes play important roles in the peripheral clearance of Aβ. Therefore, the present review summarizes our current knowledge of the potential roles of infiltrating macrophages, circulating monocytes, and Kupffer cells in the peripheral clearance of Aβ in AD.
Keywords
INTRODUCTION
Alzheimer’s disease and Aβ theory
Alzheimer’s disease (AD) is the most common neurodegenerative disorder; it leads to dementia and is responsible for approximately 80% of dementia cases [1]. An epidemiological investigation revealed that AD incidence rates were approximately 3–7% in different countries [2]. More than 47 million people worldwide currently live with dementia, and this number is projected to increase to greater than 131 million by 2050 with the aging population. Dementia also has a huge economic impact on various countries worldwide [3], and the total estimated worldwide cost of dementia was US$818 billion in 2016 [4].
The accumulation of the toxic protein amyloid-β (Aβ) in the brain is the most frequently hypothesized initiating factor of AD; however, some drugs targeting Aβ have been unsuccessful in clinical trials, possibly because the intervention was given too late [5, 6]. The primary and underlying cause of Aβ plaque deposits is an imbalance between Aβ production and clearance [7]. Dysfunctional physiological Aβ clearance may cause an increase in cerebral Aβ levels, which plays an important role in AD [8, 9].
The production, transport, and clearance of Aβ
The amyloid-β protein precursor (AβPP) is a conservative type 1 integral membrane protein that is universally expressed and exhibits important physiological functions in vivo [10]. In hippocampal neurons, AβPP plays an important role in synaptogenesis, neurite branching, and axonal morphology [11, 12]. AβPP also acts as a signaling nexus to transduce information regarding a range of extracellular conditions into the intracellular space [13]. However, AβPP can be sequentially hydrolyzed by β-secretase and γ-secretase to generate Aβ peptides in both the central nervous system (CNS) and peripheral tissues [14–16]. Peripheral tissues, such as the cerebrovasculature [10], human lymphocytes, and mononuclear phagocytes, release Aβ in adherence-activated patterns [15, 16].
Aβ is transported reciprocally through the blood-brain barrier (BBB) between the brain and the periphery [17]. Low-density lipoprotein receptor-related protein-1 (LRP1) in the brain endothelium contributes to the brain efflux of Aβ [18, 19]. Transthyretin (TTR) and amylin promote the transport of brain Aβ to the blood via the induction of LRP1 subcellular translocation to the plasma membrane of the BBB endothelium [20, 21]. The influx of Aβ across the BBB is partially mediated by the receptor for advanced glycation end products and organic anion transporter polypeptide [22]. The concentration of brain Aβ is closely related to Aβ transport through the BBB and the peripheral Aβ concentration because there is a physiological balance between brain Aβ and peripheral Aβ [23].
There are various clearance mechanisms that remove brain Aβ and peripheral Aβ in vivo. First, the resident immune cells in the brain, microglia, play a primary role in brain Aβ clearance. Second, as the cells that mainly perform phagocytosis, circulating monocytes transmigrate across the BBB and accumulate in the brain during neuroinflammation. Aβ may be an immunomodulator and induce neuroinflammation, which was characterized by the activation and infiltration of peripheral monocytes in a mouse model of AD [24, 25]. Phagocytosis via monocytes and macrophages (hereafter, the term monocytes will be used to refer to monocytes and macrophages) is the predominant mechanism of peripheral clearance of cerebral Aβ in AD mouse models [26, 27]. Third, the physiological clearance of peripheral Aβ via circulating monocytes is an important mechanism for maintaining the balance between Aβ production and Aβ clearance; this mechanism promotes the removal of brain Aβ and effectively restricts the formation of Aβ plaques in AD mouse models [24]. Finally, the Kupffer cells (KCs) of the liver may also contribute to the clearance of peripheral Aβ [28]. In fact, these four types of cells have deficient Aβ phagocytic functions, which may be involved in the progression and/or development of AD.
Monocyte system and AD
The monocyte system, a critical component of innate and adaptive immunity, responds to pathogens, dead cells, and Aβ [29, 30]. The system includes circulating monocytes in the peripheral blood and macrophages residing in tissues. Both cell types are derived from premononuclear cells in the bone marrow [31]. Premononuclear cells transform into macrophages upon entering tissues from the blood, and they assume different names in different tissues [32], such as microglia cells in the CNS, KCs in the liver, dust cells in the lungs, and macrophages in the spleen and lymph nodes.
The understanding of the involvement of the immune system in AD, especially phagocytosis by immune cells in the clearance of Aβ in AD, has increased gradually. In earlier research, the brain was considered an immune-privileged area and was excluded from the immune surveillance system in grafting experiments in the cerebral cortices of rats and rabbits [33]; the presumed possible mechanisms were the isolating function of the BBB, the lack of fully organized drainage via lymphatic vessels and the absence of major histocompatibility complex [34]. The immune system was not considered relevant to AD because it was not thought of as a disease restricted to the CNS at that time. Later, studies showed that microglia are important immune macrophages and inflammation-related cells in the brain and are closely related to brain inflammation and the central clearance of Aβ. Patrick McGeer, Dennis Dickson and others studied the role of microglia and cytokines secreted by them in AD systematically [35–41]. Because AD is a systemic disorder, the general immune system, especially microglial cells and peripheral monocytes, may participate in the pathogenesis of AD via the clearance of Aβ, as shown in both an APP/PS1 AD mouse model and AD patients [42–44].
As resident phagocytes in the brain, microglia are closely related to the clearance of brain Aβ, and this has been reviewed previously in depth [26]. Peripheral clearance of Aβ plays an important role in AD. Therefore, the present review discusses and analyses the possible role of brain-infiltrating macrophages, circulating monocytes in peripheral blood, and liver KCs in AD pathophysiology.
INFILTRATING MACROPHAGES AND AD
Aβ/AβPP causes neuroinflammation and facilitates monocyte migration into the brain in AD
Aβ and the carboxyl-terminal fragment of the AD AβPP protein lead to markedly increased production of tumor necrosis factor-alpha (TNF-α) and matrix metalloproteinases–9 (MMP-9) in a monocytic leukemia cell line (THP-1) in the presence of interferon-γ (IFN-γ) [45]. Significant hypomethylation of the TNF-α promoter in monocytes was found in AD patients’ brains, and this hypomethylation increased TNF-α levels in the blood of AD patients [46]. TNF-α and MMP-9 may trigger inflammatory processes [45]. Monocytes co-expressing the inflammasome component Nod-like receptor family pyrin domain containing 3 (NLRP3) with caspase 1 or caspase 8 were significantly increased in patients with severe AD [47]. Monocytes produce significantly higher levels of the proinflammatory cytokines IL-1β and IL-18, and the NLRP1 and NLRP3 inflammasomes are activated in AD patients [47]. Some drugs, such as acetylcholinesterase inhibitors, directly inhibit cytokine release from microglia and monocytes and might exhibit anti-inflammatory effects in AD patients [48].
AβPP may be a proinflammatory receptor on microglia that regulates the ability of these cells to acquire a proinflammatory phenotype in mouse models of AD [49]. Microglial TLR4 signaling was altered in an AD mouse model [41], and this change may contribute to Aβ accumulation in the brain.
However, some studies demonstrated that the proinflammatory cytokines IL-6, IL-12, interferon-γ, and TNF-α in whole blood cell cultures were significantly lower [50], and IL-12 was also reduced in the cerebrospinal fluid of AD patients [51].
Different concentrations of inflammatory factors have been reported because of differences between study samples, but most of the literature indicates that inflammation accompanies the progression and/or development of AD. The increased levels of inflammatory cytokines and inflammatory receptor expression support the association between neuroinflammation and AD, and neuroinflammation is a critical part of the pathogenesis of APP/PS1 mice [52]. Previous studies demonstrated that long-term use of nonsteroidal anti-inflammatory drugs, such as ibuprofen, reduced the AD-associated pathology in aged R1.40 mice [53]. One study even demonstrated that blocking TGF-β-Smad2/3 innate immune signaling mitigated the Alzheimer-like pathology in AD mice [54].
Microglia begin to lose the ability to take up and clear Aβ during AD progression [55], and accumulating evidence indicates that peripheral blood monocytes migrate into the CNS in the presence of chronic neuroinflammation to improve Aβ clearance in APP/PS1 mice [56, 57].
Chemokines and other factors promote monocyte migration into the brain in AD
Many factors influence the infiltration of circulating monocytes.
First, chemokine/chemokine receptors, such as CCL2/CCR2, CX3CL1/CX3CR1, CCL5/CCR5, CXCL10/XCR3, CXCL1/CXCR2, SDF-1α/CXCR4, RANTES/CCR1, CCR3, and CCR5, regulate circulating monocyte recruitment into the brain in mouse models of AD [58, 59]. CCR2 was the first chemokine receptor associated with AD, and it facilitates the recruitment of a large number of circulating blood monocytes into the brain. CCR2 may also play an important role in microglial precursor trafficking into the brain in a mouse model of AD [27]. Some studies demonstrated that decreased expression of monocyte CCR2 impaired the recruitment of CCR2+ monocytes in some AD patients [60, 61].
Second, the adhesion of monocytic cells to the AD brain endothelium is partially dependent on endothelial AβPP expression [10]. The cellular signaling initiated by oligomeric Aβ also greatly enhances the transmigration of circulating monocytes in the peripheral blood across the BBB and their subsequent accumulation in the brains of both AD mice and patients [62, 63]. The level of AβPP phosphorylated at tyrosine residue 682 (pAPP) was increased in the endothelial cells of AD patients, and this phosphorylation led to Aβ secretion and may promote the transmigration of monocytes [10].
Third, granulocyte-macrophage-colony-stimulating factor induced BBB opening in the presence of deposited Aβ, which facilitated the infiltration of circulating monocytes in the peripheral blood across the BBB and into the brains of APP/PS1 mice [64].
In addition, the recruitment and activation of different immune cells were significantly attenuated in heparanase-overexpressing mice. LPS stimulation decreased the expression of interleukin-1β (IL-1β), CCL2 and intercellular adhesion molecule-1 (ICAM-1) and then reduced the recruitment of CD45+ blood-borne monocytes to delay Aβ clearance [65].
Infiltration of dysfunctional monocytes in AD
Normally, functioning infiltrating monocytes are critical for the physiological clearance of Aβ. Increased cerebral infiltration of monocytes substantially attenuated disease progression in an AD mouse model via mechanisms that involved enhanced cellular uptake, enzymatic degradation of toxic Aβ and regulation of brain inflammation [66]. When CD11b+ monocytes from wild type donors were injected into AD mice, they were rapidly taken up into the brain, and the progression of AD-related pathology was limited [63, 67]. Another study demonstrated that monocytes were recruited to the brains of 5×FAD mice after chronic infection with T. gondii, and these cells exhibited highly increased phagocytic capacity for Aβ. Chronic Toxoplasma infection ameliorated Aβ in 5×FAD mice via activation of the immune system [68]. Bacillus Calmette-Guérin immunization enhanced the recruitment of monocytes to the brain and alleviated neuroinflammation and cognitive deficits in APP/PS1 mice [69].
The microenvironment influences the ability of macrophages to exert a neuroprotective role. Macrophage polarization is driven towards a proinflammatory (M1) phenotype in an inflammatory microenvironment, which is a state associated with a reduced capacity to remove debris. The M1 state is associated with enhanced proinflammatory cytokine production, which affects neuronal function and health [70]. Although macrophages from both AD and normal mice uptake soluble and oligomeric Aβ in the brain, AD macrophages uptake less Aβ and clear it more slowly than normal and even induce apoptosis [71]. The direct evidence for dysfunctional infiltrating monocytes is that the replacement of brain-resident myeloid cells with infiltrating monocytes did not alter cerebral Aβ deposition in APP/PS1 or APP23 mice models, although the infiltrating monocytes clustered around plaques [72].
In summary, the AD brain microenvironment is abnormal because of the inflammation associated with AD. The infiltrating monocytes are dysfunctional, and Aβ cannot be removed efficiently in AD despite the brain inflammatory cytokines promoting the migration of circulating monocytes into the brain.
CIRCULATING MONOCYTES AND AD
Peripheral clearance of Aβ is mediated primarily via circulating monocytes. Monocytes are selectively attracted to small Aβ aggregates in veins, and the patrolling monocytes carrying the Aβ circulate back into the bloodstream. The depletion of monocytes reduces cell crawling inside Aβ+ veins and increases Aβ deposition in APP/PS1 mice [73].
Proportion of monocytes in the peripheral circulation in AD
Aβ-induced inflammation may change the proportion of lymphocyte subsets. The reported changes in the proportions of peripheral blood monocytes in AD patients are conflicting and confusing.
Some studies reported a higher proportion of monocytes in the peripheral blood of patients diagnosed with AD [74–77]. One of these studies demonstrated that CD68 expression on the peripheral blood monocytes of AD patients was significantly increased, and treatment with antipsychotic drugs decreased CD68 expression [78]. However, some studies showed that the frequency of CD14+ monocytes did not change in AD patients, but the expression of IL-1β was increased [79]. Further studies demonstrated skewed results for peripheral blood monocytes, with a significant increase in IL-6-and IL-23-producing CD14+ cells in AD patients and a significant reduction in IL-10-producing CD14+ cells and the MFI [80].
Monocytes are heterogeneous and classified according to different markers. The complexity of monocyte subsets and the differences in AD patient samples may explain the different proportions of monocytes in the peripheral circulation across studies. The phagocytosis of Aβ is related to the proportions and, even more so, to the phagocytic function of monocytes.
Dysfunction of monocytes in the peripheral circulation in AD
Peripheral monocytes capture Aβ, which diffuses from the brain into the periphery. The defective capacity of peripheral monocytes to engulf Aβ resulted in higher Aβ40 and Aβ42 levels in an AD mouse model [81].
The expression of monocytic cell adhesion molecules, such as ICAM-3 and P-selectin, was significantly decreased in AD patients [82]. The levels of CCL15, CXCL9 and p21 in peripheral blood monocytes were also decreased in AD patients [83]. A study demonstrated that peripheral blood macrophages from AD patients exhibited a weakened ability to uptake and digest Aβ [84], and curcumins restored the phagocytosis of Aβ by macrophages from AD patients via upregulating MGAT3, vitamin D receptor, and toll-like receptors [84, 85].
The sensitivity of monocyte platelets towards Aβ peptides is decreased in AD patients [86]. The competence of monocytes to respond to inflammatory challenges is decreased, but the percentage of cells producing IL-1β, IL-6, IL-12, and TNF-α is increased in AD patients [75, 76]. These monocytes also exhibit elevated expression of the class B scavenger receptor CD36 in AD patients [87].
A functional deficit in phagocytosis was demonstrated when the monocytes of AD patients were pretreated with Copaxone (CPX, to stimulate phagocytosis) or ATP (an inhibitor of P2X7-mediated phagocytosis), and this phagocytosis was associated with the Aβ-burden status [88]. Only one study demonstrated that the increased levels of monocytes in the peripheral circulation may substantially attenuate disease progression in murine AD models via the enhanced cellular uptake and enzymatic degradation of toxic Aβ [66].
Generally, studies about macrophage proportions in the peripheral blood of AD patients revealed complicated results, which may be because of the dynamic changes during the course of AD development. More detailed and thorough research is needed, and samples need to be better characterized due to the heterogeneity of subjects. However, the consensus opinion is that there is a phagocytic functional deficit in the peripheral blood monocytes of AD patients.
Molecular mechanism of the disabled phagocytosis of Aβ in peripheral monocytes in AD
First, as a lipid-sensing activating receptor on microglia, triggering receptor expressed on myeloid cells (TREM) may affect the phagocytosis of Aβ. A TREM1 variant, rs6910730G, is associated with decreased TREM1 surface expression in monocytes isolated from peripheral blood mononuclear cells and alters the Alzheimer-related amyloid accumulation pathology [89]. TREM2 signaling helps protect against AD, and several TREM2 variants decrease binding between TREM2 and its ligands, which might be associated with increased AD risk [90].
Second, the CD33 rs3865444C risk allele is also associated with defective phagocytosis in monocytes because the increased cell surface expression of CD33 in monocytes diminished the internalization of Aβ42 and the neuritic amyloid and fibrillar amyloid accumulation pathology [91, 92]. However, one study demonstrated that the frequency of CD33-positive monocytes was lower and that the mRNA expression of CD33 was downregulated in AD patients [93].
Third, defective phagocytosis may be caused by lower expression of several lysosomal enzymes (cathepsin B, D, and S, β-galactosidase, α-mannosidase, and β-hexosaminidase) in monocytes. These lower expression levels may be a direct consequence of miR-128 upregulation because the inhibition of miR-128 improved Aβ42 degradation in monocytes from AD patients [94].
Lastly, AD pathology may contribute to telomere length shortening, which is decreased in monocytes from AD patients [95]. This shortening may lead to a decrease in the absolute number of macrophages and cause phagocytosis deficiency.
In general, the mechanisms underlying the dysfunction of monocytes are complex, and there may be other factors. Regardless, circulating monocytes exhibit functionally defective phagocytosis in AD, which may directly contribute to the reduced clearance of peripheral Aβ. The imbalance between Aβ generation and clearance increases Aβ concentrations in the peripheral circulation and may further aggravate Aβ deposition in the brain.
Improved phagocytosis in peripheral monocytes reduces Aβ deposition and alleviates AD symptoms
The deficient phagocytic function of circulating monocytes leads to inadequate peripheral clearance of Aβ, which is an important molecular mechanism in AD pathology. Improvement in the phagocytic function of circulating monocytes would reduce brain Aβ levels and slow the development of AD.
Previous studies demonstrated that improving myelomonocytic function via increasing myelomonocytic angiotensin-converting enzyme expression reduced perivascular Aβ deposits and protected against cognitive decline in an animal model of AD [96].
ω-3 fatty acids reduce prostaglandin F2α release from blood mononuclear leukocytes, which decreases IL-6 and IL-1β levels in AD patients [97, 98]. ω-3, ω-6, 1, 25D3, epoxy fatty acids, and some herbal medicines, such as trans-crocetin [99], enhance Aβ phagocytosis by monocytes, and these agents are helpful for improving cognitive decline in AD [100].
Peripheral monocytes pretreated with soluble AβPP α (sAPPα) contributed to Aβ clearance and ameliorated cognitive impairments and reduced Aβ-associated neuropathology in an AD mouse model [101].
Monocytes also affect the pathology of AD in ways other than disabling phagocytosis. Stimulation of human monocytes with soluble AβPP/p3 (αAβPP/p3) or β-processing AβPP/Aβ (βAβPP/Aβ) induced the secretion of soluble factors from monocytes and activated recombinant N-methyl-d-aspartate (NMDA) receptor subtype NR1a/NR2B [102]. This activation is relevant in the pathogenesis of AD, and it is important for synaptic plasticity and related learning and memory in hippocampal brain regions.
The normal phagocytosis function of peripheral monocytes is important for maintaining the balance between the generation and elimination of Aβ. Abnormal Aβ peripheral clearance may participate in AD pathogenesis. Therefore, improving the phagocytosis of peripheral Aβ will be helpful for alleviating AD symptoms.
KCs IN AD
KCs are the resident macrophages of the liver and are the largest macrophage population in vivo. KCs exhibit multiple immune functions and mediate the phagocytosis of Aβ in B6SJLF1/J mice [23, 103]. The clearance of peripheral Aβ by KCs requires the assistance of complement and erythrocytes. Initially, Aβ activates serum complement, and erythrocytes then capture complement-opsonized Aβ via complement receptor 1 (CR1). Ultimately, in AD, hepatic KCs capture the erythrocytes bound to Aβ [28].
The current literature includes few studies on the role of KCs in peripheral Aβ clearance. Therefore, because of the high proportion and absolute number of KCs in the monocyte system, the phagocytic function of KCs requires further research.
CONCLUSION
Our understanding of the pathology of AD has transitioned from intracerebral to extracerebral and from local to systemic with the increase in knowledge of the roles of the inflammatory environment and immune cells. Therefore, it is acknowledged that AD is a systemic immune disease in some ways.
Brain Aβ is the presumptive initiation factor of AD, but the maintenance of its physiological concentration depends on the balance between Aβ production and clearance. Brain Aβ and peripheral Aβ are exchanged to achieve a balance across the BBB, which limits the exchange of substances inside and outside the brain. Monocytes exhibit questionable elimination functions for Aβ because of the inflammatory microenvironment in AD and depending on whether these cells are infiltrated monocytes or circulating monocytes. The impaired peripheral clearance of Aβ by macrophages may lead to abnormal brain Aβ deposition, which may be an important event in the pathogenesis of AD.
Improving monocyte phagocytosis, internalization and clearance functions using different drugs has reduced Aβ deposits in animal models of AD and protected against the cognitive decline in AD.
The role of peripheral Aβ clearance in the development of AD is undeniable. Aβ clearance clearly alleviates the AD pathology in animal models and brings hope for the future treatment of AD. The failure of drugs targeting Aβ clearance in clinical trials need not negate the role of Aβ in AD, and delayed intervention may be responsible for such failures since AD pathology induced by Aβ is irreversible in the later stages of the disease. Therefore, better markers for the early diagnosis of AD may be important for prevention and effective early intervention in AD.
Footnotes
ACKNOWLEDGMENTS
This work was supported by National Natural Science Foundation of China (grant numbers 81471415, 81801088) and the Research Foundation of Xi’an Medical University (grant numbers 12FZ28, 2017DOC14 and 2017GJFY30), Natural Science Basic Research Plan in Shaanxi Province of China (Program No.2018JM7065 and No.18JK0673). We acknowledge the helpful support of the Xi’an Medical University’s key disciplines of molecular immunology. Authors’ disclosures available online (
).
