Cytosolic Phospholipase A₂ Facilitates Oligomeric Amyloid-β Peptide Association with Microglia via Regulation of Membrane-Cytoskeleton Connectivity

Abstract

Cytosolic phospholipase A₂ (cPLA₂) mediates oligomeric amyloid-β peptide (oAβ)-induced oxidative and inflammatory responses in glial cells. Increased activity of cPLA₂ has been implicated in the neuropathology of Alzheimer’s disease (AD), suggesting that cPLA₂ regulation of oAβ-induced microglial activation may play a role in the AD pathology. We demonstrate that LPS, IFNγ, and oAβ increased phosphorylated cPLA₂ (p-cPLA₂) in immortalized mouse microglia (BV2). Aβ association with primary rat microglia and BV2 cells, possibly via membrane-binding and/or intracellular deposition, presumably indicative of microglia-mediated clearance of the peptide, was reduced by inhibition of cPLA₂. However, cPLA₂ inhibition did not affect the depletion of this associated Aβ when cells were washed and incubated in a fresh medium after oAβ treatment. Since the depletion was abrogated by NH₄Cl, a lysosomal inhibitor, these results suggested that cPLA₂ was not involved in the degradation of the associated Aβ. To further dissect the effects of cPLA₂ on microglia cell membranes, atomic force microscopy (AFM) was used to determine endocytic activity. The force for membrane tether formation (F_mtf) is a measure of membrane-cytoskeleton connectivity and represents a mechanical barrier to endocytic vesicle formation. Inhibition of cPLA₂ increased F_mtf in both unstimulated BV2 cells and cells stimulated with LPS + IFNγ. Thus, increasing p-cPLA₂ would decrease F_mtf, thereby increasing endocytosis. These results suggest a role of cPLA₂ activation in facilitating oAβ endocytosis by microglial cells through regulation of the membrane-cytoskeleton connectivity.

Introduction

Growing evidence has shown an important role of microglia in sporadic or late-onset Alzheimer’s disease (LOAD) [1], In addition, increased cytosolic phospholipase A₂(cPLA₂) activity has been observed in AD brains [2,3], although the roles of this enzyme in the pathological expression in AD is not fully understood. The goal of this study is to examine the role of cPLA₂ in soluble oligomeric Aβ (oAβ) association with microglial cells.

As the resident macrophage cells, microglia are key cells responsible for scavenging cell debris, plaques, and damaged neurons and synapses in the brain [4], In recent years, understanding the mechanism(s) for the microglia-mediated clearance of Aβ has become an important direction of AD research. Aβ phagocytosis by microglia has been shown to be governed by a range of receptors [5]. For example, macrophage scavenger receptor 1 (SCARA1) [6,7], CD36 [8-10], and a functional triggering receptor expressed on the myeloid cells 2 protein (TREM2) [11-14] facilitate Aβ uptake. However, some receptors, such as the CD14-Toll-like receptors (TLRs) [15,16], and CD33, a member of the sialic acid-binding immunoglobulin-like lectins (SIGLECS) family [17,18], reduce the ability to phagocytose Aβ. Although receptor-mediated Aβ phagocytosis has been extensively studied, how Aβ-triggered cellular pathways that determine oAβ uptake by microglia are still not fully understood.

Although increased cPLA₂ activity has been implicated in AD [2] and Aβ has been shown to activate cPLA₂in microglia [2,19,20], the role of cPLA₂ in alterations of microglial functions related to AD pathophysiology are not fully understood. There is evidence showing oAβ causes cellular membranes to become more molecularly-ordered in astrocytes, and this process involves activation of cPLA₂ [19]. cPLA₂ has also been found to mediate actin rearrangements [21,22]. Since endocytosis is both a biochemical and a mechanical process (i.e. mechanochemical processes) governed in parts by the membrane-cytoskeleton connectivity [23], we further tested the hypothesis that cPLA₂ plays a role in endocytosis of oAβ in microglia through its effects on membrane-cytoskeleton connectivity.

In this study, we demonstrate that phosphorylation of cPLA₂ in microglial cells is essential in maintaining oAβ association with microglial cells, possibly via membrane-binding and/or intracellular deposition. However, cPLA₂ activation is not involved in the degradation of the associated Aβ in microglia. Our study using cPLA₂ inhibitors and atomic force microscopy (AFM) shows that inhibiting cPLA₂ activation results in elevated force for membrane tether formation (F_mtf), which is a measure for the membrane-cytoskeleton connectivity, and represents a mechanical barrier to endocytic vesicle formation. These data suggest that cPLA₂ facilitates oAβ association with microglia in part through the regulation of the cell membrane-cytoskeleton connectivity.

Materials and Methods

Materials.

Human Aβ₄₂ was purchased from AnaSpec (Fremont, CA). Hexafluoro-2-propanol (HFIP), dimethyl sulfoxide (DMSO), cOmplete protease inhibitor cocktail, PhosSTOP phosphatase inhibitor cocktail, β-mercaptoethanol (β-ME), and Triton X-100 were purchased from MilliporeSigma (St. Louis, MO). Ham’s F-12 was from Crystalgen Inc. (Commack, NY). Dulbecco’s modified eagle medium (DMEM), Penicillin/Streptomycin (P/S), and Dulbecco’s Phosphate-buffered Saline (DPBS) were purchased from Life Technologies (Grand Island, NY). Heat-inactivated fetal bovine serum (HI-FBS) and bovine serum albumin (BSA) were from GE Healthcare Life Science (Logan, UT). Methylarachidonyl fluorophosphate (MAFP), pyrrophenone (Pyr), lipopolysaccharide (LPS) and bromoenol lactone (BEL) were purchased from Cayman Chemical (Ann Arbor, MI). Recombinant mouse and rat interferon gamma (IFNγ) were from R&D Systems (Minneapolis, MN). Radioimmunoprecipitation assay buffer (RIPA buffer), BCA protein assay kit, ProLong™ diamond antifade mountant with DAPI, SuperSignal™ west femto maximum sensitivity substrate, SuperSignal™ west pico plus chemiluminescent substrate, and Restore™ PLUS western blot stripping buffer were purchased from Thermo Scientific (Waltham, MA). Paraformaldehyde (PFA), tween^® 20, and ammonium chloride (NH₄Cl) were purchased from Fisher Scientific (Pittsburgh, PA). Laemmli sample buffer and tris buffered saline (TBS) were from Bio-Rad (Hercules, CA).

Preparation of Oligomeric Aβ.

Oligomeric Aβ₄₂ (oAβ) was prepared according to the protocol as described by Dahlgren et al. 2002 [24].

Primary Rat Microglia Isolation.

All protocols involving the use of animals were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Illinois at Chicago. Timed-pregnant Sprague-Dawley rats were purchased from Charles River (Wilmington, MA) and primary cortical microglia were isolated as described [25]. A neural tissue dissociation kit (P) from Miltenyi Biotec (Auburn, CA) was used to re-suspend cells [26]. Primary microglia were isolated using CD1 lb/c microbeads (Miltenyi Biotec) according to the manufacturer’s instruction.

Primary Rat Microglial Cell Culture.

Primary rat microglia were seeded onto a 6-well plate after isolation with a density of 5× 10⁵ cells per well and maintained in DMEM supplemented with 10% HI-FBS and 1% P/S. Cells were treated after two days of culturing.

BV2 Cell Culture.

The immortalized mouse microglia cell line, BV2 cells, were provided by Dr. Rosario Donato (University of Perugia, Italy). Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) (Life Technologies, Grand Island, NY) with 10% HI-FBS and 1% P/S (100 U/ml penicillin and 100 mg/ml streptomycin). At 80-90% confluency, BV2 cells were detached from the culture flasks with a cell scraper and re-seeded (~10⁵ cells) in new T-75 flasks. Cells were used between passages 15-25.

Measurement of Aβ association with cells.

Primary rat microglia were cultured in 6-well plates and incubated with serum free DMEM for 1 h before treatments. Cells were pretreated with 10 μM MAFP or Pyr for 30 min and remained in the medium. Cells were then stimulated with 1 μg/ml LPS + 10 ng/ml IFNγ for 1 h, followed by incubation with 1μM oAβ for 1h. BV2 cells were cultured in 35 mm dishes to a confluence of 70-80% and incubated with serum free DMEM for 6 h. Cells were pretreated with 1 μM, 5 μM, 10 μM MAFP or Pyr or 2 μM BEL for 30 min and remained in the medium thereafter. Cells were then stimulated with 1 μg/ml LPS +10 ng/ml IFNγ for lh, followed by incubation with 1 μM oAβ for 15, 30, 45 or 60 min. Cells were washed, lysed and collected using RIPA buffer supplemented with complete protease inhibitor cocktail and PhosSTOP phosphatase inhibitor cocktail. Total protein concentrations were determined with BCA kits (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. ELISA assay of human Aβ₄₂ (Life Technologies) was performed according to the manufacturer’s instructions. Data were represented as the total amount of Aβ divided by the total protein per sample (pg Aβ/μg total protein) and normalized by control groups (i.e. Cells treated with oAβ alone or LPS + IFNγ + oAβ).

Western Blot Analysis.

Cells were cultured in 35 mm dishes at a density of 1 × 10⁵ cells/mL overnight achieve 70-80% confluence. Before treatment, BV2 cells were starved with serum free media for 6h. Cells were stimulated with 1 μM oAβ for lh, 0.1 or 1 μg/mL LPS and/or 10 ng/mL IFNγ for various time points to identify the time course for activation of cPLA₂. After treatment, cells were washed twice with cold DPBS and then lysed with 300 μL cold RIPA buffer supplemented with cOmplete protease and PhosSTOP phosphatase inhibitor cocktails for 15 min at 4 °C. Cell lysates were collected into 1.5 mL Eppendorf tubes by a cell scraper. Cell lysates were then centrifuged at 13,000 rpm for 20 min and supernatants were collected. Western blot analysis of p-cPLA₂ was carried out as previously described [26].

Measurement for the Depletion of Associated Aβ with Cells by ELISA Assay.

BV2 cells were starved in serum free DMEM for 6 h and pretreated with 1 μg/ml LPS + 10 ng/ ml IFNγ for 1 h or 20 mM NH₄Cl for 30 min. Cells were then treated with 1 μM oAβ for 15 min and replaced with serum-free DMEM. At time 0, the cells were either collected or further incubated for 5, 15, 30 or 60 min in the presence of 1 μg/ml LPS + 10 ng/ ml IFNγ, 10 μM MAFP, 10 μM Pyr, 2 μM BEL or 20 mM NH₄Cl before cell lysis. Total protein and Aβ quantification analysis was performed in the same manner as the previous section, “Measurement of Aβ Association with cells”.

Immunofluorescence Microscopy of Aβ at the Cell Surface.

Cells were grown on poly-D-lysine pre-coated coverslips in 12-well plate to a density of 8× 10⁴ or 5× 10⁴ cells/well respectively overnight in culture medium. In the next day, cells underwent the aforementioned treatments with LPS, IFNγ, 10 μM MAFP or Pyr, BEL and/or Aβ₄₂. After treatment, cells were washed twice with cold DPBS and fixed with 3.7% PFA in DPBS for 15 min. Cells were washed 3 times with DPBS and then blocked with 5% BSA (w/v) in DPBS for 1 h at room temperature and incubated with Alexa Fluor 488-6E10 (BioLegend, Dedham, MA, 1:200) in 1% BSA in DPBS at 4°C overnight in dark then washed 3 times with DPBS and mounted onto slides with ProLong™ Diamond Antifade Mountant with DAPI and cured at room temperature in dark for 24 h before imaging. Fluorescent images were acquired by a Nikon Eclipse Ti fluorescence microscope with an oil immersion 60X objective lens using a CCD camera. At least 12 images were acquired from each sample. Analysis was performed with CellProfiler software (Carpenter Lab at the Broad Institute of Harvard and MIT) and data was normalized by the cell number.

Mass Conservation for Aβ₄₂ in BV2 Cell Culture.

To quantify p-cPLA₂ mediated uptake and processing of Aβ within microglia, a simple analytical equation was developed based on the principle of mass conservation. The rate of change of intracellular concentration of Aβ can be defined as

$$\frac{d{C}_i}{dt}=k_{up}{C}_o-k_d{C}_i$$ (1)

, where k_up is the Aβ uptake rate constant, k_d the Aβ depletion rate constant, C_i the intracellular concentration of Aβ, and C_o the concentration of Aβ in the medium. In this model, the following assumptions are made: 1) C_o is considered constant, since C_o » C_i ; and 2) k_d is considered to be constant because the depletion of the intracellular Aβ was not affected by cPLA₂ inhibition via MAFP treatment (Fig. 5b). By defining the relative intracellular Aβ concentration, $C_i^{\ast }=\frac{C_i}{C_o}$, with an initial condition, $C_i^{\ast }(t=0)=0$, solving Eqn. 1 yields:

$$C_i^{*}=\frac{K_{up}}{K_d}\left[1-exp(-K_{d}t)\right]$$ (2)

Fig. 5.

Depletion of associated Aβ in BV2 cells. (a) Quantification of associated Aβ in BV2 cells treated with NH₄Cl using ELISA assay. BV2 cells were treated with 1 μg/ml LPS + 10 ng/ml IFNγ for 1 h or 20 mM NH₄Cl for 30 min, followed by adding 1 μM oAβ for 15 min. Afterwards the media was replaced with standard serum-free media and allowed to incubate for 5, 15, 30 and 60 min before lysis. In the case of the 0 min treatment groups, the cells were immediately lysed after incubating cells with 1 μM oAβ for 15 min. Data are represented as the ratios of Aβ to total protein, and normalized by the oAβ group at 0 min. Data are shown as mean ± SD from at least 3 independent experiments (n ≥ 3). ***P<0.001 compared with the oAβ group at 0 min; # P < 0.05, ## P < 0.01 compared with the LPS + IFNγ + oAβ group. (b) Quantification of associated Aβ in BV2 cells treated with MAFP, BEL or Pyr using ELISA assay. BV2 cells were treated with 1 μg/ml LPS +10 ng/mL IFNγ for 1 h, followed by 1 μM oAβ incubation for 15 min. Afterwards, cells were treated with serum-free media, 10 μM MAFP, 10 μM Pyr or 2 μM BEL for 5, 15, 30 and 60 min before cell lysis. Data are represented as the ratio of Aβ₄₂ to total protein, and normalized by the LPS + IFNγ + oAβ group at 0 min. Data are shown as mean ± SD from at least 3 independent experiments (n ≥ 3).

Definitions and units of all parameters in this mass conservation-based model are provided in Table 1. Fitting data from Fig. 4b with Eqn. 2 yields $\frac{k_{\mathit{up}}}{k_d}$ for different concentrations of MAFP (Fig. 6a). Plotting $\frac{k_{\mathit{up}}}{k_d}$ against MAFP concentration exhibits a negative linear relationship between $\frac{k_{\mathit{up}}}{k_d}$ and MAFP concentration (Fig. 6b). Since Fig. 5b suggested k_d was not changed by MAFP, k_up decreased linearly with increasing dose of MAFP.

Table. 1.

Defining the variables of the mass conservation-based model.

Variable	Definition	Units
Ci	Concentration of Aβ associated in the cell	pg Aβ/μg total protein
Co	Concentration of oAβ in medium	μM oAβ
$C_i^{\ast }$	The ratio of Aβ in the cell to oAβ in medium (Ci/Co)	(pg Aβ/μg total protein)/(μM oAβ)
kup	oAβ uptake rate constant	(pg oAβ/μg total protein)/(μM Aβ)(min)
kd	Aβ depletion rate constant	(min−1)
kup.Co	Rate of oAβ uptake	(pg oAβ/μg total protein)/(min)
kd.Ci	Rate of Aβ depletion	(pg Aβ/μg total protein)/(min)
dCi/dt	Rate of change of associated Aβ concentration in the cell	(pg Aβ/μg total protein)/(min)
t	Time	min

Fig. 4.

Aβ association with BV2 cells surface. Cells were pretreated with or without 10 μM MAFP or Pyr or 2 μM BEL for 30 min and treated with 1 μg/ml LPS + 10 ng/ml IFNγ for 1 h, followed by incubation with 1 μM oAβ for 15 min. Fluorescent intensities per cell were normalized by the oAβ group. Data are shown as mean ± SD from 3 independent experiments (n = 3) (at least 12 images were analyzed for each group per experiment). * P < 0.05 compared with the LPS + IFNγ + oAβ group. (a) Representative immunofluorescent images of Aβ association with BV2 cell surface. Aβ was stained with Alexa Fluor 488-6E10 antibody without cell permeabilization. (b) Quantification of immunofluorescent images of Aβ association with BV2 cell surface. MAFP and Pyr did not impose any effect on Aβ association with BV2 cell surface.

Fig. 6.

Application of the mass conservation model to the experimental ELISA Aβ association with BV2 cells data. (a) Normalized concentration of Aβ in cells, $C_i^{\ast }$, from Fig. 4B were plotted against time and fit with a mathematical model, $C_i^{\ast }=\frac{k_{\mathit{up}}}{k_d}(1-\mathit{exp}(-k_d\cdot t\left)\right)$, to yield the ratio of the oAβ uptake rate constant to the Aβ depletion rate constant (k_up/k_d) for different doses of MAFP. (b) The ratio of the oAβ uptake rate constant to Aβ depletion rate constant (k_up/k_d) linearly decreased with increasing dose of MAFP. Data are represented as the mean ± SEM.

Membrane Tethering Force Measurement.

BV2 cells were cultured on poly-d-lysine pre-coated coverslips. Cells were pretreated with 10 μM MAFP or without for 30 min and then treated with 1 μg/ml LPS + 10 ng/ml IFNγ for 1 h. Then the coverslip was washed with PBS and mounted onto a glass slide with Epoxy (3M, St. Paul, MN), cells were kept wet during all the procedures. Tethering forces were measured using a MFP-3D Bio atomic force microscopy (AFM) (Oxford Instruments Asylum Research, Santa Barbara, CA). TR400PSA cantilevers (also known as OTR4, Oxford Instruments Asylum Research) were used. Spring constant of the AFM tips were range from 82.69 to 100.24 pN/nm. Indentation for each measurement is 1 μm with a rate of 1 Hz and the force curve for each measurement was recorded by the software (AR14, Oxford Instruments Asylum Research). At least 10 cells and 5 force curves per cell were analyzed for each group.

Statistical Analysis.

Data is presented as the mean ± standard deviation (SD) or standard error (SEM) from at least three independent experiments. Statistical analysis was carried out with one-way ANOVA and Bonferroni post hoc tests in GraphPad Prism (Version 6.01). P-values less than 0.05 are considered statistically significant.

Results

cPLA₂ Inhibitors Decrease Aβ Association with Primary Rat Microglia.

As compared to the monomeric and fibrillar forms of Aβ, oAβ has been reported to be the most neurotoxic form of aggregates [24]. In this study, we focus on oAβ association with microglia. oAβ was prepared as previously described [24] and was characterized using AFM for quality control as previously described [27]. To demonstrate the role of cPLA₂ in the association of oAβ with microglia, primary rat microglia were pretreated with cPLA₂ inhibitors, MAFP and Pyr, followed by treatment with 1μM oAβ for lh. Aβ association with cells was assessed using ELISA. Pre-stimulation of cells with LPS + IFNγ, known to activate cPLA₂, did not increase Aβ association with cells (Fig. 1a). Interestingly, both Pyr and MAFP decreased Aβ association with primary rat microglia (Fig. 1a), suggesting a role for cPLA₂ in Aβ association with microglia. However, when we performed fluorescent immunostaining of Aβ without cell permeabilization (i.e. fluorescently labeling Aβ at the cell surface only), both cPLA₂ inhibitors did not affect Aβ association with primary rat microglia at the cell surface (Fig. 1b and c). These results suggest that cPLA₂ inhibitors decrease oAβ internalization in primary microglia.

Fig. 1.

Aβ association with primary rat microglia. (a) Quantification of Aβ association with primary rat microglia using ELISA assay. Cells were pretreated with or without 10 μM Pyr or MAFP for 30 min, followed by treatment with 1 μg/ml LPS + 10 ng/ml IFNγ for 1 h, and then cells were exposed to 1 μM oAβ for 1 h. Data are represented as the ratio of Aβ to total protein, and normalized by the control group (oAβ group). Data are shown as mean ± SD from at least 3 independent experiments (n ≥ 3). *** P < 0.001 comparing with the LPS + IFNγ + oAβ group. (b) Representative immunofluorescent images of Aβ association with primary rat microglia at the cell surface. Aβ was stained with Alexa Fluor 488-6E10 antibody without cell permeabilization. (c) Quantification of immunofluorescent images of Aβ association with primary rat microglia at the cell surface. MAFP and Pyr did not impose any effect on Aβ association with primary rat microglial cell surface.

Activation of cPLA₂ in BV2 cells by oAβ and LPS + IFNγ.

To further study the roles of cPLA₂ in Aβ association with microglia, we utilized immortalized mouse microglia (BV2 cells). BV2 cells are the most frequently used cell lines in studies of microglia and have been shown to demonstrate similar inflammatory responses and Aβ accumulation to primary microglia [28,29,26]. Here we demonstrate the activation of cPLA₂ in BV2 cells. When BV2 cells were exposed to 1μM oAβ for lh, activation of cPLA₂ was reflected by the increase in phosphorylated cPLA₂ (p-cPLA₂) (Fig. 2a). cPLA₂ was also activated when cells were exposed to 1μg/mL LPS + 10ng/mL IFNγ and 1γg/mL LPS for 1 h (Fig. 2b). When cells were exposed to 1γg/mL LPS + 10ng/mL IFNγ, the increase in p-cPLA₂ was activated in a time-dependent manner with a maximal activation after 2h (Fig. 2c). p-cPLA₂ was suppressed when cells were pretreated with MAFP for 30 min (Fig. 2d). Based on these results, BV2 cells were stimulated with 1 γg/mL LPS + 10ng/mL IFNγ for 1 h, and different concentrations (5 and 10 μM) of MAFP were applied to further study the role of cPLA₂ in oAβ association with microglia.

Fig. 2.

CPLA₂ activation in BV2 cells. Cells were treated with (a) 1 μM oAβ₄₂ for 1 h, (b) LPS and/or IFNγ (10 ng/ml in all cases except control) for 1 h and (c) 1 μg/mL LPS and 10 ng/mL IFNγ from 0-4 h. (d) BV2 cells were pretreated with 0, 1, 5 or 10 μM MAFP for 30 min, then all groups were treated with LPS (1 μg/ml) + IFNγ (10 ng/ml) for 1 h. The ratio of p-cPLA₂/cPLA₂ is represented as the fraction percentage of the control group. Data are shown as mean ± SD from 3-4 independent experiments (n = 3 or 4), at least 6 independent experiments for (d) (n ≥ 6). *P < 0.05, **P < 0.01, ***p < 0.001, compared with the control group; # P < 0.05, ### P < 0.001 compared with the LPS + IFNγ group in (d).

cPLA₂ Inhibitors Decreases Aβ Association with BV2 cells.

Consistent with the results using primary rat microglia (Fig. 1a), we found that pre-stimulation of BV2 cells with LPS + IFNγ did not increase Aβ association with cells (Fig. 3a). Inhibition of cPLA₂ with MAFP and Pyr decreased Aβ association with BV2 cells (Fig. 3b and d). MAFP and Pyr did not affect Aβ association with BV2 cells at the cell surface (Fig. 4a and b). These results suggest that inhibition of cPLA₂ by MAFP and Pyr reduced Aβ association with microglia, and that mainly occurred intracellularly. Since MAFP at high concentrations can also inhibit calcium-independent PLA₂ (iPLA₂), we found that BEL, a specific iPLA₂ inhibitor, also had no effect on Aβ association with cells (Fig. 3c), thus ruling out the involvement of iPLA₂.

Fig. 3.

Effect of CPLA₂ inhibition on Aβ association with BV2 cells. (a) Quantification of associated Aβ in BV2 cells pre-treated with LPS + IFNγ using ELISA assay. Cells were treated with or without LPS (1 μg/ml) + IFNγ (10 ng/ml) for 1 h, followed by 1 μM oAβ₄₂ treatment for 15, 30, 45 and 60 min. Data are represented as the ratio of Aβ₄₂ to total protein, and normalized by the oAβ group at 15 min. Data are shown as mean ± SD from 3 independent experiments (n = 3). No significant difference was observed between these two groups at any time point. (b) Quantification of associated Aβ in BV2 cells pre-treated with MAFP using ELISA assay. Cells were pretreated without or with MAFP (1 μM, 5 μM and 10 μM) for 30 min, followed by 1 μg/ml LPS +10 ng/ml IFNγ treatment for 1 h, then cells were incubated with 1 μM oAβ for 15, 30, 45 and 60 min. Data are represented as the ratio of Aβ to total protein, and normalized by the LPS + IFNγ + oAβ group at 15 min. Data are shown as mean ± SD from 3 independent experiments (n = 3). **P < 0.01, ***P < 0.001 compared with the LPS + IFNγ + oAβ group at the corresponding time. (c) Quantification of associated Aβ in BV2 cells pre-treated with BEL using ELISA assay. Cells were pretreated without or with 2 μM BEL for 30 min, followed by 1 μg/ml LPS +10 ng/ml IFNγ treatment for 1 h, then cells were incubated with 1 μM oAβ for 15, 30, 45 and 60 min. Data is represented as the ratio of Aβ to total protein, and normalized by the LPS + IFNγ + oAβ group at 15 min. Data are shown as mean ± SD from 3 independent experiments (n = 3). ** P < 0.01 compared with the LPS + IFNγ + oAβ group at the corresponding time. (d) Quantification of associated Aβ in BV2 cells pre-treated with Pyr using ELISA assay. Cells were pretreated without or with Pyr (1 μM, 5 μM and 10 μM) for 30 min, followed by 1 μg/ml LPS + 10 ng/ml IFNγ treatment for 1 h, then cells were incubated with 1 μM oAβ for 15 and 60 min. Data is represented as the ratio of Aβ to total protein, and normalized by the LPS + IFNγ + oAβ group at 15 min. Data are shown as mean ± SD from 3 independent experiments (n = 3). *P<0.05, **P < 0.01, ***P < 0.001 compared with the LPS + IFNγ + oAβ group at the corresponding time.

Fig. 3a also suggests that intracellular Aβ content had already reached equilibrium after exposing cells to oAβ for 15 min. However, Aβ association with cells was decreased by MAFP and Pyr in a dose-dependent manner (Fig. 3b and d). These results suggest that an equilibrium of associated Aβ was reached within 15 min, when the rate of oAβ uptake equaled the rate of Aβ depletion in cells. Inhibition of cPLA₂ by MAFP and Pyr decreased Aβ association with cells, indicating a shift of this equilibrium.

Inhibition of cPLA₂ does not Affect Depletion of Associated Aβ in BV2 Cells.

Aβ association with cells is governed by both the rates of oAβ uptake and depletion of associated Aβ. Although the rate of oAβ uptake cannot be measured directly, it can be indirectly evaluated by measuring both oAβ association with cells (Fig. 3) and the depletion of associated Aβ in cells (Fig. 5). To study the depletion of associated Aβ in BV2 cells, cells stimulated by LPS + IFNγ were exposed to 1μM oAβ for 15 min, followed by replacing oAβ-containing media with standard serum-free media. Cells were then incubated in standard serum-free media for 5, 15, 30 and 60 min and lysed to quantify associated Aβ in cells using ELISA. Fig. 5a shows that pre-stimulation of cells with LPS + IFNγ did not affect depletion of associated Aβ. Associated Aβ decreased and leveled off at ~35% within 5 min and remained unchanged even at the 60-min time point (Fig. 5a). The depletion of associated Aβ can be the result of Aβ recycling back to the media and/or hydrolysis of Aβ in lysosomes. To examine the cause of associated Aβ depletion, we treated cells with 20 mM NH₄Cl, an inhibitor of lysosomal function [30-32] after the treatment of cells with oAβ. We found that the depletion of associated Aβ was completely abrogated in cells pretreated with NH₄Cl (Fig. 5a). These results indicate that degradation of associated Aβ in lysosomes is the major cause of the depletion of associated Aβ in cells. MAFP, Pyr and BEL did not impose any effects on the depletion of associated Aβ (Fig. 5b), indicating that neither cPLA₂ nor iPLA₂ plays a role in the degradation of associated Aβ.

Inhibition of cPLA₂ Decreases the oAβ Uptake Rate Constant.

Since cPLA₂ did not affect the rate of associated Aβ depletion in cells (Fig. 5b), data depicting the effects of cPLA₂ inhibition on Aβ association with cells in Fig. 3b suggest that oAβ uptake rate constant decreases with increasing dose of MAFP. By applying the principle of conservation of mass (see detailed derivation in the Methods section), the normalized concentration of associated Aβ in cells, $C_i^{\ast }$, at different time points can be fit with a mathematical model, $C_i^{\ast }=\frac{k_{\mathit{up}}}{k_d}(1-\mathit{exp}(-k_d\cdot t\left)\right)$, to yield the ratio of the oAβ uptake rate constant to the Aβ depletion rate constant, $\frac{k_{\mathit{up}}}{k_d}$, for different doses of MAFP (Fig. 6a). Plotting $\frac{k_{\mathit{up}}}{k_d}$ against the concentration of MAFP, showing that $\frac{k_{\mathit{up}}}{k_d}$ decreased linearly with an increasing dose of MAFP (Fig. 6b). Since Fig. 5b suggests that k_d was independent of cPLA₂ activation (i.e. k_d remained unchanged with different dose of MAFP), Fig. 6b suggests that the oAβ uptake rate constant, k_up, decreased linearly with an increasing dose of MAFP.

Inhibition of cPLA₂ Increases the Membrane-Cytoskeleton Connectivity.

Membrane-cytoskeleton connectivity governs many mechanochemical processes [23], including endocytosis [23,33,34], Membrane-cytoskeleton connectivity can be quantitatively evaluated by the measurement of force for membrane tether formation (F_mtf) using laser tweezers [33,35,36] and atomic force microscopy (AFM) [37,38], Here, we employed AFM to study the mechanical process underlying the effects of cPLA₂ inhibition on oAβ association with BV2 cells. Fig. 7a is a typical force curve obtained from AFM measurement. The red curve represented the AFM cantilever approaching to the cell in three steps: (1) the cantilever was approaching to the cell without contact from A to B; (2) the cantilever contacted the cell surface at B; (3) the cantilever moved further to make an indentation from B to C (Fig. 7a). The red curve from B to C, where cell indentation was made, is typically fitted with the Hertz model to estimate cell stiffness, which is not the focus of this study. The blue curve represented the cantilever retracting from the cell (Fig. 7a). Along the retraction curve, a sudden release of force was observed at point F, where the rupture of a membrane tether occurred (Fig. 7a). The step change in the force curve during the rupture of a membrane tether was used to measure F_mtf. Fig. 7b showed that MAFP increased F_mtf in unstimulated cells and in cells stimulated with LPS + IFNγ, indicating that inhibition of cPLA₂ activation results in increased membrane-cytoskeleton connectivity in cells. These results suggest that activated cPLA₂ helps attenuate the increase in the membrane-cytoskeleton connectivity to maintain endocytosis of oAβ in stimulated microglia.

Fig. 7.

Membrane tethering force of BV2 cells. (a) A typical AFM force curve from the cell. Red line is the approaching curve and blue line is the retraction curve. Magnified the retraction curve at point F shows a sudden release of force as a membrane tethering force where a membrane tethering rupture event happened. The membrane tethering force measured from this event is around 50 pN. (b) Membrane tethering force in BV2 cells. Cells were pretreated with or without 10 μM MAFP for 30 min, followed by 1 μg/ml LPS + 10 ng/ml IFNγ treatment for 1 h. Data are represented as the mean ± SEM from 45-128 membrane tethering events (n= 45 - 128). *** P < 0.001 compared with the control group; #P < 0.05 compared with the LPS + IFNγ group.

Discussion

Microglia have been found to be immunoreactive for cPLA₂ in central nervous system (CNS) injury and neurodegenerative diseases, including Alzheimer’s disease [2]. Upregulation of cPLA₂ in microglia can be induced through the redox-sensitive NF-κB activation, and in turn, cPLA₂ plays a major role in Aβ-induced NADPH oxidase activity, superoxide production, prostaglandin E2 (PGE₂) formation, iNOS expression, and NO production in microglia [39], However, AD-related cPLA₂ function has yet to be fully elucidated. Our data clearly show that cPLA₂ plays a role in oAβ association with microglia through regulation of the membrane-cytoskeleton connectivity.

The soluble, monomeric, and oligomeric forms of Aβ aggregate to produce fibrillar Aβ (fAβ). However, microglia do not effectively clear fAβ [40,41], and oAβ has been found to be the most neurotoxic among other forms of Aβ [24]. Therefore, understanding the mechanism(s) involving in the maintenance of oAβ homeostasis provides important information regarding the development of AD pathology. Although receptor-mediated phagocytosis of fAβ has been studied extensively, the mechanism(s) underlying oAβ association with microglia are largely unknown. It has been reported that microglia mediate the clearance of soluble Aβ through fluid phase macropinocytosis [42]. The uptake of oAβ is dependent on both actin and tubulin dynamics, but does not involve clathrin assembly, coated vesicles or membrane cholesterol, and the uptake of soluble Aβ and fAβ occurs largely through distinct mechanisms and upon accumulation are segregated into separate subcellular vesicular compartments [42,43]. These previous findings suggest that oAβ uptake by microglia is intimately related to the composition and physical properties of the plasma membrane, and the interaction between the plasma membrane and cytoskeleton beneath the plasma membrane. Since cPLA₂ is a lipid-modifying enzyme, and increased activities have been found in AD brains, it is important to investigate the role of cPLA₂ in oAβ association with microglia.

Lipid-modifying enzymes induce membrane curvature for biogenesis of transport carriers [44,45]. As activated cPLA₂ targets cellular membranes to hydrolyze phospholipids to produce fatty acids and lysophospholipids, it has been hypothesized that converting cylinder-shaped phospholipids into wedge-like lysophospholids by activated cPLA₂ alters lipid geometry favorable for the generation of positive membrane curvature, and supports the budding of tubular transport [46,47]. In turn, cPLA₂ was found to be a key driver for recycling through the clathrin-independent endocytic route [48]. These previous findings suggest that activated cPLA₂ facilitates clathrin-independent fluid phase macropinocytosis of oAβ in microglia.

In addition to membrane bending or curvature, the mechanochemical process of endocytosis, including receptor-mediated and fluid phase endocytosis, is controlled not only biochemically through interaction with regulatory proteins, but also physically through an apparently continuous adhesion (or connectivity) between plasma membrane lipids and cytoskeletal proteins [23]. Membrane-cytoskeleton adhesion is the key mechanical resistance for the formation of endocytic vesicles, and has been found to be the major factor in the force for membrane tether formation (F_mtf) [33]. Therefore, F_mtf can be a measure for the membrane-cytoskeleton connectivity. Modifying the plasma membrane by exposing cells to amphiphilic compounds including detergents, lipids, and solvents, has been demonstrated to cause a parallel decrease in F_mtf and rise in endocytosis rate [35,49]. Here we employed atomic force microscopy (AFM) to measure F_mtf for studying the mechanical mechanism underlying the effects of cPLA₂, a lipid-modifying enzyme, on oAβ association with microglia. F_mtf of (LPS+IFNγ)-stimulated cells was higher than that of control cells (Fig. 7b). This result is consistent with the notion that the inflammatory environment in the AD brain may increase membrane-cytoskeleton connectivity to lower the ability of microglia to macropinocytose oAβ from the milieu, acting to promote disease pathogenesis. In fact, anti-inflammatory agents have been reported to enhance the ability of microglia to process Aβ [50-53], and prevent neuroinflammation, lower Aβ levels and improve cognitive performance in Tg APP mice [54]. Interestingly, inhibition of cPLA₂ by MAFP further increased F_mtf both unstimulated and stimulated cells (Fig. 7b). These results suggest that stimulated cells have higher membrane-cytoskeleton connectivity, and cPLA₂ activation attenuated the increase in membrane-cytoskeleton connectivity to help maintain the endocytosis rate.

Although cPLA₂ did not play a role in the degradation of associated Aβ (Fig. 5b), our data clearly show that ~ 65% of associated Aβ was rapidly digested through lysosomal activities (Fig. 5a). ~35% of associated Aβ remained in the cells and suggested that there was no significant Aβ recycling back to culture medium (Fig. 5a). These results are consistent with previous findings that soluble Aβ is rapidly trafficked into late endolysosomal compartments and subject to degradation without significant Aβ recycling back to culture medium [42]. Once accumulated, Aβ microaggregates colocalize to cellular fractions containing the lysosomal markers β-hexosaminidase and acid phosphatase [55].

cPLA₂ has been implicated in neurodegenerative diseases and brain injury including AD and cerebral ischemia [56-58,3]. Therefore, examining the effects of cPLA₂ activities in brains using animal models have been a research focus. For example, genetic ablation or reduction of cPLA₂ protected human amyloid precursor protein (hAPP) mice against Aβ-dependent deficits in learning and memory, behavioral alterations and premature mortality [59]. Lithium activated brain phospholipase A₂ and improves memory in rats [60]. In vivo inhibition of cPLA₂ in rat brain decreased the levels of total Tau protein [61]. Another line of research on cPLA₂ is to address its functional roles in AD pathology. Modulations of cPLA₂ activation by Aβ stimulation and inhibitors altered membrane fluidity and molecular order [62,19,63]. Therefore, Aβ-activated cPLA₂ may also alter functions of organelles through its ability to alter cellular membrane properties. For example, cPLA₂ has been found to mediate Aβ-induced mitochondrial dysfunction in astrocytes [64]. Consistently, inhibition of cPLA₂ diminished Aβ-induced neurotoxicity [59]. Both NMDA and oAβ activated cPLA₂ leading to release of arachidonic acid (AA) in primary rat neurons [20]. In turn, AA served not only as a precursor for eicosanoids, but also as a retrograde messenger which plays a role in modulating synaptic plasticity [20]. Aβ also induced elevation of APP protein expression mediated by cPLA₂, PGE₂ release, and cAMP response element binding protein (CREB) activation via protein kinase A pathway in primary rat cortical neuronal cultures [65]. In this study, our results contribute to the functional roles of cPLA₂ in AD pathology by demonstrating its role in facilitating oAβ association with microglia through its regulation of membrane-cytoskeleton connectivity. Better understanding the functional roles of cPLA₂ in AD pathology should provide critical insights into the development of therapeutic strategies for AD treatment.