Pharmacological inhibition of APP processing and knock-down of APP in primary human macrophages impairs the secretion of cytokines

Background The amyloid precursor protein (APP) and amyloid β (Aβ) peptides support the innate immune defense as an immune receptor, an antimicrobial peptide and an opsonine. In APP-deficient mouse models, a reduced secretion of cytokines has been observed. It is unclear whether this can be attributed to the lack of APP or to the missing secretion of Aβ peptides.Methods In primary human monocyte derived macrophages the secretion of Aβ peptides was inhibited by the γ-secretase inhibitor DAPT or by the β-secretase inhibitor GL-189, applied as a tripartite substance. Alternatively, APP was knocked down by transfection with siRNA. TNFα, IL-6 and IL-10 were measured by ELISA and the phagocytotic activity was evaluated by flow cytometry.Results Reduced concentrations of TNFα and IL-6 were observed in the media of APPk/d macrophages and after inhibition of the β-, or γ-secretase, especially after additional immunological activation with LPS. Secretion of IL-10 was increased after pharmacological inhibition of APP processing when the macrophages were not immunologically activated but was decreased during LPS-induced inflammation in APPk/d macrophages. No changes of the phagocytotic activity were conclude

Alzheimer's disease, little is known about their physiological functions and their role within the immune system. Inflammatory processes such as the activation of microglia and peripheral macrophages are increasingly considered in the research of AD pathophysiology (3)(4)(5). However it is still not clear, whether neuroinflammation is the cause or the consequence of AD and whether it is harmful or beneficial (3,6,7).
The anti amyloid antibody Aducanumab was associated with an increased incidence of urinary-tract and lung infections (8). Also, a knock-out of APP or the β-site amyloid cleaving enzyme (BACE-1) in mice was associated with a reduced activity of microglia and a reduced secretion of pro-inflammatory cytokines (9)(10)(11)(12)(13). Vice versa, reduced concentrations of Aβ peptides in cerebrospinal fluid (CSF) were also found during brain infections (14,15). Astrocytes express higher amounts of the APP processing enzymes BACE-1 and presenilin 1 upon infection with C. pneumonia (16). Therefore, an immunological function of APP and Aβ peptides can be assumed.
Brain microglia and peripheral macrophages both belong to the mononuclear phagocyte system and part of the microglia seems to be recruited from peripheral monocytes transmigrating into the brain (17,18). Although microglia and peripheral monocyte-derived macrophages differ to some extent, they still share many features (17,19). As primary human microglia is hard to obtain, monocyte derived macrophages are therefore a frequently used model for certain aspects of microglial biology (19)(20)(21). We previously reported that monocytes express APP and that its metabolisation into Aβ peptides depends on their immunological activation (22)(23)(24). Phagocytosis of polystyrene particles and E. coli was shown to be improved by coating the particles with different Aβ peptide variants (25).
Furthermore, an antimicrobial activity of Aβ peptides in cultures of gram positive and gram negative bacteria as well as candida spp. has been observed (26,27). Especially the more hydrophobic Aβ peptide variants seem to agglutinate microorganisms and form channels in their cell membranes (27)(28)(29). These findings could be confirmed in an in vivo model of experimental bacterial meningitis resulting in an improved survival of APP transgenic mice, and a reduced survival of APP k/o mice (30).
The question arises, whether Aβ peptides only support the immune system as opsonin and antimicrobial agent or if they have additional functions as co-stimulatory factors that induce a pro-inflammatory immune response. During inflammation, macrophages secrete a plethora of cytokines (21). Key cytokines indicating a pro-inflammatory reaction are besides others IL-1β, IL-12A, IL-12B and   IL23, IL-6 and TNFα. One of the most important anti-inflammatory cytokines of macrophages is IL-10 (21). We tested, whether the autologous Aβ peptides, secreted by macrophages during inflammatory processes support the immune defense by increasing the secretion of IL-6 and TNFα and by improving the phagocytosis of polystyrene particles.

Preparation and cultivation of monocytes
Monocytes were isolated from buffy coats of anonymous healthy erythrocyte donors (Transfusionsmedizin, Suhl, Germany) by density gradient centrifugation and adhesion to polystyrene cell culture dishes in Dulbecco's modified minimal essential medium (DMEM, Pierce biotechnology, Rockford, USA) without serum. Lymphocytes were removed by thorough washing with 4°C Dulbecco's modified phosphate buffered saline (PBS). Cultures only included monocytes of a single donor. All experiments were replicated with the indicated number of donors (biological replicates). Monocytes were then cultivated at 37°C and 5% CO 2 in RPMI medium (Promocell, Heidelberg, Germany) containing 10% fetal calf serum (FCS, Biochrome, Berlin, Germany) and differentiated into macrophages by adding 40 ng/ml GM-CSF (Immunotools, Friesoythe, Germany). 50% of the medium was exchanged after four days. To avoid interference of endogenous Aβ peptides with those contained in FCS, the medium was changed to serum-free AIM-V medium (Thermo scientific, Dreieich, Germany) at the 7th day in vitro (div).
An inflammatory reaction was induced either by adding 1 µm polystyrene particles (7/cell) (Polysciences, Hirschberg, Germany) or 10 ng/ml lipopolysaccharide (LPS, Sigma-Aldrich, Munich, Germany) to the cell culture medium at the 9th div, 24 h before measuring cytokine secretion or phagocytotic activity.
All cell culture experiments were carried out in duplicates and the viability of the cells was assessed with the Cytotox96 non-radioactive assay (Promega, Mannheim, Germany) (Additional figure 1) as well as the MTT-test according to the manufacturer's instructions (Sigma-Aldrich, Munich, Germany).

Transfection of macrophages
Macrophages were transfected with validated silencer ® select siRNA directed towards APP (ID s1500, Thermo Scientific, Dreieich, Germany) using the viromere blue transfection system (Lipocalyx, Halle, Germany) according to the manufacurer's instructions. On the 7th div. the medium was exchanged with serum free AIM-V medium. APP siRNA was diluted to 2.8 µM with buffer BLUE. Viromer ® BLUE was mixed with buffer BLUE at a ratio of 1:90 and added to the siRNA dilution. After 15 minutes of incubation, 100 µl of the siRNA mix was added to 1 ml of cell culture medium resulting in a final siRNA concentration of 0.14 µM. Non-silencing silencer ® select negative control No. 1 siRNA (Thermo Scientific, Dreieich, Germany) served as control (mock). All experiments were carried out in duplicates.

Phagocytosis-Assay -Flow Cytometry
To assess the phagocytotic activity of macrophages, fluorescent 1 µm polysterene particles were

ELISA of TNFα, IL-6, IL-10
The concentrations of TNFα, IL-6 and IL-10 in the conditioned macrophage media were quantified 48 h 6 after transfecting the macrophages with APP siRNA or adding the secretase inhibitors by commercially available antibody sets (Catalog numbers: IL-6: 31670069, IL-10: 31670109, TNFα: 31673019; all Immunotools, Friesoythe, Germany). Optimized working concentrations of the respective antibodies were established before the experiments. All measurements were run in duplicates. The samples were diluted to be measured within the detection range of the assays and the coefficient of variation of all measurements was below 20%.

Immunoprecipitation, SDS-page and immunoblot
The concentrations of APP and Aβ peptides in macrophage cultures were assessed with SDS-PAGE followed by immunoblotting. Aβ peptides were immunoprecipitated with the n-terminal anti Aβ peptide antibody 1E8 and separated on Tris/Bicine SDS-Pages containing 8 M urea (35). Peptides were transferred to PVDF membranes using a semi-dry westernblot with a discontinuous buffer-system (35). Immunolabeling was performed with the anti-Aβ antibody clone 1E8 and the signal was enhanced by a two-step 7 labeling with a biotinylated goat-anti-mouse antibody and streptavidine conjugated horseradish peroxidase. Finally, membranes were developed with ECL ® advance (GE-Healthcare, Freiburg, Germany) and recorded with the Amersham Imager 600 (GE-Healthcare, Freiburg, Germany).
Quantification of the blots was carried out with the quantity one software (Bio-Rad, Munich, Germany).

Statistical analysis
Statistical analysis was carried out using Prism 6.0 (GraphPad Software Inc., La Jolla, CA, USA). As each experiment was carried out with cells from the same donor, pairwise comparisons were calculated with the ratio paired t-test. Results are presented as mean with standard deviations and were considered to be significant at a p-value < 0.05.

Reduced secretion of TNFα and IL-6 after inhibition of APP processing:
Primary human monocyte derived macrophages were cultivated in serum-free media. The secretion of  Figure 3). Interestingly, IL-10 was found elevated after inhibition of APP processing, but only in cultures without an immunological activation ( Figure 3).

Reduced secretion of IL-6 and IL-10 after inhibition of APP expression
To discriminate the impact of the APP from Aβ peptides, the expression of APP was inhibited by a siRNA knock-down of APP in the same macrophage cultures. Transfection with a non-binding siRNA (mock) served as control and viability was tested as indicated above (Additional figure 1). The reduced concentration of APP in cell lysates 72 h after the transfection is shown in figure 2B ( Figure   2B). The medium remained on the cells for 24 h, 48h after the transfection. The knock-down of APP reduced the concentration of IL-6 in the media of LPS activated macrophages (Figure 4).
Unexpectedly, the secretion of IL-6 and IL-10 was also reduced after transfection with non-binding siRNA. However, the effect of the transfection with siRNA directed towards APP was significantly stronger than that of the transfection with non-binding siRNA. The transfection with APP siRNA did not change the cytokine secretion in cells that were unchallenged or activated by phagocytosis of polystyrene particles (Figure 4). While the pharmacological inhibition of APP processing resulted in increased concentrations of IL-10 in LPS activated cultures, the knock down of APP reduced the concentration of IL-10 in the medium (Figure 4). Again, no change of IL-10 was found in unchallenged or particle-challenged cultures of APP k/d macrophages (Figure 4).

No change in the phagocytic activity of monocytes after inhibition of APP processing or knock-down of APP.
To evaluate the impact of APP expression and Aβ peptide secretion on phagocytosis, APP processing was either pharmacologically inhibited or APP was knocked down by siRNA as detailed above. After establishing the optimal concentration of fluorescent particles and time of measurement, phagocytosis was determined by flow cytometry 240 minutes after adding fluorescent 1 µm microparticles (20 particles/cell) to the cultures (Additional figure 2). However, neither the inhibition of the β-or g-secretase nor the APP knock-down affected the amount of intracellular particles as indicated by the mean fluorescent intensity (MFI) or the fraction of macrophages that is associated with at least one fluorescent particle ( Figure 5). 9 We showed that the pharmacological inhibition of APP processing by the tripartite β-secretase inhibitor T GL-189 and an established g-secretase inhibitor (DAPT) reduced the secretion of IL-6 and increased the anti-inflammatory IL-10 in primary human monocyte-derived macrophages. During LPS induced inflammation reduced concentrations of IL-6 and TNFα were observed. After an APP knock/down, IL-6 and IL-10 were reduced in macrophages which were activated by LPS.

Discussion
A major weakness of this work is the incomplete inhibition of APP expression and processing. Neither To increase the probability of our reported findings not being due to pharmacological side effects, we used two different substances (T GL-189 and DAPT) with two different mechanisms (inhibition of β-and g-secretase). It was described previously that the application of GL-189 as a tripartite substance (T GL-189 ) reduces unspecific side effects by directing the pharmacophore to the catalytic center of the βsecretase (31,32,36). The reported reduction of IL-6 and TNFα as well as the increased secretion of IL-10 are therefore very probable induced by the reduced production of Aβ peptides.
Blockage of the β-secretase pathway normally increases processing via the α-secretase pathway, resulting in increased concentrations of sAPPα (32). While we have not measured sAPPα, our results still suggest that macrophage sAPPα is not able to replace the missing Aβ peptides. This is opposing earlier publications which found that sAPPα activates microglia (39)(40)(41). However, this discrepancy might be an issue of concentration and the impact of amyloid peptides was not assessed in former experiments.
It is long known that Aβ fibrils and oligomers activate macrophages and microglia (39,42). However, our data suggests that not only external Aβ but also the Aβ peptides produced by macrophages themselves have an activating effect on the secretion of proinflammatory cytokines. As a consequence, the missing ability to produce Aβ peptides impaired the proinflammatory reaction induced by LPS. We and others previously reported that the expression of APP and the secretion of Aβ peptides by monocytes/macrophages depends on their immunological activation (22,23,43).

Expression of APP and secretion of Aβ peptides was increased during phagocytosis and LPS-induced
inflammation. In this context it seems possible that the Aβ peptides are part of a self-energizing circuit initiating an immune response.
Further functions of Aβ peptides within the immune defense as antimicrobial agent and opsonine have been shown (26,27,30). The reason, why an inhibition of Aβ peptide generation had no impact on phagocytosis although it changed the concentrations of IL-6, TNFα and IL-10 in this study might be that phagocytosis is strongly affected by opsonines and the expression of receptors involved in phagocytosis but poorly by the investigated cytokines (44). Effects caused by Aβ peptides as an opsonine could probably not be seen in this study because the changes in Aβ peptide concentrations were too small to induce a measurable effect.
Kumar and his colleagues could show that APP knockout mice had a reduced survival, while mice transgenic for APP had an improved survival in a model of infectious meningitis (30). Fitting into this hypothesis, an increased expression of APP, an accumulation of Aβ peptides in the brain and reduced concentrations of Aβ peptides in the CSF were not only observed in patients with Alzheimer's disease but also with meningitis and other inflammatory diseases (14,15,(45)(46)(47).
Regarding AD this would indicate, that the Aβ peptide deposition could be the consequence and not the cause of neuroinflammation. This idea is supported by epidemiological data showing a reduced risk of AD in patients using non-steroidal anti-inflammatory drugs (48). TNFα antagonists also seem to improve cognitive performance in AD patients (49,50). Some even hypothesize an infectious agent as the cause of AD (51)(52)(53)(54)(55).
Pharmacological inhibition of Aβ peptide generation reduced Aβ 1-x but not n-terminal modified Aβ peptides. This indicates, that the Aβ 1-x species are responsible for the observed differences. As we did not analyze the aggregation state of the Aβ peptides in our cultures we are not able to differentiate whether Aβ monomers, oligomers or fibrils are responsible for the observed effects. However, Aβ aggregation takes place within few hours and aggregation of Aβ peptides in cultures of macrophages has been shown (56). Therefore, it seems very likely, that at least part of the secreted Aβ peptides aggregate to oligomers and fibrils. Several receptors expressed by macrophages have been shown to bind Aβ peptide fibrils or oligomeres, e.g. CD14, CD36, macrophage scavenger receptor 1, N-formylpeptide receptor like-1 and APP (10,57). Binding of these receptors triggers downstream thyrosin kinases, release of Ca ++ and ultimately activation of NFkB and CREB (39,(57)(58)(59)(60)(61)(62)(63)(64).
In microglial cultures of APP knock-out mice as well as in brains and intestines of these mice a reduced motility of macrophages as well as reduced concentrations of several cytokines, including IL-6, TNFα and IL-10 were observed which is in accordance to our findings (9, 10, 12). However, due to their methodology, the authors could not discriminate between the effects caused by APP and those caused by Aβ peptides. Consequently, they discuss the role of APP as a receptor for LPS or a transcription factor. The different effects of the APP knock-down and pharmacological inhibition of APP processing concerning the IL-10 concentrations after stimulation with LPS support this assumption.
Pro-and anti-inflammatory activities are reduced in APP k/o /APP k/d macrophages. When APP as a cell bound protein remains intact, the pharmacological inhibition of Aβ peptide generation removes a proinflammatory peptide and might result in a more anti-inflammatory state of the macrophages with reduced secretion of IL-6 and increased secretion of IL-10.

Conclusions 12
Taken together, the presented data supports the hypothesis that APP and Aβ peptides expressed and secreted by macrophages are involved in initiating and regulating immune responses. In clinical trials testing Aβ lowering therapies, dysfunctions of the immune system should be closely monitored. No ethic approval was necessary for this study. The human material (buffy coats) was purchased at the Transfusionsmedizin, Suhl. The identity of the healthy donors is not known to the researchers.

Consent for publication
Not applicable

Availability of data and materials
The datasets generated and/or analysed for the current study are available from the corresponding author on reasonable request

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