Animals

Female and male young and aged (4–5-month and 17–22-month-old) C57BL/6 mice, and neonatal C57BL/6 mice, were used in these experiments. All experiments were performed under license from the Health Products Regulatory Authority of Ireland in accordance with EU regulations and with local ethical approval (Trinity College Dublin). Animals were housed under controlled conditions (20–22 °C, food and water ad lib) and maintained under veterinary supervision.

Preparation of Primary Microglia from Neonatal Mice

Microglia were prepared as previously described [21]. Briefly, mixed glia were isolated from triturated and filtered cortical tissue of C57BL/6 neonatal mice. Tissue was centrifuged (800 × g, 5 min), the pellet was resuspended in pre-warmed media, triturated and cells were seeded (1 ml) and cultured in T25 cm2 flasks in cDMEM (31,330–038, Bioscience, USA) supplemented with macrophage colony stimulating factor (M-CSF; 100 ng/ml; 416-ML, R&D Systems, UK), and granulocyte macrophage colony stimulating factor (GM-CSF; 100 ng/ml; 415-ML, R&D Systems, UK) for 10–12 days with media change every 3–4 days. Cells were shaken for 2–3 h (37 °C) to dislodge non-adherent microglia and, after centrifuging (800 × g, 5 min), the pellet was harvested, cells counted, seeded in 24-well plates (1 × 105 cells/well), and cultured for a further 2 days.

Cells were incubated for 24 h in the presence or absence of IFNγ (12.5 ng/ml) + and amyloid-β (Aβ; 10 μM comprising 4.8 μM Aβ1–40 and 5.8 μM Aβ1–42; Invitrogen, USA). Aβ was prepared as follows: lyophilised Aβ1–40 and Aβ1–42 peptides were dissolved in HPLC grade water to give a stock solution (6 mg/ml), which was diluted using sterile PBS (final concentration 1 mg/ml). This was allowed to aggregate (24 h, 220 rpm, 37 °C) to give a mix that was shown, by enhanced Thioflavin T binding, to contain both oligomers and fibrils (Supplementary Fig. 1).

Isolation of Microglia from Young and Aged Mice

Tissue from whole brain of young and aged mice was homogenised using the gentle-MACS Dissociator (Miltenyi Biotec, UK) in combination with the AdultBrain Dissociation Kit (130–107-677; Miltenyi Biotec UK); filtered through a MACS SmartStrainer (70 µm); washed with D-PBS containing PBS, calcium (100 mg/l), magnesium (100 mg/l), glucose (1000 mg/l) and pyruvate (36 mg/l), and centrifuged (3000 × g; 10 min, 4 °C) yielding a pellet that contained the microglia as described [22]. Pellets were resuspended in cold D-PBS (3.1 ml), debris removal solution (900 µl) was added, and the suspension was gently overlaid with cold D-PBS (4 ml) and (3000 × g, 4 °C, 10 min). The third phase that contained microglia was resuspended in cold D-PBS (15 ml) and centrifuged (1000 × g, 10 min, 4 °C), and pellets were resuspended in red blood cell removal solution (1 ml) and incubated (10 min, 4 °C). After addition of cold D-PBS (10 ml), samples were centrifuged (300 × g, 10 min, 4 °C) and the pellet was resuspended in D-PBS; microglia were incubated with CD11b microbeads (130–126-725; Miltenyi Biotec, UK), magnetically separated using the QuadroMACS separator (Miltenyi Biotec, UK), and resuspended in PB buffer (PBS + 0.5% foetal bovine serum (10,270–106, Bioscience, USA)). Samples were centrifuged (300 × g; 10 min), the final microglial pellet was resuspended in PBS (75 μl), and cells were seeded in 48-well plates (2.5 × 104 cells/well; final volume 200 μl) and cultured in cDMEM supplemented with M-CSF (100 ng/ml; 416-ML, R&D Systems, UK) and GM-CSF (100 ng/mL; 415-ML, R&D Systems, UK) for 5 days with a change of media on alternate days. At the time of seeding, purity of the microglia isolated was assessed by flow cytometry in aliquots stained with a Hoechst/Propidium Iodide Stain (ThermoFisher Scientific, 62249, P3566), PE/Cy7-CD11b (Biolegend, 101215), and Alexa-fluor-647-GFAP (BD Bio, 561470) antibodies. At the time of seeding, the purity of the microglia isolated with magnetic-CD11b beads was routinely assessed by flow cytometry. Typically, > 95% of live cells were CD45+, CD11b+, and GFAP− (Supplementary Fig. 2).

Analysis of ECAR and OCR

The SeaHorse Extracellular Flux (XF96) Analyser (SeaHorse Bioscience, USA) was used to assess the ECAR in cultured microglia from neonatal mice and OCR in microglia isolated from young and aged mice. For analysis of ECAR, microglia (6 × 104 cells/well) were seeded (180 μl/well) on SeaHorse cell culture microplates (101,085–004, SeaHorse Bioscience, USA) and treated with IFNγ (12.5 ng/ml) + Aβ (10 μM) as above. SeaHorse XF Calibrant solution (200 μl; 100,840–000, SeaHorse Bioscience, USA) was added to each well of the utility plate which was kept overnight in a CO2-free incubator at 37 °C. Cells were washed with assay medium (200 μl) in accordance with the manufacturer’s instructions, and samples were incubated in assay media (final volume, 200 μl/well; 37 °C; 1 h) in a CO2-free incubator. Glucose (10 mM), oligomycin (20 μM), and 2-deoxy-D-glucose (2-DG; 500 mM; all Sigma-Aldrich, UK) were prepared in glycolytic flux assay media and loaded into the appropriate ports for sequential delivery to the sample wells. Following calibration, ECAR was measured every 8 min for 96 min during which time the appropriate compounds were injected at 24-min intervals. ECAR was automatically calculated using the SeaHorse XF96 software; 3–5 replicates for adults and 3 for neonates were assessed for each separate sample.

OCR was assessed in microglia from young and aged male and female mice using the mitochondrial stress test. Microglia (6 × 104 cells/well) were seeded (180 μl/well) on SeaHorse cell culture plates, and the sensor cartridge was hydrated as described above by adding SeaHorse XF Calibrant solution (200 µl; 100,840–000, SeaHorse Bioscience, USA) to each well of the utility plate and left overnight in a CO2-free incubator at 37 °C. Cells were washed twice with assay medium (90 µl), and cells were incubated (37 °C, 1 h) in a CO2-free incubator in a final volume of 180 μl/well. Oligomycin (20 μM; AbCam, UK), carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (20 μM; FCCP; Sigma-Aldrich, UK), and antimycin A/rotenone (40 μM; Sigma-Aldrich, UK) were loaded into the appropriate ports and, after calibration, delivered sequentially. These events allow OCR to be measured every 8 min for 96 min (at 24min intervals). Oligomycin is a complex V inhibitor and allows ATP-linked respiration to be assessed (by subtracting basal OCR from the value following oligomycin). FCCP collapses the inner membrane gradient maximising electron function and allows assessment of maximal respiratory capacity (by subtracting non-mitochondrial respiration from the FCCP rate) and antimycin A/rotenone (40 μM; Sigma-Aldrich, UK) which inhibit complexes III and I, shuts down the electron transport chain, and therefore collapses mitochondrial respiration. OCR was automatically calculated using the SeaHorse XF96 software, and 3–5 replicates were assessed for each separate sample.

Assessment of PFKFB3 by Gel Electrophoresis

Microglia were prepared for western immunoblotting as previously described [23]. Briefly, cells were incubated in lysis buffer (composition in mM: Tris–HCl 10, NaCl 50, Na4P2O7.H2O 10, NaF 50, containing 1% each of Igepal, phosphatase inhibitor cocktails I and II, and protease inhibitor; Sigma, UK), equalised for protein, added to 4 × SDS sample buffer (composition: Tris–HCl 100 mM, pH 6.8, 4% SDS, 2% bromophenol blue, 20% glycerol; Sigma, UK), and boiled (95 °C, 5 min).

Samples were applied to 10–15% SDS gels, proteins were transferred to PVDF membrane, non-specific binding was blocked (5% Marvel in TBS containing 0.05% Tween 20), and membranes were incubated overnight at 4 °C with rabbit anti-PFKFB3 antibody (1:750 in 5% non-fat dried milk/TBS-T; ab181861, AbCam, USA). Membranes were washed and incubated (room temperature, 2 h) with a secondary HRP-linked anti-rabbit antibody (1:2000 in 5% milk in TBS-T; 111–035-003, Jackson, USA)). Bands were detected using WesternBright ECL chemiluminescent substrate (Advansta, USA) and images (Fujifilm LAS-4000 imager) evaluated using ImageJ (http://rsb.info.nih.gov/).

Uptake of Fluorescently Labelled Beads

Uptake of latex beads (1.0 μm mean particle size; L4655, Sigma-Aldrich, Ireland) in a fluorescent yellow-green aqueous suspension was used to analyse phagocytic capacity in microglia isolated from young and aged mice. Cells, prepared as described above, were incubated (4 h) with 0.025% latex beads in fresh cDMEM media. Cells were washed, fixed in 4% PFA, and washed in PBT (PBS + 1% Triton X-100) prior to staining. Cells were blocked (1 h; PBT + 3% BSA), incubated with rabbit anti-Iba1 (19–7141, Wako, Japan 1:1000; overnight; 4 °C), washed, incubated with Alexa Fluor® 546 donkey anti-rabbit IgG (1:1000; 2 h; room temperature, A10040, Invitrogen, USA), and mounted in ProLong Gold with DAPI (P36941, Thermo Scientific, USA). Images (8 fields; 40 × magnification) were acquired with a Zeiss AX10 Imager A1 microscope. Analysis of images was undertaken with the ImageJ software, and the number of latex beads+ microglia was assessed.

Analysis of Phagocytosis of Synaptosomes

Brain tissue was homogenised (10% weight/volume 0.32 M sucrose in HEPES; pH 7.4) and centrifuged (5 min, 5000 rpm, 4 °C) to yield supernatant which was collected and centrifuged (15 min, 14,000 rpm, 4 °C) to provide a synaptosome-enriched pellet which was resuspended in complete culture media (DMEM-F12, FBS 0.5%, penicillin/streptomycin 50 U/ml (15,070–063, Biocience, USA)). Microglia (60,000 cells per well on coverslips in 48-well plates) were incubated (24 h) with synaptosomal preparations (final concentration 1 µg/µl), washed × 3 in complete culture media (1 ml), rinsed in PBS, fixed in PFA (4%, 20 min, RT), and washed before staining.

Coverslips were incubated (1 h, RT) in PBS containing Triton (0.1%) and horse serum (10%), incubated (2 h, RT) with anti-Iba1 (1:1000; 19–7141, Wako) and anti-CD68 (1:100; MCA1957, AbD Serotec) and thereafter with secondary antibodies (1 h, RT; donkey anti-rabbit 1:1000, A10040, Invitrogen for Iba1, and goat anti-rat 1:1000, A21094, Invitrogen for CD68). Coverslips were then incubated with anti-synaptophysin (1:100; overnight, 4 °C; S5748, Sigma UK) and secondary antibody (donkey anti-mouse 1:1000, A21202, Invitrogen; 1 h, RT) with washes between incubations. Coverslips were mounted on slides using Vectashield antifade mounting medium with Dapi (H-1500–10, Vector laboratories) and observed under a confocal microscope (Leica SP8). To quantify synaptosomal engulfment by microglia, z-stacks were taken with a 63 × oil objective. Quantitative 3D in silico modelling (q3DISM) of synaptosome uptake by microglia was adapted from a method previously developed using the Imaris BitPlane software (version 9.1) [24].

Analysis of Microglial Motility

To assess motility of microglia prepared from young and aged mice, we used the wound healing assay. Microglia were cultured on inserts in 48-well plates for 5 days. Two scratches at right angles were drawn using a 10 ml pipette tip to create “the wound.” Images were taken using an Olympus IX51 light microscope with a built-in camera (Olympus, Japan) 8 h after the scratch and the width of the scratches were measured to quantify the distance and speed travelled by the cells.

PFKFB3 Staining

Microglia, prepared as described above, were washed, permeabilised (10 min; PBS with 0.1% Triton-X100), blocked (1 h; 3% BSA with 0.1% Triton-X100), and incubated in the presence of rabbit anti-PFKFB3 (1:250; ab181861, Abcam, UK) and goat anti-Iba1 (1:750; LS-B2645-50, LSBio Inc., USA; overnight; 4 °C) and thereafter with Alexa Fluor® 546 donkey anti-goat IgG (1:1000, A11056, Invitrogen) and Alexa Fluor® 488 donkey anti-rabbit IgG (1:1000; A21206, Invitrogen, 2 h, room temperature). Coverslips were mounted in ProLong®Gold with the nuclear marker DAPI (P36941, ThermoScientific, USA); images were captured using a Leica SP8 scanning confocal microscope (40 × ; 5 fields of view) and processed using the ImageJ software. Total cellular fluorescence (minus DAPI) was corrected by subtracting the product of the area of selected cell fluorescence and the mean background fluorescence from the integrated density as previously described [17].

PFKFB3 mRNA

RNA, isolated from cells using a Nucleospin® RNAII kit (Macherey–Nagel GmbH, Germany), was reverse transcribed into cDNA using a high-capacity cDNA Archive kit (Applied Biosystems, UK), and real-time PCR was performed on duplicate samples with predesigned Taqman gene expression assay for PFKFB3 (Mm00504650_m1) using an ABI Prism 7300 instrument (Applied Biosystems, UK); the endogenous control was β-actin, which was used to normalise gene expression data. Gene expression was calculated relative to the endogenous control samples and to the control sample giving an RQ value (2−DDCt, where CT is the threshold cycle).

Statistical Analysis

Data are reported as the mean ± SEM, and the number of experiments is indicated in each case. Statistical analysis was carried out using Student’s t-test for independent means, or 2-way ANOVA followed by post hoc Tukey’s multiple comparisons test, as appropriate. The significance level was set at p < 0.05.