Monocytic cells were cultivated as described before Magga et al. (2012). Briefly, bone marrow was isolated from 6- to 8-week-old C57BL mice. When needed to obtain greater amount of HSCs, or to obtain HSCs from mice over 8-weeks-old, adult mice were treated s.c. with a single dose of granulocyte colony stimulating factor 500 μg/kg (Pegfilgrastim, Neulasta, Amgen, diluted in sterile 0.15 M sodium acetate, pH adjusted to 7.4. with acetic acid) 3–4 days prior to the sacrifice to mobilize HSCs. Then, bone marrow mononuclear cells were isolated by gradient centrifugation with Ficoll paque (GE Healthcare) and HSCs were isolated by immunomagnetic cell separation using CD117 mouse HSC positive selection kit (EasySep, StemCell Technologies). CD117⁺ cells were plated at 100,000 cells/cm² and proliferated in serum-free conditions in a humidified atmosphere at 37°C in 5% CO₂ as described (Malm et al., 2008). Non-adherent cells were replated every 2 days when half of the medium was refreshed. For differentiation, non-adherent cells were collected and plated at 100,000 cells/cm² in Iscove’s modified Dulbecco’s medium (IMDM) in the presence of low endotoxin serum, L-glutamine, penicillin-streptomycin (all products from Gibco, Thermo Fisher Scientific, Waltham, MA, USA), 100 μM β-mercaptoethanol (Sigma) and 10 ng/ml macrophage colony stimulating factor (MCSF; R&D Systems, Minneapolis, MN, USA). After differentiation, cells were collected in PBS when needed.

Primary mouse postnatal day P0-P1 microglia cultures were prepared from cortices and hippocampi and cultivated as a mixed astrocyte/microglia culture as described before Magga et al. (2012). Nonadherent microglia present above the astrocyte layer were collected by shaking the plates 10–15 min at 120 rpm at 37°C and collection of supernatant. Adherent microglia below the astrocyte layer were collected by removal of astrocyte layer with mild trypsinization and collection of remaining microglia from bottom of the flask with repeated pipetting in PBS, as described earlier (Magga et al., 2012). After collection and when plated as microglia culture, both cell types adhered to surface. Microglia were cultivated in IMDM, 10% low endotoxin serum, L-glutamine, penicillin-streptomycin (all products from Gibco, Thermo Fisher Scientific, Waltham, MA, USA).

Aβ species were prepared as described before Magga et al. (2010). Aβ42 (American Peptide) was dissolved into a stock solution of 1 mg/ml in sterile water (soluble Aβ termed as “sAβ”). To obtain fully fibrillized Aβ (termed as “fAβ”), the dissolved peptide was incubated at 37°C for a week. We have previously analyzed these Aβ preparations with immunoblotting for human Aβ (clone 6E10, Signet, Covance) after cross-linking the samples with glutaraldehyde (Sigma; Kanninen et al., 2008; Magga et al., 2010). Immediately after dissolving the peptide, the Aβ42 peptide preparation contained monomers, dimers, various forms of oligomers and large high molecular weight aggregates. Within time, the amount of low molecular weight forms decreased, as analyzed 24 h to 48 h after solubilization. In in vitro studies, we used the concentration of Aβ42 preparation which has been neurotoxic in our previous studies in primary hippocampal neurons and neural stem cells, including dose-response assays for cell viability and neural stem cell migration (Kanninen et al., 2008; Karkkainen et al., 2012; Kärkkäinen et al., 2014).

To obtain native Aβ deposits (termed as “native Aβ”), brains were excised from aged APdE9 mice (Jankowsky et al., 2004) overexpressing human amyloid precursor protein APP695 Swedish mutation and human presenilin-1 deletion in exon 9 (dE9) genes. Brains were frozen on dry ice and processed as previously described (Pihlaja et al., 2008; Magga et al., 2010). Briefly, cryostat-cut 10-μm-thick sagittal brain sections were mounted on glass coverslips and transferred onto 48-well cell culture plates and stored at −20°C until used. Brain sections were thawed shortly prior to the use and rehydrated with cell culture medium before application of cells on top of the sections.

Cells were differentiated for 2 days, and then replated in differentiation medium onto 25-mm glass coverslips (Merck) for another 3 days. Calcium imaging was performed as described previously (Magga et al., 2006). Briefly, cells were loaded with 4 μM fura-2 acetoxymethyl ester (Molecular Probes) at 37°C for 20 min in HEPES-buffered medium (137 mM NaCl, 5 mM KCl, 0.44 mM KH₂PO₄, 4.2 mM NaHCO₃, 10 mM glucose, 20 mM HEPES, 1 mM probenecid, 1 mM CaCl₂ and 1.2 mM MgCl₂ (all reagents from Sigma), pH 7.4). The coverslip was attached to the bottom of a 37°C thermostated perfusion chamber and stimulants were applied by perfusion in HEPES-buffered medium. The cells were excited by alternating 340 and 380 nm UV light with a filter exchange in the control of an InCyt2 system (Intracellular Imaging) and emission was measured through a dichroic mirror and a 510 nm barrier filter with a Cohu CCD camera. A new ratio image (340/380 nm) was collected every second. Primary microglia were applied to calcium imaging 2–3 days after their plating to coverslips. Microglia were labeled with 2 μM fluo-3 acetoxymethyl ester and imaged with TillVision equipped with rapid shifting monochromator Polychrom V with wavelength 488 nm (TILL Photonics GmbH) and perfused with HEPES-buffered medium. Cells were viewed via Olympus IX-70 microscope with specific filter using 10× objective. Images were collected using Sensicam VGA digital camera (PCO AG) at sampling frequency of two frames per second. The data were analyzed with Origin (OriginLab) for quantification of responsive cells and their peak amplitude. The cell was determined viable if it responded to any of the stimuli used.

After 2 days of differentiation, cells were replated onto 24-well plate at 70,000 cells/cm² for another 2 days. Cells were cultivated in differentiation medium and treated with 50 ng/ml LPS (from Escherichia Coli, serotype O111:B4, Sigma) or 20 μM sAβ or 20 μM fAβ. After 24 h of incubation, media were collected, centrifuged to remove any cell debris, and the supernatant was frozen at −70°C. Cytokine concentrations were determined with tumor necrosis factor α (TNFα) enzyme-linked immunosorbent assay (ELISA; R&D Systems, Minneapolis, MN, USA). Cytokine production was also determined from cell culture medium collected after 100 ng/ml LPS, 10 μM sAβ or 10 μM fAβ stimulation. Cytokines were analyzed with cytometric bead array mouse inflammation kit (BD Biosciences) according to manufacturer’s instructions. Primary microglia were treated with stimulants 2 days after their separation from astrocyte co-culture. Cytokine production was analyzed from medium with cytometric bead array kit.

After 2 days of differentiation, cells were replated onto 96-well plate at 100,000 cells/cm² for another day. Cells were cultivated in phenol red–free differentiation medium and treated with 100 ng/ml LPS, 10 μM sAβ, 10 μM fAβ or 300 μM paraquat (Sigma). After 24 h of incubation, media were collected and mixed with an equal volume of Griess reagent (2% phosphoric acid, 1% sulfanilamide, 0.1% naphthylethylene dihydrochloride; Sigma), incubated for 10 min. Absorbance was measured at 540 nm with a multiscan reader. Sodium nitrite (Sigma) was used as a standard.

After 2 days of differentiation, cells were replated onto 96-well plate at 100,000 cells/cm² for another day. Cells were cultivated in differentiation medium and treated with 100 ng/ml LPS, 10 μM sAβ, 10 μM fAβ or 300 μM paraquat (Sigma). After 4 h of incubation, medium was removed and replaced with ROS assay buffer (HBSS, 15 mM HEPES, 5.5 mM glucose and 1.8 mM CaCl₂, Sigma) supplemented with 20 μM 2’7’-dichlorofluorescein (Sigma) and incubated at 37°C for 30 min. Thereafter, cells were washed once with ROS assay buffer and ROS production was measured with excitation 485 nm, emission 535 nm with multi-label microplate reader (Victor Wallac).

After 2 days of differentiation, cells were replated onto 48-well plate as 100,000 cells/cm² for another day. Cells were cultivated in differentiation medium and treated with 100 ng/ml LPS, 10 μM sAβ, 10 μM fAβ or 300 μM paraquat (Sigma). After 24 h of incubation, 10 μM resazurin (Sigma) was applied and incubated for 4 h. Medium samples were collected into a 96-well plate and measured by excitation at 544 nm and emission at 590 nm on a multiplate reader (Victor Wallac).

Cell migration was analyzed by live cell imaging system Cell-IQ, equipped with a phase-contrast microscope and a camera (Chip-Man Technologies, Tampere, Finland). Monocytic cells growing on a 48-well plate were left untreated or treated with either 100 ng/ml LPS, 10 μM sAβ or 10 μM fAβ and cultured in the Cell-IQ incubator for 12 h. Eight replicate wells were analyzed per treatment. Time-lapse images were taken every 15 min and the trajectory and the speed of cell migration were analyzed for 20 cells per well with Cell-IQ Analyzer software. The total number of analyzed cells per group is 160.

Cells were counted and stained as described (Malm et al., 2005). Briefly, cells were fixed with 4% paraformaldehyde (Sigma) for 20 min at room temperature and then permeabilized with 0.05% saponin (Sigma) and stained with CD68 (Serotec). A minimum of 10,000 events were acquired on FACSCalibur flow cytometer equipped with a 488 laser (BD Biosciences). Data analysis was performed using Cellquest Pro software (BD Biosciences).

The cells were incubated in differentiation medium on top of brain sections obtained from aged APdE9 mice containing native Aβ deposits. After 5 days of incubation, Aβ quantification was performed with Aβ ELISA or pan-Aβ immunostaining (both from Biosource) as described (Malm et al., 2008).

Intrahippocampal transplantation was performed as described (Magga et al., 2012). Briefly, 2-year-old transgenic APdE9 mice and their non-transgenic littermates were anesthetized with isoflurane (Baxter) and placed in a stereotaxis apparatus. Monocytic cells (300,000 in 1 μl of HBSS, 2% fetal bovine serum (both from Gibco, Thermo Fisher Scientific, Waltham, MA, USA)) were injected with a Hamilton syringe into the right hippocampus and the same amount of vehicle was injected into the left hippocampus according to following coordinates: ±0.25 mm medial/lateral, 0.27 mm anterior/posterior, 0.25 mm dorsal/ventral from bregma. Brains were excised 4 days post-transplantation and processed as described (Magga et al., 2012). All animal experiments were conducted according to the national regulations for the use and welfare of laboratory animals and approved by the Animal Experiment Committee in State Provincial Office of Southern Finland.

Histological analysis was performed as described (Malm et al., 2005). Briefly, 20-μm coronal sections were stained with Anti-pan Aβ (rabbit polyclonal, Biosource) or Congo red (Sigma), which was applied as 0.2% Congo red in 0.01% NaOH, saturated NaCl/EtOH solution. Stainings were visualized with fluorescence or light microscope and immunoreactive area was quantified using Image Pro Plus software (Media Cybernetics) from the hippocampal region.

The data are expressed as mean ± standard deviation and analyzed with SPSS software using Student’s t-test or one-way analysis of variance when appropriate, followed by Tukey’s post hoc test. The occurrence of calcium responsiveness was evaluated with Fisher’s exact test. *p < 0.05, **p < 0.01, ***p < 0.001.

Monocytic cells were differentiated from HSCs in the presence of MCSF. After a few days of differentiation, these cells represent phagocytic activity as well as response to inflammatory stimulus LPS in terms of increased cytokine production and NO release (Magga et al., 2012). Herein, we aimed to further characterize their response to specific inflammatory stimulus induced by Aβ. First, we exposed monocytic cells to a well-known pro-inflammatory stimulus LPS and measured cellular calcium responses by intracellular calcium imaging. After loading the cells with a fluorescent calcium indicator Fura-2, the cells were perfused with a stimulant and individual cells were imaged for determination of intracellular calcium response. LPS induced a minor elevation in intracellular calcium in monocytic cells. This elevation was either a transient elevation (in 46% of the responsive cells) or slow elevation on the in intracellular calcium level (in 54% of the responsive cells; Figures 1A,B). Both the transient elevation and the low elevation in intracellular calcium level were responses of low amplitude (Figures 1A,B). This minor LPS-induced calcium elevation was accompanied with activation of cell mobility. Monocytic cells were static, but within minutes after LPS application they changed into exploring cells. LPS induced monocytic cell protrusions to elongate to random direction in a non-organized manner (experimental observation, data not shown).

MCSF is a differentiation and growth factor for monocytes (Ginhoux and Jung, 2014). When stimulated with a recombinant MCSF, monocytic cells responded with a transient calcium elevation or with an oscillatory calcium response (Figure 1C). As seen in Figure 1A, LPS induced only a minor calcium response in monocytic cells. LPS stimulation, however, reduced both the occurrence and amplitude of subsequent MCSF-induced calcium response either when LPS was inducing a calcium response of its own or when LPS induced no detectable elevation in intracellular calcium at all (Figure 1D). As comparison, LPS had no effect on adenosine triphosphate (ATP)-induced elevation of intracellular calcium (Figure 1E). This indicates that inflammatory stimulus may specifically interfere with monocytic growth factor signaling while leaving at least some other cellular signaling pathways intact. We have previously shown that omission of MCSF reduces monocytic cell viability but if stimulated with LPS, the cell survival increases back to normal level (Magga et al., 2012). LPS and MCSF may thus induce some competitive cellular pathways leading to enhancement of monocytic proliferation/survival. On the other hand, ATP serves as a potential danger signal in CNS and activates microglia. LPS and ATP-mediated signaling may both induce similar types of signals, i.e., microglial motility and the release of various biologically active substances, such as cytokines and inflammatory proteins (Kettenmann et al., 2011). Since we did not observe an overlap in LPS and ATP signaling, they probably utilize different calcium signaling pathways in monocytic cells.

Distinct from LPS-induced calcium response, freshly solubilized Aβ (termed as sAβ) induced a pronounced transient calcium response or, interestingly, an oscillatory calcium response (Figure 2A). Calcium response to freshly solubilized Aβ was detected in majority of cells while 90% of these responses were oscillatory by nature, and the amplitude was far greater than induced by LPS (Figures 2B,C). Surprisingly, the amplitude of calcium response induced by Aβ decayed after the solubilization event time-dependently (Figure 2B). Also, the occurrence of calcium response decreased after the solubilization event (Figure 2C). As a comparison, the occurrence of calcium responses to ATP in the same cells remained constant and did not decrease time-dependently (Figure 2C). Prefibrillized Aβ (termed as fibrillar Aβ, fAβ) induced a small calcium response (in 50% of the stimulated cells, with Δ 340/380 ratio value of 190) resembling the time-dependent decayed calcium response of sAβ (Figure 2D). As seen in Figure 2A, calcium oscillations induced by freshly solubilized Aβ persisted even after removal of Aβ stimulus. Therefore, calcium responses induced by MCSF could not be reliably determined after Aβ stimulus since they were mainly covered by oscillatory calcium responses induced by sAβ. Also, the magnitudes of ATP-induced calcium responses could not be reliable determined since they were partly masked by persistent oscillatory response induced by sAβ. Aβ may utilize same calcium signaling pathways thus underestimating the magnitude of ATP response. To be noted, ATP did not induce oscillatory calcium signaling but calcium transients.

To examine whether sAβ-induced calcium oscillations occur in other mononuclear phagocytes, we performed calcium imaging for primary microglia, collected from mixed astrocyte/microglia culture as nonadherent microglia cells above the astrocyte layer and as adherent microglia cells growing below astrocyte layer. Microglia had less stable basal intracellular calcium level (Figures 2E,F) compared to monocytic cells (Figures 1C, 2A,D). Microglia produced spontaneous calcium transients without stimulus (Figures 2E,F), which interfered with further detection of possible sAβ-induced responses. However, we did not detect a clear oscillatory pattern in microglia (Figures 2E,F) similar to that observed in monocytic cells. In addition, calcium levels were not further increased after sAβ stimulus, compared to basal levels without stimulus (Figures 2E,F). Both microglia types produced prominent calcium response to ATP (Figures 2E,F), served as a control stimulus. In general, it is very difficult to maintain microglia in resting state in vitro; any change in their culture environment tends to transform them into activated state. This may cause difficulties in sensitive measurements as calcium signaling.

To test whether sAβ-induced calcium oscillations could be reduced with anti-inflammatory treatment, monocytic cells were preincubated with neuroprotective and neuroinflammatory drug minocycline (Tikka et al., 2001; Malm et al., 2008). One day or immediate preincubation with minocycline (5 μM), however, had no effect on sAβ-induced calcium oscillations (data not shown). Taken together, the data suggest that sAβ induces oscillatory, persistent calcium signaling in monocytic cells. We next aimed to decipher how calcium responses induced by Aβ affected monocytic cell activation and viability.

Interestingly, sAβ induced a small, significant production of anti-inflammatory cytokine IL-10 (p < 0.05, Figure 3A) which was transient and decayed until 6 h incubation. There was no significant production of IL-12 (active heterodimer IL-12p70) in response to sAβ (p = 0.06, Figure 3B). Neither synthetic sAβ nor fAβ caused any production of pro-inflammatory cytokines IL-6 (Figure 3C) or TNFα (Figure 3D). Monocytic cells had high production of monocyte chemoattractant protein-1 (MCP-1), but it was not affected by sAβ (Figure 3E). LPS, on the other hand, increased production of all aforementioned cytokines (Figures 3A–E). Cytokine production did not, however, affect cell migration (Figure 3F). Furthermore, when monocytic cells were incubated on top of brain sections from APdE9 mice containing substantial amount of native Aβ deposits as well as inflammatory environment present in the AD mouse brain, no TNFα production was observed (Figure 3G). Synthetic Aβ or native Aβ deposits induced no NO production while LPS induced a notable production of NO (Figure 3H). This is in line with our previous results where we have shown these HSC-derived monocytic cells to produce inflammatory cytokine TNFα and release NO after inflammatory stimulation with LPS (Magga et al., 2012).

In comparison, we also studied how Aβ species activate primary microglia. Unlike monocytic cells, the production of IL-10 and IL-12p70 cytokines in microglia remained under detectable level. Aβ species did not induce any anti-inflammatory cytokine production in microglia. LPS induced production of IL-10 (42 ± 6 pg/ml in nonadherent, 33 ± 2 pg/ml in adherent microglia, respectively) but no IL-12p70. sAβ did not produce any inflammatory activation in microglia, but fAβ induced production of proinflammatory cytokine TNFα in microglia (Figures 3I,J). LPS increased production of IL-6, TNFα and MCP-1 in microglia as expected (Figures 3I,J).

Since sAβ induced calcium oscillations and anti-inflammatory cytokine production but did not induce pro-inflammatory cytokine or NO production, we further investigated whether Aβ induces ROS production or affects monocytic cell viability. Aβ did not induce ROS production while LPS induced a pronounced production of ROS (Figure 4A). Nevertheless, fAβ treatment triggered a slightly reduced cell viability (Figure 4B). Although LPS induced pro-inflammatory activation and clear ROS production, the cell viability was rather increased (Figure 4B). We have previously demonstrated that HSC-derived monocytic cell growth is greatly dependent on MCSF as a growth factor, or on the inflammatory stimulus LPS (Magga et al., 2012).

Finally, we investigated how the persistent oscillatory response of freshly solubilized sAβ (Figure 2A) affects monocytic cell function. We stimulated the cells with freshly solubilized sAβ for 1 h, then washed the cells and changed the medium to exclude further stimulation with aggregating Aβ. When analyzed 24 h later, short stimulation with freshly solubilized sAβ, capable of inducing persistent calcium oscillations (Figure 2A) was not found to reduce cell viability (Figure 4C). While LPS induced the classical ameboid phenotype, sAβ did not change cell morphology into ameboid inflammatory phenotype (Figure 4D). Taken together, sAβ did not induce pro-inflammatory cytokine, NO or ROS production and did not affect cell viability but induced a production of anti-inflammatory cytokine. This finding suggests that the oscillatory calcium response induced by freshly solubilized sAβ is not a detrimental stimulus to monocytic cells.

To investigate the monocytic function in the presence of native Aβ deposits, we cultivated the cells on top of brain sections obtained from aged APdE9 and wild type mice. The environment rich in native Aβ deposits did not affect cell viability (Figure 5A). Monocytic cells retained also their characteristics as phagocytes. This was evidenced by flow cytometry analysis that showed no difference in the expression of lysosomal and phagocytosis marker CD68 on cells incubated on top of brain sections prepared from aged wild type or APdE9 mice (Figures 5B–D). These findings are-in line with our previous report on the potential of monocytic cells to internalize or clear Aβ in vitro, ex vivo and in vivo (Magga et al., 2012). Furthermore, in the areas of high rate of phagocytosis, monocytic cells tended to eventually phagocytose both diffuse and hard Aβ plaques, leaving vascular Aβ plaques intact (Figures 5E,F). In ex vivo assay, monocytic cells are exposed to both insoluble and sAβ during cell attachment and phagocytosis, the level of sAβ42 (19.1 ± 12.1 pg/ml) roughly corresponding to concentrations of Aβ42 in human plasma.

Our previous studies have shown that intrahippocampally transplanted monocytic cells survive well in APdE9 mouse brain and reduce Aβ deposits as detected within pan-Aβ immunoreactive area (Magga et al., 2012). In line with the earlier study, transplanted monocytic cells detected by their green fluorescent protein (GFP) fluorescence (Figure 6A) were observed in the hippocampus 4 days after transplantation. In addition to reduction of diffuse Aβ plaques in the hippocampus (Magga et al., 2012), transplanted monocytic cells also reduced Congo red positive Aβ plaques in the areas occupied by the cells (Figures 6B,C). When Aβ was quantified from the areas devoid of transplanted cells, there was no reduction of Congo red positive Aβ plaques. This finding further suggests that monocytic cells reduce Aβ plaques by phagocytosis and emphasizes the fact that monocytic cells should be drawn into the lesion site at sufficient quantity.