High Plasma Levels of Aβ1-42 Affect Monocytes and Macrophages via Biphasic Effects on Myeloid-Derived Suppressor Cells and Granulocyte-Monocyte Progenitors in Mouse Models of Alzheimer’s Disease

Background: Microglia play a crucial role in the pathogenesis of Alzheimer’s disease (AD). Plasma Aβ 1-42 levels signicantly increase 15 years before the onset of dominantly inherited AD. The effects of high plasma levels of Aβ 1-42 on monocytes and macrophages, the hematogenous counterparts of microglia, remain unclear. Methods: We investigated the effects of plasma Aβ 1-42 on peripheral monocytes and macrophages in three animal models, including 7- and 11-month-old female APPswe/PS1dE9 (APP/PS1) transgenic (Tg) mice, wild-type (Wt) parabiotic with Tg (paWt(Wt-Tg)) mice, and Wt mice. Results: We found that high plasma levels of Aβ 1-42 , in younger (7-month) AD mice signicantly decreased the amounts of pro-inammatory macrophages, myeloid derived suppressor cells (MDSCs), granulocyte-monocyte progenitors (GMP), and the plasma levels of interleukin-6 (IL-6). In older (11-month) AD mice, high plasma levels of Aβ 1-42 signicantly increased the amounts of pro-inammatory macrophages, MDSCs, GMPs, the plasma levels of IL-6 and TNF-α, and the brain inltration of bone marrow-derived macrophages (BMDMs). However, high plasma levels of Aβ 1-42 consistently increased the amounts of monocytes and the proliferation of bone marrow cells (BMCs) without affecting the phagocytic function of macrophages on Aβ 1-42 . Conclusion: The response of mouse AD model suggests that a high plasma level of Aβ 1-42 affects monocytes and macrophages via its biphasic effects on MDSCs and GMPs. We suggest that intervening in the effects of plasma Aβ 1-42 on monocytes and macrophages might offer a new therapeutic approach to AD. CCK-8:cell counting kit-8 assays; SPSS: Statistical Package for Social Sciences; IL-12p70: interleukin-12p70; IL-10: interleukin-10; IL-1β: interleukin-1β; TGF-β: transforming growth factor-β; MPP: multipotent progenitors; CMP: common myeloid progenitors; MEP: megakaryocyte–erythroid progenitors; MCI: mild cognitive impairment; ST-HSC: short-term hematopoietic stem cells; LT-HSC: long-term hematopoietic stem cells.


Background
Alzheimer's disease (AD), the most common neurodegenerative disorder with progressive memory and cognitive loss, is affecting almost 50 million people worldwide, and the incidence of AD is increasing rapidly with the ageing of the world population [1]. The medical care and nursing cost of AD is enormous [2][3][4].
Accumulating studies suggests that neuroin ammation plays an early and crucial role in the genesis of AD pathology [12]. Microglia, the main resident immune cells in the central nervous system (CNS) act as vigilant housekeepers in the adult brain; they activate immediately if the blood-brain barrier (BBB) is disrupted, and they switch their behavior from patrolling to shielding the injured site [13]. Misfolded and aggregated Aβ 1−42 and p-Tau, which can bind to pattern recognition receptors (PRRs) on microglia and astrocytes, trigger innate immune response to release in ammatory mediators, which in turn contribute to disease progression and severity [12,[14][15][16][17][18]. Genome-wide analysis suggests that several genes, encoding for glial clearance of Aβ 1−42 and the in ammatory reaction, increase the risk of sporadic AD [19][20]. In the pathogenesis of AD, microglia might have a double-edged sword role. At the early stage of AD, microglia protect the brain from the toxic effects of Aβ 1−42 by phagocytizing and clearing Aβ [21].
While, as AD progresses, microglia lose their Aβ-clearing capabilities, the expression of microglial Aβ receptors and Aβ-degrading enzymes with persistent production of pro-in ammatory cytokines, results in Aβ deposition. Moreover, complement and microglia mediate the early loss of synapses AD mouse models [22].
While the brain traditionally has been regarded as an immune-privileged organ protected by the BBB, there are interactions between the brain and peripheral organs that have a signi cant role in the development and progression of AD [23]. Neuroin ammation is closely related to peripheral immunity, especially in the late stage of AD, because the BBB is impaired and, hence, peripheral immune cells and in ammatory molecules enter the brain parenchyma [24,25]. There is evidence that circulating neutrophils can extravasate and surround Aβ deposits, where they secrete interleukin-17 (IL-17) and neutrophil extracellular traps (NETs). Moreover, inhibiting neutrophil tra cking or depleting these cells reduces ADlike neuropathological changes and improves memory in AD Tg mice [26].
In recent years, the role of peripheral innate immunity in the pathogenesis of AD has gained more attention [27,28]. Several studies have found that circulating bone marrow-derived macrophages (BMDMs) can enter brain tissue, where they serve as bone marrow-derived microglia that more e ciently phagocytize Aβ 1−40/1−42 compared to resident microglial cells [29,30]. Selective ablation of bone marrowderived dendritic cells increases amyloid plaques in AD mouse models [31]. Increased cerebral in ltration of monocytes, either by elevating the level of circulating monocytes or by weekly treatment with glatiramer acetate (which simulates myelin basic protein), substantially attenuated disease progression in an AD mouse model [32]. Indeed, prior to these observations, monocytic cells derived from bone marrow stem cells had been used to treat AD [33], while long-term use of nonsteroidal anti-in ammatory drugs (NSAIDs) prior to the onset of AD offered protection against AD [34][35][36]. In November of 2019, sodium oligomannate (GV-971), a marine algae-derived oral oligosaccharide, was approved for AD treatment by the Chinese Food and Drug Administration based on its ability to alleviate neuroin ammation by regulating gut microbiota and inhibiting the brain in ltration of peripheral T helper (Th) 1 cells [37].
In dominantly inherited AD and Down syndrome, plasma Aβ 42 levels increased signi cantly 15 years before the onset of symptoms compared with normal people [38,39]. Moreover, in the parabiosis animal model of APPswe/PS1dE9 (APP/PS1) transgenic (Tg) mice and wild-type (Wt) mice and the mice model of transplanting bone marrow cells (BMCs) from APP/PS1 Tg mice into Wt mice, the Aβ 42 from Tg mice plasma or BMCs of Tg mice could signi cantly increase the plasma Aβ 1−42 levels of Wt mice and enter into Wt mice brain to form cerebral amyloid angiopathy (CAA) and Aβ plaques similar to that of Tg mice brain [40,41]. The present study employed three mouse models of AD to examine the effects of high plasma levels of Aβ 1−42 on monocytes and macrophages.

Materials And Methods
All experiments involving mice were approved by the Laboratory Animal Welfare and Ethics Committee of Shenzhen Luohu Hospital Group (SLHG), Shenzhen, China. Detailed materials and methods are presented in the supplementary information. Brie y, 6-8 week-old female pathogen-free APP/PS1 mice and C57BL/6 wild-type (Wt) mice were purchased from Shanghai Model Organisms Center, Inc (SMOC, Shanghai, China). Animals acclimatized for 2-5 days in the pathogen-free animal facility of the SLHG Precision Medicine Research Institute. Mice were housed in a room with a 12-hour light-dark cycle and provided with food and water adlibitum. Parabiosis surgery resulting in shared blood circulation was performed on pairs of mice after they had adapted to each other by living together in a cage for 1 month.
Surgery was performed at 3 months of age on female Tg-Wt littermates and female Wt-Wt littermates (n = 6, per pair) according to the procedure from a previous study [40]. Parabiosis was maintained for 4 and 8 months, while age and weight-matched female Wt mice (n = 6, per group) and female Tg mice (n = 6, per group) without parabiosis were used as controls (Fig. 1a).
Exogenous Aβ 1−42 peptide labeled with HiLyte Fluor 488 (AnaSpec, Fremont, CA, USA) 200ul of 100uM in phosphate buffered solution was intravenously injected into 6-8 week-old female Wt mice via the tail vein (n = 6) three times in a week, and the age and weight matched female Wt mice (n = 6) were used as controls. The mice in control group received the same volume of vehicle. Bioluminescent imaging was Differences among groups were tested by the rank-sum test. Differences of categorical data were tested by Chi-square. For all statistical tests, two-sided P-values less than 0.05 were de ned as statistically signi cant. All analyses were carried out using Statistical Package for Social Sciences (SPSS) version 23.0 software (IBM, West Grove, PA, USA).

3.1
The plasma levels of Aβ 1−42 and brain amyloid deposition in Tg mice and paWt(Wt-Tg) mice.
At 7 months old, Aβ plaques were found in the neocortex and hippocampus of Tg mice and paTg(Wt-Tg) mice, but they were scarce in the neocortex and hippocampus of paWt(Wt-Tg) mice (Fig. 1d). However, at 11 months old, Aβ plaques were very obvious in both the neocortex and hippocampus of Tg mice, paTg(Wt-Tg) mice, and paWt(Wt-Tg) mice (Fig. 1d).

Effects of Aβ 42 on amounts of splenic monocytes and macrophages
At the age of 7 months and 11 months, the amounts of monocytes in the spleen of Tg mice and paWt(Wt-Tg) mice were signi cantly increased compared with those in Wt mice and paWt(Wt-Wt) mice, respectively (p < 0.001 and 0.01) (Fig. 2a), while the amounts of macrophages and anti-in ammatory macrophages in the abdominal cavity among all the groups showed no signi cant difference (p > 0.05) (Fig. 2b, 2d). However, the amounts of pro-in ammatory macrophages in the abdominal cavity of Tg mice and paWt(Wt-Tg) mice were signi cantly decreased at the age of 7 months compared with Wt mice and paWt(Wt-Wt) mice respectively (p < 0.01), and were signi cantly increased at the age of 11 months in Tg mice and paWt(Wt-Tg) mice (p < 0.01) (Fig. 2c).
The IVIS spectrum imaging system identi ed that Aβ 1−42 peptide labeled with HiLyte Fluor 488 entered the circulation of Wt mice 1h after tail intravenous injection (Fig. 1b). Amounts of monocytes and macrophages exhibited no signi cant differences between Wt mice with and those without Aβ 1−42 injection (data not shown).

The effects of Aβ1 1−42 on the in ltration of BMDMs in mice brain
BMDMs were rarely found in brain tissue of 7 month-old Wt mice, paWt(Wt-Wt) mice, Tg mice and paWt(Wt-Tg) mice (data not shown). However, the in ltration of BMDMs in 11 month-old Tg and paWt(Wt-Tg) mouse brains was signi cantly higher when compared with those in Wt and paWt(Wt-Wt) mouse brains, respectively (p < 0.001, and 0.01) (Fig. 3a-3b).
3.5 The effects of Aβ 1−42 on secretion of in ammatory factors by macrophages in the abdominal cavity At the age of 7 months, only the plasma levels of IL-6 were found to be signi cantly reduced in Tg mice and paWt(Wt-Tg) mice compared with Wt mice and paWt(Wt-Wt) mice, respectively (p < 0.01 and 0.05) (Fig. 5a). At the age of 11 months, however, the plasma levels of IL-6 and TNF-α were signi cantly increased in Tg mice and paWt(Wt-Tg) mice when compared with Wt mice and paWt(Wt-Wt) mice, respectively (p < 0.01 and 0.05) (Fig. 5b). The levels of interleukin-12p70 (IL-12p70), interleukin-10 (IL-10), interleukin-1β (IL-1β), and transforming growth factor-β (TGF-β) in all groups showed no signi cant differences (p > 0.05).

The effects of Aβ 42 on MDSCs in spleen
At the age of 7 months, the amounts of MDSCs in the spleen of Tg mice and paWt(Wt-Tg) mice decreased signi cantly compared with those in Wt mice and paWt(Wt-Wt) mice (p < 0.001), while the M-MDSCs proportions showed no alterations (Fig. 6a-6c). At the age of 11 months, the proportions of MDSCs and M-MDSCs in the spleen of Tg mice and paWt(Wt-Tg) mice increased signi cantly compared to the proportions in Wt mice and paWt(Wt-Wt) mice (p < 0.001 and 0.01) (Fig. 6a-6c). The proportions of MDSCs in the spleen of Wt mice was signi cantly increased following intravenous injection of Aβ 1−42 (p < 0.05) (Fig. 6d).
The proportions of MPP, CMP, GMP and MEP in Wt mice given an intravenous injection of Aβ 1−42 remained comparable with those of normal Wt mice (data not shown).

Discussion
In the present study, the plasma Aβ 1−42 levels and Aβ plaques in paWt(Wt-Tg) mice signi cantly increased similar to those of Tg mice. Thus, paWt(Wt-Tg) mice might be a reliable model to investigate the effects and mechanisms of high plasma Aβ 1−42 levels on monocytes and macrophages, which could eliminate or at least attenuate the direct in uence of genetic background of Tg mice. In the early stages (7 monthold) and late stages (11 month-old) of both Tg mice and paWt(Wt-Tg) mice, high levels of plasma Aβ 1−42 could consistently increase monocytes in the spleen without affecting the amount of macrophages in the abdominal cavity. However, the proportions of pro-in ammatory macrophages in the abdominal cavity of Tg mice and paWt(Wt-Tg) mice reduced signi cantly at 7 months, but increased signi cantly at 11 months of age. The proportions of anti-in ammatory macrophages in the abdominal cavity consistently remained stable both in the early and late stages of Tg mice and paWt(Wt-Tg) mice.
We further studied the effects of the high plasma levels of Aβ 1−42 on the secretion of pro-in ammatory and anti-in ammatory cytokines. To that end, we found that the high levels of plasma Aβ 1−42 inhibited the secretion of pro-in ammatory cytokines IL-6 in the early stage of both Tg and paWt(Wt-Tg) mice, whereas increased secretion of IL-6 and TNF-α was observed in the late stage of both animal groups. The secretion of anti-in ammatory factors, including IL-10 and TGF-β, showed no signi cant alterations at either age in any animal group. Additionally, we found that the in ltration of BMDMs into the brain was signi cantly increased only in the late stages of paWt(Wt-Tg) mice and Tg mice (Fig. 9). Collectively, these results suggest that high plasma levels of Aβ 1−42 in the early stage of AD inhibit peripheral innate immunity, whereas in the late stage of AD, they over-activate peripheral innate immunity. These conclusions may be in line with a previous study showing that peripheral monocyte gene expression is pro-in ammatory throughout the course of AD, while the pro-in ammatory gene expression is suppressed at the prodromal stage of disease [42].
We found that the capacity of macrophages to phagocytose Aβ 1−42 in Tg mice and Wt mice, paWt(Wt-Tg) mice and paWt(Wt-Wt) mice, showed no signi cant difference both in animals aged 7 months and 11 months; however, the phagocytosis ability of macrophages was increased after in vitro Aβ 1−42 peptide coculture. In a previous study, Aβ was found to be an effective stimulant of macrophage/microglia phagocytosis [43]. However, the phagocytosis ability of macrophage/microglia in AD patients is controversial and complicated. The basal levels of phagocytosis in all three subsets of monocytes, such as non-classic, intermediate and classic monocytes, were reported to be similar between healthy controls and AD patients, while a signi cant increase in basal phagocytosis was found in subjects with high Aβamyloid burden as assessed by PET scans [44]. However, in another study, Aβ 1−42 uptake by blood monocytes was reduced with ageing and AD [45]. Our study results suggest that although high plasma Aβ 1−42 can activate peripheral monocytes, leading to their in ltration into the brain, these BMDMs might not engulf Aβ 1−42 in the brain and could possibly be harmful by secreting pro-in ammatory cytokines.
Thus, it might not be reasonable to expect that the activated peripheral monocytes and macrophages that in ltrate the brain parenchyma could replace the ageing microglia in AD.
We also investigated the mechanisms by which high plasma levels of Aβ 1−42 affect monocytes and macrophages. In previous studies, Aβ and Tau were found to deposit in bone marrow; simultaneous Aβ treatment promoted osteogenic differentiation via Wnt/β-catenin signaling and inhibited osteoclast differentiation via the OPG/ RANKL/RANK system. Osteoclast activation was regulated by Aβ in an agedependent manner [46][47][48]. To the best of our knowledge, we are the rst to demonstrate that high plasma levels of Aβ 1−42 signi cantly increase the proliferation of BMCs in the early and late stages of AD.
In addition, we also found that MDSCs in spleen and GMP proportions in bone marrow (the precursor cells of monocytes and macrophages in the plasma) signi cantly decrease in the early neurodegenerative stage of Tg mice and paWt(Wt-Tg) mice, but increase in the later stage of Tg mice and paWt(Wt-Tg) mice; these results are consistent with the increase of monocytes in spleen and proin ammatory macrophages in abdominal cavity.
MDSCs are believed to be the most important immune modulatory cells of the innate immune system. Interestingly, in line with our ndings, a signi cant increase in MDSCs was reported in the peripheral blood of patients with amnesic mild cognitive impairment (MCI) when compared with healthy controls or mildly affected AD patients [24]. To date, there are few studies of the changes of HSCs in AD patients and AD animal models. A previous study reported a signi cant decrease of short-term hematopoietic stem cells (ST-HSC) proportion in 12 month-old 3×Tg mice [49]. In our study, the proportions of long-term hematopoietic stem cells (LT-HSC) decreased but ST-HSC proportions increased signi cantly at the age of 7 months in Tg mice, whereas no signi cant difference was observed in the proportions of LT-HSC and ST-HSC among 11 month-old Tg mice, Wt mice and paWt(Tg-Wt) mice (data not shown).
There are some limitations in our study. Firstly, the mechanisms of plasma Aβ 1−42 on monocytes, macrophages, MDSCs and HSCs remain unclear. Secondly, we did not study the effects of plasma Aβ 1−42 on the subtype of splenic monocytes, including non-classic (CD14 dim CD16 + ) monocytes, intermediate (CD14 + CD16 + ) monocytes and classic (CD14 + CD16 − ) monocytes. Lastly, we also did not study the effects of plasma Aβ 1−42 on the amounts and functions of monocytes, macrophages, MDSCs and HSCs in the stages of preclinical, prodromal and clinical AD patients.

Conclusion
We propose that high plasma levels of

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare no competing interests.

Funding
This work was supported by Sanming Project of Medicine in Shenzhen City (SZSM201801014) and Key project of Shenzhen Science and Technology Innovation Committee (JCYJ20200109143431341).
Authors' contributions CL conducted the study, analyzed the data and wrote the manuscript, JZ took part in the study design and reviewed the manuscript, KL analyzed the data and reviewed the manuscript, FZ designed and funded the study, analyzed the data, wrote and nalized the manuscript. All authors read and approved the nal manuscript. Figure 1 Wt mice showed chronic Aβ42 stimulation from Tg mice after parabiosis. a: Schematic diagram depicting the parabiotic pairings. A pair of 3 month-old female Tg mice and age-matched female Wt littermates were used for parabiosis. A pair of 3 month-old female Wt mice were used for parabiotic controls.  BMDMs entered into the brain under AD-like pathology conditions. a: Immuno uorescence image of microglia and BMDM co-stained with CD68, IBAI, and P2Y12 antibodies in Wt and Tg mice brain at the age of 11 months. Scale bars: 500 µm. b: Immuno uorescence image of microglia and BMDM co-stained with CD68, IBAI, and P2Y12 antibodies in paWt(Wt-Wt) and paWt(Wt-Tg) mice brain at the age of 11 months. Scale bars 500 µm. (***P<0.001, **P<0.01, ns denotes no statistical signi cance).    Tg, paWt(Wt-Wt), and paWt(Wt-Tg) mice. MEP were stained with CD34-CD16/32-. (n=6 for each group, Mean±SD, one-way analysis of variance, **P<0.01, *P<0.05, ns denotes no statistical signi cance).

Figure 8
Changes in the proliferation of BMCs. a: The proliferation of BMCs in Wt and Tg mice at the age of 7 and 11 months. b: The proliferation of BMCs in paWt(Wt-Wt) and paWt(Wt-Tg) mice at the age of 7 and 11 months. c: The proliferation of BMCs in Wt and Wt with Aβ42 peptide mice group. d: The proliferation of Page 22/22 BMCs following co-culture with Aβ42 peptide in vitro. (n=6 for each group, Mean±SD, one-way analysis of variance, ***P<0.001, **P<0.01, *P<0.05, ns denotes no statistical signi cance). Schematic diagram depicting the effects of high plasma levels of Aβ42 on peripheral innate immune cells and HSCs In the early stage (7 month-old) of AD in these mice models, the high levels of plasma Aβ42 signi cantly decreased the amounts of peripheral pro-in ammatory macrophages and MDSCs, GMP, as well as the plasma levels of IL-6. In the late stage (11-month-old), the high levels of plasma Aβ42 signi cantly increased the amounts of peripheral pro-in ammatory macrophages and MDSCs, GMP, and plasma levels of IL-6 and TNF-α, as well as the brain in ltration of BMDMs. In addition, the high plasma levels of Aβ42 consistently and signi cantly increased the amounts of peripheral monocytes in both the early and late stages of AD in these mice models.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.