Background
Alzheimer's disease (AD) is a progressive dementia in which one of the defining characteristics is the deposition of extracellular plaques in the brain [
1]. While beta-amyloid (Aβ) fibrils, a key component of the neuritic plaques, have been demonstrated to be neurotoxic
in vitro [
2‐
4], there is a weak correlation between the severity of dementia and plaque load [
5‐
7]. This suggests that something other than fibrillar Aβ is also contributing to the cell loss and dysfunction characteristic of AD. Recently, the focus has shifted somewhat to the soluble form of Aβ, which has also been found in the cortex of AD patients [
8‐
11]. Interestingly, there is a direct correlation between the levels of soluble oligomers isolated from AD brain and the degree of synaptic loss and cognitive impairment [
12]. Additionally, Aβ oligomers are demonstrated to be neurotoxic
in vitro [
13‐
16]. Increasing evidence suggests that neuronal dysfunction in AD may occur prior to the deposition of fibrillar Aβ and it may be mediated by Aβ oligomers [
17]. These data are beginning to unravel the contribution of both the oligomers and the fibrils in Aβ-mediated neurotoxicity in AD.
Although some
in vitro preparations have demonstrated that Aβ peptides can undergo transitions from monomer to oligomer to protofibril
in vitro [
18,
19] other studies have indicated that physiologically secreted forms of oligomeric Aβ are much more resistant to extracellular multimerization [
20]. Regardless, a number of studies have now demonstrated that oligomeric Aβ is found in varying molecular weight multimers extracellularly in human disease and its rodent models [
8,
21‐
23]. The extracellular multimers have been reported to have a plethora of autocrine effects on neurons. Small molecular weight dimer-trimers have been demonstrated to alter LTP formation both
in vitro and
in vivo [
24‐
26]. In agreement, other small molecular weight forms ranging from approximately 8–42 kDa, depending upon the study, have demonstrated reversible effects on decreasing LTP, dendritic spine density and length
in vitro and direct neurotoxicity as well [
14,
27‐
29]. These multimers reportedly interact with a specific plasmalemmal protein complex involving activation of the NMDA receptor and subsequent activation of the tyrosine kinase, fyn, to carry-out their detrimental effects [
14,
30‐
33]. Larger molecular weight multimers up to 100 kDa have also been reported to have the ability to bind to neurons and decrease neuron viability although the mechanism remains less described [
29,
34]
In addition to direct neurotoxic effects of Aβ peptides they also have the ability to stimulate glia. In particular, fibrillar Aβ has been demonstrated to further contribute to cell loss via stimulating microglia to release neurotoxic mediators that propagate an inflammatory cycle [
35‐
38]. As with the neuronal toxicity data, there is also accumulating evidence that soluble oligomeric intermediates also mediate a portion of this inflammatory response [
39,
40]. These data demonstrated microglia are activated differentially by soluble and protofibrillar Aβ compared to fibrillar [
40,
41]. Moreover, astrocytes are also differentially responsive to the unique peptide conformations [
42]. This suggests that a comprehensive study of the effects of nonfibrillar peptide on microglia activation state is warranted analogous to the observations that have now characterized oligomeric neuron stimulation. Microglial activation is a prominent component of AD histopathology and thus it is of significance to understand the various mechanisms through which these cells are stimulated to acquire a reactive phenotype. These data would provide insight into discreet and perhaps early pathophysiology of the AD brain.
In this study we have compared the ability of oligomeric and fibrillar forms of the Aβ peptide to modulate proinflammatory activation of microglia. We compare the in vitro microglial response and demonstrate a unique activation profile stimulated by both oligomers and fibrils, including tyrosine kinase activation, mitogen-activated protein kinase (MAPK) activation and secretion of cytokines and chemokines.
Methods
Materials
The 4G10 monoclonal anti-phosphotyrosine antibodies were purchased from Millipore (Billerica, MA). Anti-Lyn, anti-Syk, anti-COX-2, anti-phospho-ERK, anti-ERK2 antibody, horseradish peroxidase conjugated secondary antibodies, and protein A/G PLUS-Agarose beads were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-phospho-p38, anti-p38, anti-phospho-JNK, anti-JNK, and anti-phospho-Lyn were from Cell Signaling (Beverly, MA). Anti-CD68 antibody was from Serotec (Raleigh, NC). KC ELISA, IL-6 ELISA, and MCP-1 ELISA were from R&D Systems (Minneapolis, MN). Anti-Aβ, clone 6E10, was obtained from Covance (Emeryville, CA). Anti-oligomer antibody, A11 was purchased from Invitrogen. (Camarillo, CA). DMEM/F12, Neurobasal media and B27 supplements were purchased from Invitrogen (Rockville, MD). BSA was purchased from Serologicals Corporation (Norcross, GA).
Preparation of peptides
In order to generate fibrillar and oligomeric peptides for cell stimulation, Aβ1–42 was purchased from Bachem (Torrance, CA) or American Peptide (Sunnyvale, CA). Oligomers were generated as described in Chromy et al [
43]. Briefly, Aβ1–42 peptide was dissolved in cold hexafluoro-2-propanol (HFIP) to a final concentration of 1 mM. The peptide was aliquoted and dried under vacuum. The aliquoted peptides were stored at -80degreeC until use. For use in cell experiments, the peptide was dissolved in DMSO to a final concentration of 5 mM then diluted to 100 μM in Ham's/F12 media. The peptide oligomers were then incubated 24 hours, 4 degree C then spun 14,000 rpm, 4 degree C, 10 min. The supernatant was collected as the oligomeric Aβ peptide. Fibrils were prepared by dissolving Aβ1–42 peptides in ddH2O to a final concentration of 250 μM then incubated at 37 degree C for 1 week [
44]. Fibrils were resuspended with vigorous trituration prior to removing aliquots for cell stimulation. In order to assess the peptide states under our bioassay conditions, a portion of each preparation was diluted to 20 μM in DMEM/F12 and incubated an additional 48 hours, 37 degree C for subsequent Western analysis. In order to verify that the fibril concentrations employed were accurate following the 1 week fibrillization procedure, five different aliquots of prepared fibril were quantified during different days by Bradford assay to calculate molarities of the solution. This calculated molarity was compared to the predicted molarity based upon the volume of water added to the known mass of purchased peptide used. The difference between predicted molarity and mean calculated molarities from the volume of fibril assayed varied only by 2.4% (predicted 2.5 nM; calculated 2.4 nM ± 0.3). This verified that even though fibrils formed an insoluble precipitate in the solution it was being adequately resuspended for use as a stimulant.
Tissue culture
Microglia were derived from the brains of postnatal C57BL/6J mice as described previously [
45]. Neurons were cultured from cortices of embryonic day 16 (E16) mice (C57BL/6J) as described previously (Sondag and Combs 2006). For co-cultures, neurons were plated at 260 cells/mm
2 in 48 well plates. At day 14, media was removed and replaced with Neurobasal media supplemented with B27 components containing microglia (26 cells/mm
2) and 20 μM Aβ oligomers or fibrils.
Neurons were cultured from cortices of E16 mice (C57B1/6J). Briefly, meninges-free cortices were isolated, trypsinized and plated onto 0.05 mg/mL poly-L-lysine coated tissue culture wells (260 cells/mm2) for 14 days in vitro before use. Neurons were grown in Neurobasal media with glutamine and B27 supplements (Life Technologies, Rockville, MD) to consistently provide neuronal cultures able to survive for at least one month in vitro. Neuron purity was increased up to 98–99% through a transient treatment with 1 μmol/L AraC during days 1–3. Culture purity is routinely evaluated by cell counting after immunostaining to identify the neuronal cytoskeletal protein, microtubule associated protein 2 (MAP2).
Cell stimulation
Microglia were stimulated by removing them from growth media into serum-free DMEM/F12 media (2 × 106cells/mL) containing Aβ oligomer or Aβ fibril. Neurons were stimulated by removing growth media and replacing it with serum free media containing Aβ oligomer or Aβ fibril. Cells were stimulated for 5 minutes or 24 hours and total cell lysates were prepared as described below, or cells were stimulated for 24 hours at 37°C and media was collected.
Non-denaturing electrophoresis
Sample buffer containing no SDS or β-mercaptoethanol was added to Aβ oligomers or Aβ fibrils. Unheated samples were resolved on a 15% polyacrylamide gel in the absence of SDS. Western blotting was performed as described below.
Immunoprecipitation
To perform immunoprecipitations, cells were lysed in ice-cold radioimmunoprecipitation assay (RIPA) buffer. Lysates were then vortexed, incubated on ice for 15 min, briefly pulse sonicated, and centrifuged (10 min, 4 °C) to remove insoluble material. Primary antibody (1 μg/mg protein) was added to the equal protein amounts from supernatants and incubated 4 h at 4°C. Protein A/G beads (Santa Cruz) were added and incubated an additional 4 hr at 4°C. Beads were washed three times with lysis buffer, and immunoprecipitates were resolved and Western blotted as described below.
Western blotting
Western blotting of cell or tissue lysate was done as described previously [
45]. Briefly, ice cold RIPA buffer was used to lyse cells. Lysates were sonicated and centrifuged (14,000 × g, 4°C, 10 min) to remove insoluble material. Protein concentrations were quantitated and proteins were resolved by SDS-PAGE and transferred to PVDF membranes. Western blots were blocked and incubated in primary antibodies. Blots were washed followed by incubation with HRP-conjugated secondary antibodies. Blots were washed again followed by detection with enhanced chemiluminescence (Pierce, Rockford, IL).
Enzyme linked-immuno-sorbent assay (ELISA)
Media was collected from microglia following 24 hour stimulation. Levels of mouse interleukin-6, KC, and MCP-1 in the media were determined using commercially available ELISA kits according to the manufacturer's protocol.
Cell viability assays
Lactate dehydrogenase release (LDH) Assay
Media was collected following 24 hour cell stimulation and centrifuged (14,000 × g, 2 min, 25°C). Aliquots were then added to a 96-well plate and LDH concentrations assayed according to manufacturer's instructions (Promega Corporation, Madison, WI). Background absorbance was subtracted from each condition. Conditions were performed with 12 replicates and each experiment was repeated 3 times. Values were averaged (± SD) as a percent of control release.
Cell counting assay
As an additional means to assess cell viability, neurons were fixed following 72 h stimulation, stained using a neuron specific anti-MAP2 antibody, and a counting grid was placed under the wells to count stained neuron numbers from four identical fields per condition. MAP2 immunoreactive cells with visible nuclei and processes at least one cell diameter in length, were counted as neurons. The average number of neurons (± SEM) was calculated for each condition. Each experiment was performed in quadruplicate 3 times.
Discussion
There is increasing evidence that Aβ oligomers have an early role in AD pathology before the appearance of amyloid deposits [
17]. Importantly, oligomers have been demonstrated to potentiate synaptic loss and inhibit hippocampal long-term potentiation (LTP)
in vitro and
in vivo in the absence of any fibril formation [
24,
57]. Furthermore, transgenic mice overexpressing APP have substantial presynaptic loss before the appearance of Aβ deposits [
58,
59]. Thus, amyloid fibril formation may be an end stage event in a disease process that is mediated by the effects of the oligomers. To date, there are few reports as to how microglia react to Aβ oligomers and how that compares with fibril-mediated activation [
40]. While it is known that Aβ fibrils stimulate microglial activation
in vitro, it is unclear where this fits in the neuroinflammatory timeline of events in AD. Here we examined the effect of oligomeric and fibrillar Aβ1–42 on proinflammatory activation of microglia
in vitro. Microglial cultures treated with oligomeric Aβ exhibited different profiles for changes in tyrosine phosphorylated proteins, MAPK activation, and subsequent cytokine and chemokine production than fibril-treated cultures. These results suggest that oligomeric and fibrillar Aβ1–42 may play distinct roles in the proinflammatory activation of microglia demonstrated in AD. While both conformations of the peptide stimulated increased levels of tyrosine phosphorylated proteins, they did so at qualitatively different levels and the resultant phenotypes differed between the two stimuli. We demonstrated that oligomers and fibrils activate the microglia through unique signaling pathways that include activation of specific MAPKs for each form of the peptide. In addition, the propagation of this signaling response through subsequent activation of Lyn and Syk tyrosine kinases is specific to the oligomeric peptide. Again, this reinforces the differences between the ability of the fibrils and oligomers to stimulate microglia and provide potential therapeutic targets to alleviate inflammation associated with AD. Although the peptide comparisons were based upon molarity calculations, it is possible that the surface area of peptide in the two different states is a variable influencing ability to stimulate microglia. For example, the aggregated, insoluble nature of the fibril may diminish the surface area for stimulating microglia in spite of a comparable or even higher molarity comparison to oligomers. With this caveat in mind, fibrils did induce direct, significant neurotoxicity, assayed by LDH release, at concentrations that were inadequate for oligomers (Fig.
4B). This differential neuron effect at similar concentrations, together with the ELISA and signaling data supports the notion that similar concentration comparisons with microglia are reasonable, at least as a starting point, for oligomers versus fibrils.
It is well-documented that Aβ1–42 stimulates increased cytokine secretion from microglia and cytokine upregulation is a key feature demonstrated in AD [
60,
61]. IL-6 expression is largely increased in AD brain and is believed to have a prominent role in the inflammatory cycle associated with the disease [
62]. Interestingly, we show that both Aβ
o and Aβ
fstimulate increased IL-6 release from microglia but the soluble assemblies stimulate significantly more than the fibrils. In contrast, the fibrillar conformation stimulates a significantly greater amount of the proinflammatory chemokine, KC, than the oligomer. KC is the murine homolog to IL-8, a chemokine with demonstrated proinflammatory effects that is also upregulated in AD brain [
63,
64]. These data demonstrate that Aβ1–42 is indeed a stimulus for increased cytokine production in microglia but more importantly it is the specific conformation of the peptide that dictates the nature of the release. The decrease in MCP-1 release upon oligomer stimulation, while somewhat unexpected, is yet another example of how the oligomers and fibrils differ in the means by which they stimulate microglia. It is possible that these data will provide relevant information regarding when in the disease process some of the inflammatory mediators are released. For example, it has been suggested that IL-6 upregulation occurs relatively early in AD, prior to neuritic changes [
17]. This is supported by our data which demonstrates the oligomer is a potent stimulus for microglial IL-6 release relative to the fibrils.
We and others have demonstrated that Aβ1–42 is neurotoxic
in vitro [
2‐
4]. However, there are still conflicting reports as to whether it is the oligomers or fibrils that are the more potent neurotoxin. It is likely that the different assay parameters are affecting the clarity of this outcome. For example, our results demonstrated that oligomeric peptide was toxic to neurons when viability was assessed via cell counting but was not toxic when cell death was determined by measuring levels of LDH released into the media. In addition, primary neuronal cultures and cell lines appear to respond differently to stimulation with Aβ1–42 [
56,
65]. Also, there may be some discrepancies that can be attributed to the use of the synthetic Aβ1–42 and the consistency of its preparation. We have characterized by denaturing electrophoresis our oligomeric preparation to be approximately dimer-trimer forms initially and multimerizing to several larger molecular weight species with
in vitro incubation. It is important to point out that our species of Aβ may not be the only disease relevant multimers. For example, recent work by Lesne et al has demonstrated that memory deficits in a transgenic AD mouse model can be produced by a 56 kDa multimer of Aβ [
22]. Reports by others have also demonstrated that varying molecular weight oligomers, detected by immunostaining, co-localize with neurons in AD brains [
34,
66] and bind to neuronal membranes
in vitro to affect changes in gene expression [
34]. Others have demonstrated that smaller molecular weight oligomers have distinct effects on neuron dysfunction [
24‐
26]. Although our preparation is consistent with that employed by others [
33] we can not be entirely specific regarding the particular species involved in the stimulated cytokine secretion
in vitro or the toxicity induced during the 48 hour stimulation due to the clear ability of the peptide to multimerize to higher molecular weight species in our hands. Indeed, the fact that large concentrations (10–20 μM) were needed to stimulate signaling responses and cytokine secretion may well suggest that the particular species involved in stimulating microglia or neuron death may be one of the less abundant molecular weight species in our preparation. Future efforts involving chromatographic separation of the different oligomeric forms followed by acute stimulation of microglia may help define any particular receptor interactions as well as determine whether a distinct form is responsible for stimulating the tyrosine kinase-mediated signaling response.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CS performed the majority of experiments and data analysis, and wrote the initial version of the manuscript. GD performed neuron/microglia co-culture stimulations and assessed viability via immunostaining and counting. CC was involved in conceiving the study and coordinating the experiments. He was responsible for editing and revising the manuscript for the final version.