Reactive microglia are commonly observed in association with the β-amyloid (Aβ) plaques of Alzheimer's disease brains. This localization supports the hypothesis that Aβ is a specific activating stimulus for microglia. A variety of in vitro studies have used postnatal derived rodent microglia cultures to characterize the ability of Aβ to stimulate these cells. However, it is unclear whether this paradigm accurately models conditions in aged animals. To determine whether Aβ stimulatory phenotypes differ between young and adult microglia, we established cultures of acutely isolated adult murine cortical microglia to compare with postnatal derived microglial cultures. Although cells from both ages expressed robust immunoreactivity for CD68 and CD11b, their responses to activating stimuli differed. Fibrillar Aβ was rapidly phagocytosed by postnatal microglia and both oligomeric and fibrillar peptide stimulated increased tumor necrosis factor α (TNFα) secretion. However, Aβ oligomers but not fibrils stimulated TNFα secretion from adult microglia. More importantly, adult microglia had diminished ability to phagocytose Aβ fibrils. These findings demonstrate that adult microglia respond to Aβ fibril stimulation uniquely from postnatal cells and suggest that adult rather than postnatal microglia cultures are more appropriate for modeling proinflammatory changes in the aged CNS.
The senile plaques in Alzheimer's disease (AD) brains contain fibrillar Aβ peptides and are surrounded by activated microglia (Itagaki et al., 1989; Cotman et al., 1996). These observations support the hypothesis that microglia are stimulated to acquire a reactive phenotype in response to interaction with Aβ fibrils (Akiyama et al., 2000). A plethora of consequences of amyloid-mediated microglial activation has been characterized in vitro including increased secretion of proinflammatory products such as reactive oxygen species, cytokines, and neurotoxins. (Giulian et al., 1995; Meda et al., 1995; Tan et al., 2000; Combs et al., 2001; Xie et al., 2002). Collectively, data such as these support the hypothesis that Aβ-stimulated microglial activation contributes to the pathophysiology of AD. Not surprisingly, strategies to limit the Aβ-microglial response are of therapeutic interest.
However, many in vitro studies are performed using microglia derived from brains of postnatal rodents. Microglial preparation from postnatal rodent brains is a reliable, established method that yields a cell population well characterized in the literature (Giulian and Baker, 1986). These cells are taken from the postnatal brain during the period of active microglial proliferation in vivo (Dalmau et al., 2003) and grow as a loosely adherent cell population atop a bed of mixed glia over several weeks in culture. After isolation from the mixed glial population, a nearly pure culture of microglia is obtained. However, these purified cells are somewhat reactive in culture in comparison with in vivo adult microglia based on morphology and increased immunoreactivity for a host of microglial marker proteins (Carson et al., 1998; Nimmerjahn et al., 2005). This discrepancy suggests that the microglial response to Aβ derived from postnatal culture studies may differ significantly from that used by adult or aged microglia. In fact, in aged humans and rodents, it has been demonstrated that microglia undergo a morphologic dystrophy characterized by enlargement, deramification, and adoption of a more phagocytic phenotype (Sheng et al., 1998; Streit et al., 2004). These differences suggest that microglial cultures from aged brains may offer a more relevant model system for determining cellular responses to Aβ stimulation during disease. Importantly, adult microglia culture protocols demonstrate that acutely isolated cells retain a quiescent in vivo phenotype, although increased culture time results in a more reactive phenotype (Becher and Antel, 1996; Slepko and Levi, 1996; De Groot et al., 2000; Frank et al., 2005; Ponomarev et al., 2005).
Based on the previous work demonstrating the feasibility of adult microglia preparations, we sought to determine whether the Aβ stimulatory response we had previously characterized from postnatal derived microglia was similar to that induced in adult cells (Floden et al., 2005). To minimize phenotype conversion of the microglia induced by prolonged culturing, we chose to use cells acutely isolated from the adult mouse brain.
Materials and Methods
The anti-β-amyloid IgG1 clone 6E10 antibody (residues 1–17) was purchased from Signet Laboratories (Dedham, MA). Anti-β-amyloid IgG1 clone BAM-10 (residues 1–12), and lipopolysaccharide (LPS) were purchased from Sigma (St. Louis, MO). The CD11b and CD68 antibodies were purchased from Serotec (Raleigh, NC). FITC-labeled Escherichia coli (K-12 strain) bioparticles were from Invitrogen (Eugene, OR).
Microglia were derived from postnatal day 1 (P1) to P3 mouse brains (C57BL/6) as described previously (Floden et al., 2005). Postnatal microglia were isolated from 14 d in vitro cultures for immediate use. Adult microglia were taken from 5–8 month mouse brains (C57BL/6) as previously described and used immediately after isolation (Carson et al., 1998).
Adult and postnatal derived microglia were plated onto 96-well tissue culture plates (20,000 cells/well; 75 μl of serum-free DMEM/F12) for 6 or 24 h. Cultures were stimulated with fibrillar or oligomeric Aβ peptide or 25 ng/ml LPS. Experiments were performed with eight replicates per condition, three to four independent times. Data are presented as mean ± SD. Values statistically different from controls were determined using one-way ANOVA. The Tukey–Kramer multiple-comparison post test was used to determine p values.
Human Aβ1–42 preparation.
To generate fibrillar and oligomeric peptides for cell stimulation, Aβ1–42 was purchased from Bachem (Torrance, CA) or American Peptide (Sunnyvale, CA). Aβ oligomers were generated as described by Chromy et al. (2003). Fibrils were generated as described previously (Floden et al., 2005). A FITC conjugate of human Aβ1–42 (FITC-Aβ; rpeptide, Athens, GA) was used to prepare fluorescent aggregates for the phagocytosis assays according to the manufacturer's instructions. All peptides were resolved via SDS-PAGE and Western blotted to confirm an oligomeric (dimer/trimer) or fibrillar conformation before use. All peptide stimulations were performed using the same purchased lot and preparations of peptide conformations for both adult and postnatal cultures to insure an accurate comparison between cell ages.
Quantitation of secreted tumor necrosis factor α.
After 24 h stimulation, 70 μl of media was removed from the individual culture wells for tumor necrosis factor α (TNFα) quantitation. Concentrations of secreted TNFα were then determined using commercially available mouse TNFα colorimetric sandwich ELISA plates (R & D Systems, Minneapolis, MN).
Peptide phagocytosis was quantitated by measuring the uptake of FITC-labeled Aβ1–42. Briefly, Aβ aggregates (500 nm) or FITC-E. coli bioparticles (positive control, 0.25 mg/ml) were incubated with the microglia in 96-well plates for 6 h. To quench the signal from extracellular peptide, medium was removed and the cells were rinsed with 0.25 mg/ml trypan blue in PBS. Application of trypan serves to quench any remaining peptide on the plate as well as any bound to the external leaflet of the plasmalemma. Intracellular fluorescence was read (480 nm excitation and 520 nm emission) via fluorescent plate reader (Bio-Tek, Winooski, VT).
Cell viability assay.
To determine cell viability after 24 h stimulation, cellular release of lactate dehydrogenase (LDH) was measured from culture media using a commercial nonradioactive assay (Promega, Madison, WI). Absorbance measurements were taken at 490 nm.
To perform culture immunocytochemistry, microglia were plated on glass chamberslides for 24 h then fixed in 4% paraformaldehyde (37°C, 30 min) and immunostained using anti-CD68 and anti-CD11b antibodies using Vector VIP as the chromagen (Vector Laboratories, Burlingame, CA) or Texas Red secondary (Santa Cruz Biotechnology, Santa Cruz, CA) with 4′,6′-diamidino-2-phenylindole (DAPI) (Invitrogen) to visualize the nucleus. Microglial purity was determined by placing a counting grid under the wells and counting CD68-positive microglia compared with total hematoxylin (Sigma) nuclear stained cells from four identical fields per well for each condition. The average number of microglia (±SD) was calculated for each condition. Each experiment was performed in quadruplicate three to four times.
Animal care and use.
All procedures were reviewed and approved by the University of North Dakota Institutional Animal Care and Use Committee (protocol no. 0012-3). Mice were housed at the Center for Biomedical Research on a 12 h light/dark cycle and were allowed food and water ad libitum. Adult mice were killed via CO2 asphyxiation for collection of brains. Postnatal pups were killed via decapitation.
Purified adult microglia are morphologically different from postnatally derived microglia
To begin comparing primary adult and postnatal mouse microglia cultures, we first verified culture purity as well as immunoreactivity for two well established microglial markers, CD68 and CD11b (Fig. 1). Postnatal microglia were used directly after isolation from a mixed glial bed at 14 d in vitro, whereas the adult microglia were used immediately after isolation from the brain. Although both cell types were immunoreactive for CD11b and CD68, the postnatal cells displayed a dramatically different morphology than adult cells (Fig. 1). Postnatal cells demonstrated a larger diameter, flattened phenotype in comparison with adult microglia. Importantly, both culture types were of similar purity based on percent CD68 immunoreactivity (≥99% for postnatal cells and ≥97% for adult cells) (data not shown). These data verified our system for subsequent analysis.
Fibrillar Aβ stimulates TNFα secretion only from postnatal microglia
Because proinflammatory cytokines such as TNFα are associated with plaques in AD (McGeer and McGeer, 1995), and previous data have demonstrated that Aβ fibrils stimulate microglial TNFα secretion in vitro (Meda et al., 1995; Tan et al., 2000; Floden et al., 2005), we compared the ability of Aβ fibrils to stimulate TNFα secretion from postnatal versus adult microglia. Prior data have demonstrated that the oligomeric form of Aβ is also a potent microglia stimulus, suggesting that either conformation can contribute to the reactive microgliosis observed in AD (Roher et al., 1996). Therefore, we also compared the ability of oligomeric Aβ to stimulate the cells. We stimulated both cell types with increasing concentrations of either peptide for 24 h. The cells were stimulated for only 24 h to minimize the probability that any secretory changes resulted from secondary, autocrine stimulation of the cells. Although the peptides appeared toxic at highest concentrations, both Aβ oligomers and fibrils stimulated a dose-dependent, significant increase in TNFα secretion from postnatal cells (Fig. 2A). Conversely, adult cells responded in a different manner to peptide stimulation. Oligomeric peptides stimulated TNFα secretion from adult cells in a dose-dependent manner, whereas fibrils had no effect on cytokine secretion (Fig. 3A).
Adult microglia have reduced ability to phagocytose Aβ in vitro
We next determined whether adult and postnatal microglia differed in their ability to phagocytose Aβ fibrils. Significant amyloid uptake was observed in postnatal cells after 6 h of stimulation and was increased after opsonization with two different anti-Aβ antibodies (Fig. 4A). In contrast, adult microglia were unable to take up Aβ fibrils regardless of antibody opsonization (Fig. 4A). However, adult microglia phagocytosed E. coli bioparticles demonstrating specificity for the diminished Aβ response (Fig. 4B). A modest increase in Aβ uptake was observed in adult cells after opsonization and a 24 h incubation demonstrating some limited ability for adult microglia to take up the peptide (data not shown).
In this study, we demonstrate that microglia acutely isolated from adult mouse brain differ dramatically in their response to Aβ fibril stimulation when compared with microglia from standard postnatal brain culture preparations (Giulian and Baker, 1986). This is not surprising given the fact that microglia in aged human and rodent brains undergo morphologic dystrophy characterized by enlargement, deramification, and adoption of a more phagocytic phenotype (Sheng et al., 1998; Streit et al., 2004). More importantly, studies using both human and rodent adult microglial cultures have demonstrated that adult cells respond differently than postnatal cells regarding upregulation of cyclooxygenase expression (Hoozemans et al., 2002) and interleukin 6 (IL-6) and IL-1β secretion (Xie et al., 2003) after stimulation with Aβ (Hoozemans et al., 2002) and LPS (Xie et al., 2003). In contrast to these studies, we used adult microglia acutely isolated from the brain rather than implementing a prolonged culture protocol, because it is known that this stimulates conversion to an active phenotype (Becher and Antel, 1996; Slepko and Levi, 1996). It is important to point out that we cannot rule out the possibility that the phenotype of the adult cells is altered during the isolation procedure and not precisely reflective of in vivo phenotype. However, the adult cells clearly respond to oligomeric peptide stimulation and phagocytose E. coli bioparticles indicating that they are not fundamentally compromised. As already mentioned, we have not allowed a recovery period before stimulation, because prolonged culturing leads to a histological phenotype similar to that observed in postnatal cultures (data not shown), subsequent activation and therefore conditions even further removed from in vivo.
In addition to the histological differences between ages, we also observed a significant difference in stimulated secretion between the two cell types. Postnatal cultures secreted TNFα in response to stimulation with both Aβ fibrils and oligomers, whereas adult cultures increased cytokine secretion only after oligomeric peptide stimulation. Although fibril-stimulated secretion is in agreement with our previous work as well as that of others performed using postnatal derived microglia (Meda et al., 1995; Tan et al., 2000; Floden et al., 2005), our data from the adult microglia suggests that postnatal cells may not represent the most accurate experimental model for adult or aged paradigms. Interestingly, our results suggest that oligomeric peptide serves as a proinflammatory ligand independent of animal age, perhaps supporting the hypothesis that this peptide conformation represents the more relevant disease entity mediating cell death and dysfunction (Selkoe, 2002).
The difference in ability to phagocytose Aβ was another unexpected observation between culture ages. Although postnatal cells effectively phagocytosed Aβ fibrils with and without opsonizing antibodies, adult cells were unable to do this. However, the adult cells efficiently phagocytosed bacterial bioparticles. This demonstrates that the diminished uptake of Aβ was not an issue of incomplete recovery from the isolation procedure, but rather, the microglia were truly unable to phagocytose the fibrillar peptide. It is curious that opsonizing antibodies afforded little improvement in adult cell uptake, particularly in light of the in vivo demonstrations of the ability of Aβ immunization to decrease plaque load (Schenk et al., 1999; DeMattos et al., 2001; Bacskai et al., 2002; Wilcock et al., 2003). It is intriguing to speculate that microglia have reduced capacity to clear Aβ plaques in adult brains, perhaps contributing to the age dependency of plaque deposition observed during disease.
Finally, there is some evidence that microglia from AD brains have phenotypes unique to disease. Lue et al. (2001) have reported that 13–15 d in vitro cultures of microglia derived from human AD patients secrete elevated levels of IL-1β, TNFα, IL-6, and MCP-1 (monocyte chemoattractant protein 1) compared with control microglia. Others have demonstrated that peripheral blood-derived macrophages from AD patients have reduced capacity to phagocytose Aβ compared with cells from age-matched controls (Fiala et al., 2005). Together, these data illustrate that adult AD microglia appear to respond fundamentally different to Aβ stimulation compared with control cells and complement our observations of differences between aged and postnatal cells. Although still in vitro, our results suggest that studies performed using microglial cultures from aged or diseased brains offer a more relevant understanding of microglial responses to Aβ stimulation and phagocytosis.
This work was supported by National Institutes of Health–National Center for Research Resources Grant 1 P20 RR17699-01. We are grateful to Dr. M. Carson for helpful discussion and advice.
- Correspondence should be addressed to Dr. Colin K. Combs, Department of Pharmacology, Physiology, and Therapeutics, School of Medicine and Health Sciences, University of North Dakota, 504 Hamline Street, Neuroscience Building, Grand Forks, ND 58202. Email: