Inflammatory changes parallel the early stages of Alzheimer disease

doi:10.1016/j.neurobiolaging.2006.08.014

Neurobiol Aging. Author manuscript; available in PMC 2008 December 1.

Published in final edited form as:

Neurobiol Aging. 2007 December; 28(12): 1821–1833.

Published online 2006 October 18. doi: 10.1016/j.neurobiolaging.2006.08.014

PMCID: PMC2198930

NIHMSID: NIHMS33065

Inflammatory changes parallel the early stages of Alzheimer disease

A. Parachikova,¹ M.G. Agadjanyan,³ D.H. Cribbs,^1,² M. Blurton-Jones,¹ V. Perreau,¹ J. Rogers,⁴ T.G. Beach,⁴ and C.W. Cotman^1,²^*

¹ Institute for Brain Aging & Dementia, University of California, Irvine, Irvine, CA 92697-4540, USA

² Department of Neurology, College of Medicine, University of California, Irvine, Irvine, CA 92697, USA

³ Institute for Molecular Medicine, Department of Immunology, Huntington Beach, CA 92647, USA

⁴ L,J, Roberts Center, Sun Health Research Institute, Sun City, AZ 85351, USA

Corresponding Author: Anna I. Parachikova, Institute for Brain Aging & Dementia, University of California, 1113 Gillespie Neuroscience Research Facility, Irvine, CA 92697-4540, USA, Email: aparachi/at/uci.edu, (949) 824-6071, (949) 824-2071 Fax

The publisher's final edited version of this article is available at Neurobiol Aging

See other articles in PMC that cite the published article.

Publisher's Disclaimer

Abstract

Alzheimer disease (AD) is the most prominent cause of dementia in the elderly. To determine changes in the AD brain that may mediate the transition into dementia, the gene expression of approximately 10,000 full-length genes was compared in mild/moderate dementia cases to non-demented controls that exhibited high AD pathology. Including this latter group distinguishes this work from previous studies in that it allows analysis of early cognitive loss. Compared to non-demented high-pathology controls, the hippocampus of AD cases with mild/moderate dementia had increased gene expression of the inflammatory molecule major histocompatibility complex (MHC) II, as assessed with microarray analysis. MHC II protein levels were also increased and inversely correlated with cognitive ability. Interestingly, the mild/moderate AD dementia cases also exhibited decreased number of T cells in the hippocampus and the cortex compared to controls. In conclusion, transition into AD dementia correlates with increased MHC II⁺ microglia-mediated immunity and is paradoxically paralleled by a decrease in T cell number, suggesting immune dysfunction.

Keywords: MHC class II, microglia, inflammation, dementia, amyloid beta, T cell

1. Introduction

AD is a chronic neurodegenerative disorder characterized by progressive memory deterioration. To elucidate the etiology of AD dementia, attempts have been made to correlate cognitive dysfunction with the classical neuropathological changes in AD, extracellular β-amyloid (Aβ) plaques, intraneuronal neurofibrillary tangles (NFTs), and synaptic/neuronal loss. However, some studies suggest that plaques and NFTs may not directly cause AD dementia since both pathologies can manifest in brains of patients that are not cognitively impaired [10, 14, 49, 56]. Another potential mechanism underlying dementia is synaptic loss, since synaptic markers are decreased in AD brains [16, 54, 59, 62]. However, evidence suggests that synaptic loss may occur only in moderate to severe clinical grades of dementia while early AD cases exhibit an increase in presynaptic markers [44]. Synaptic deficiency also correlates with the accumulation of soluble intraneuronal Aβ in AD vulnerable brain regions [34, 45]. Thus in accordance with the amyloid cascade hypothesis [19, 20], the accumulation of soluble Aβ with age may impair synaptic pathways associated with learning and memory.

Cognitive dysfunction in AD may also be critically influenced by Aβ-induced brain inflammation [3, 13, 37, 41, 61]. Specifically, Aβ accumulation leads to a site-specific activation of glia resulting in the secretion of pro-inflammatory cytokines [3, 13]. The inflammatory response may be an attempt to clear Aβ deposits; however, the progressive accumulation of Aβ and its aggregation into insoluble plaques may induce a chronic pro-inflammatory response leading to compromised neuronal function [8]. In support of a role for neuroinflammation in AD dementia, one study found a higher correlation between synapse loss and activated microglia than between Aβ deposits and NFTs [33]. Furthermore, some epidemiological studies demonstrate that non-steroidal anti-inflammatory drugs (NSAIDs) can prevent or retard AD cognitive decline [29, 40, 52, 58], although other studies have found little improvement [1, 50]. This slowing of cognitive decline may be attributed to decreased inflammation since NSAID therapy substantially reduces the number of activated microglia associated with plaques [22, 36].

To better understand the factors contributing to AD dementia, this study aimed to identify critical changes in neuropathology and gene expression that occur during the transition into AD dementia, a very important and fundamental question in the field. Toward this goal, we compared the gene expression profiles followed by protein analyses in non-demented control patients and AD patients with mild to moderate clinical grades of dementia. Importantly, the majority of the cases comprising the control group were non-demented patients with high degree of AD-related pathology, thereby enabling us to identify factors that correlate with early cognitive decline in AD.

2. Materials and methods

2.1. Case Selection

The study employed stringent criteria for case selection based on Mini Mental State exam scores (MMSE), plaque and tangle density, sex, age, post-mortem interval (PMI), and RNA quality. Dementia severity was evaluated using the MMSE with scores of 25-30 indicating unimpaired cognition, 17-22 indicating mild/moderate dementia and less than 10 indicating severe dementia. BRAAK staging was employed to characterize pathology with stages I -II being normal to mild, stages III -IV being moderate and V - VI being severe AD [9]. Plaque pathology was determined based on Aβ load quantification in the hippocampus [15].

A group of AD patients exhibiting mild to moderate memory impairment and extensive pathology (N=10, MMSE 17-22, amyloid load 2.7-13.5%, BRAAK IV-V) was compared to a non-demented group with varied levels of pathology, including 10 cases with high plaque and/or tangle pathology (N=14, MMSE 25-30, amyloid load 0-13.5%, BRAAK I-V) (Table 1). The two groups were matched for PMI (range 1.25-5.2 hours), sex, and age. The clinical drug histories of the cases in study are unknown and therefore we have controlled for this by using high sample numbers.

Table 1

Clinical and Pathological Features of Cases

2.2. Tissue Processing

At autopsy, brains were bisected; half frozen at −80°C and half fixed in 4% paraformaldehyde. RNA was isolated from the frozen hippocampus and pre-frontal cortex (Brodmann's area 46) by the guanidium thiocyanate method [11] with TRIzol Reagent (LifeTechologies). After RNA isolation, proteins were extracted from the phenol phase following the manufacturer's recommendations (LifeTechologies).

2.3. Gene Chip

Total RNA samples were purified using RNeasy quick spin columns (Qiagen) and quantified using a UV spectrophotometer. RNA quality was assessed by Bioanalyzer on DNA 500 chips (Agilent Technologies). High quality RNA samples (mean 28S/18S = 1.4 +/- 0.3 std) were hybridized to the high-density human oligonucleotide microarray U95Av2 GeneChip (Affymetrix). Cases were processed on individual GeneChips using 5ug of total RNA. In both, the hippocampus and the cortex, we employed 10 GeneChips. An additional 14 genechips were used to validate the hippocampus data. First strand cDNA was synthesized using the SuperScript Choice System (Life Technologies) and an HPLC-purified T7-(T)24 DNA primer encoding poly(dT) and T7 RNA polymerase promoter sequence (Integrated DNA Technologies, Inc.). In vitro transcription (IVT) reactions were carried out using the Enzo High Yield RNA transcript labeling kit (Affymetrix and Enzo Diagnostics). cRNA was phenol-chloroform extracted on Phase-Lock Gels (Eppendorf), ethanol (100%) precipitated, purified using RNA easy spin columns (Qiagen) and fragmented by heating the samples to 94°C in fragmentation buffer. Biotin labeled cRNA was hybridized to the GeneChips for 16 hours, washed on a fluidics station (Affymetrix). Probe arrays were scanned on a Hewlett Packard Genearray Scanner at 540nm. The extracted image files (.dat) were imported into DChip for data analysis.

2.4. Statistical Analysis of Gene Expression Data

Two statistical methods, DChip and CyberT, were used in conjunction with GeneSpring to identify significant changes in gene expression level in cases with mild/moderate dementia compared to control. Expression values were generated in DChip employing a model-based expression algorithm and the perfect match/mismatch model (PM/MM). DChip performed a student t-test with probe sets filtering criteria of P <0.05, fold change >1.2, and a 90% confidence bound. DChip filtering incorporated an error value for each probe set allowing for the identification of poorly performing probe sets. The expression values were also imported into the Web-based CyberT program for further analysis without considering the DChip generated standard error. CyberT incorporated a Baysian algorithm and took into account the variance of 101 probe sets expressed at a similar level and used that to estimate the variance of each probe set. Filtering criteria in CyberT were P <0.01, Lnp <0.01, and fold change >1.2. CyberT is a statistical program shown to enhance the identification of genes expressed at low levels [5]. The lists of significantly changed genes generated with Dchip and CyberT were entere into GeneSpring (Silicon Genetics). Venn diagrams were employed to compare the DChip and CyberT lists of significantly altered probe sets and to generate a common list representing the genes with heightened statistical confidence and significance.

2.5. Immunohistochemistry

Plaque/tangle pathology

The hippocampus was dissected from the paraformaldehyde-fixed hemispheres stored in phosphate buffered saline (PBS) with .02% sodium azide at 4°C. Tissue blocks were sectioned at 50μm using a vibratome. Sections were double immunostained for 6E10 (Signet) and AT8 (Endogen) to identify plaque and tangle histopathology respectively. 6E10 is a monoclonal antibody reactive to amino acid residue 1-17 of the human beta amyloid (Aβ) peptide. AT8 is a monoclonal antibody which recognizes a phosphate-sensitive epitope on PHF-tau at serine residue 202.

Sections were pre-treated with formic acid (90%) for a period of 4 minutes. Following pre-treatment, sections were washed in Tris and treated with hydrogen peroxide to quench endogenous peroxidase activity [57]. Next, sections were blocked using normal goat serum and 3% bovine serum albumin (BSA) then incubated with anti-Aβ_1-17 (6E10) primary antibody (1:5000; overnight at 4°C). On the following day, sections were washed in Tris containing 0.1% TritonX100 (TBS-A) then blocked in TBS-A with 3% BSA (TBS-B) for 1 hour and placed in biotinylated secondary anti-mouse antibody (1:150; 1 hr; Vector Laboratories). After the same sequences of washes and blocking, sections were incubated with avidin-biotinylated horseradish peroxidase complex (ABC) using the vectastain Elite ABC kit (Vector Laboratories). ABC was followed by visualization using 3,3′diaminobenzidine tetrahydrochloride (DAB) reaction (Vector Laboratories). Since the double label is performed using two primary antibodies raised in the same species, before the AT8 antibody was added, sections were pre-treated in 37% formalin at 37°C for two hours then washed in Tris and treated with hydrogen peroxide [57]. Sections followed the same staining procedure as for the first label and were incubated with the AT8 primary antibody (1:500; overnight at 4°C) then the antigen was visualized by incubating in biotinylated anti-mouse secondary (Vector Laboratories) followed by an ABC step with a blue SG substrate kit (Vector Laboratories). Elimination of either primary or secondary antibodies yielded negative immunoreactivity.

MHC II and microglia

Fluorescent double-label immunohistochemistry for MHC class II and microglia employed a monoclonal CR3/43 mouse antibody against the MHC class II that recognizes HLA-DR, DP and DQ sub-regions (DAKO) and the polyclonal rabbit antibody IBA1 (WAKO) as a marker of microglia. IBA1 is a calcium-binding protein, which is restrictedly expressed in macrophages/microglia [24], and IBA1 expression is up-regulated in activated microglia [26]. Hippocampal sections from mild/moderate dementia and control cases were washed in PBS, blocked with 5% normal goat and horse serum for one hour, and incubated overnight at 4°C with IBA1 and CR3/43, both at a 1:100 dilution. On the following day, sections were washed and incubated in secondary antibody goat anti-rabbit Alexa 488 for IBA1 (Molecular Probes) and anti-mouse Alexa 555 for CR3/43 (Molecular Probes) (1:200; 1hr). Following three washes in Tris, sections were mounted and coverslipped with Fluoromount-G (Southern Biotechnology). Controls were performed to rule out nonspecific immunostaining by eliminating the primary or secondary antibody. Controls were negative.

MHC II and Aβ

Fluorescent immunohistochemistry was performed for MHC class II and Aβ_1-42 using the monoclonal CR3/43 (DAKO) and polyclonal rabbit Aβ_1-42 (BioSource) antibodies respectively on hippocampal sections from mild/moderate AD cases and controls. Briefly, sections were washed in Tris and incubated for one hour in block containing 5% normal goat and horse serum. Sections were then placed in primary antibodies CR3/43 and Aβ_1-42 diluted in block (1:100; overnight at 4°C). Following Tris rinses, sections were incubated with biotinylated anti-mouse secondary antibody (1:150; 1hr; Vector Laboratories). Lastly, sections were placed in streptavidin-conjugated Alexa 488 and anti-rabbit Alexa 555 (1:200; 1hr, Molecular Probes). Controls were performed to rule out nonspecific immunostaining by eliminating the primary or secondary antibody. Controls were negative.

T Cell

Hippocampal and cortical sections from mild/moderate dementia and control cases were immunostained for T cells using a mouse anti-human monoclonal CD3 antibody (Chemicon). Sections were washed in Tris, treated with hydrogen peroxide [57], blocked using normal goat serum and 5% bovine serum albumin (BSA) and incubated with the primary CD3 antibody (1:1000; overnight at 4°C). On day two, sections were washed and incubated with biotinylated goat anti-mouse secondary antibody (1:150; 1hr; Vector Laboratories). Following Tris washes, sections were placed in an avidin-biotinylated horseradish peroxidase complex (ABC) using the vectastain Elite ABC kit (Vector Laboratories) and visualized using the diaminobenzidine tetrahydrochloride (DAB) reaction (Vector Laboratories). Quantification of T cell number was conducted independently by three blind scorers on hippocampal sections from 12 control and 7 mild/moderate AD cases. T cell numbers were obtained by counting T cells in the entire hippocampus from sections at the same level. T cell quantification in the cortex of 8 controls and 6 mild/moderate AD cases was based on T cell counts in a 1 cm square area from comparable cortical regions. Statistical analysis employed a Student T test.

Sections from control cases were additionally stained for CD8 using the monoclonal CD8 antibody (1:100; Chemicon) to confirm T cell specificity.

Confocal and light microscopy

Light-level labeled sections were imaged on an Olympus BX61 microscope using an Olympus U-TV1X digital camera. Fluorescently labeled sections were imaged on a Bio-Rad Radiance 2100 confocal system using Argon and Krypton lasers and appropriate filter sets. Images were scanned using lambda-strobing to avoid non-specific cross-excitation or detection. Images represent either confocal Z-stacks or individual confocal slices.

2.6. Western blot analyses

The isolated protein pellet was sonicated in non-ionic detergent-based lysis buffer (BD Biosciences). Protein quantification was performed using the bicinchoninic acid (BCA) assay (BioRad). Five micrograms of total protein was resolved by SDS-PAGE and electrophoretically transferred onto nitrocellulose membrane. Membranes were blocked in 5% milk, washed, then incubated with the monoclonal CR3/43 primary antibody against HLA-DP, DQ, DR (1:100; overnight at 4°C). On the following day, the membrane was washed and incubated with HRP-conjugated goat anti-mouse secondary (1:1000; 1hr; BioRad). Membranes were visualized using enhanced chemiluminescence (Pierce). Subsequently, membranes were stripped and re-probed with a polyclonal β-actin antibody (Abcam) as a loading control. Quantification of MHC II protein levels was performed using Scion Image Software (NIH). Three measurements were taken for each band and MHC class II protein levels (monomer + dimer) were obtained relative to each β-actin or average β-actin level across all cases.

2.7. Dot blot analyses

Dot blot analyses were performed using the polyclonal rabbit CD3 antibody (Dako). Briefly, 10ug of protein sample, in equal volumes, was directly added onto the nitrocellulose membrane and left to dry. The membrane was then blocked in 5% milk and incubated with the primary CD3 antibody (1:1000; overnight at 4°C). On the second day, the membrane was washed and incubated with HRP-conjugated goat anti-rabbit secondary (1:1000; 1hr; BioRad). Membranes were visualized using enhanced chemiluminescence (Pierce). Quantification of CD3 protein levels was performed using Scion Image Software (NIH).

3. Results

We aimed to identify gene expression and neuropathological changes in mild to moderate AD dementia cases (N=10, MMSE 17-22, amyloid load 2.7-13.5%, BRAAK IV-V) compared to non-demented controls including 10 high plaque and/or tangle pathology controls (N=14, MMSE 25-30, amyloid load 0-13.5%, BRAAK I-V) (Table 1). Importantly, the groups were chosen based on MMSE scores of cognitive functioning since plaque and tangle pathology does not always correlate with memory impairment.

The low MMSE group (experimental group, N=10) consisted of clinically characterized AD cases with mild to moderate degree of dementia (MMSE 17-22). In the hippocampus, the plaque load ranged from 2.7 to 13.5% and the tangle pathology was determined to be BRAAK stage of either IV or V.

The high MMSE group (control group, N=14) consisted of clinically normal non-demented cases (MMSE 25-30) with a range of AD pathology. Specifically, 4 control cases (C5, C6, C13, C14) exhibited significant Aβ load (6.15 - 13.5%); 8 cases (C7 – C14) had significant NFTs (BRAAK III-V); 2 cases (C13, C14) were characterized by significant plaque and tangle pathology (see Table 1 for case demographics).

3.1. Changes in gene expression in mild/moderate AD dementia cases compared to non-demented controls

Gene expression in the hippocampus and pre-frontal cortex of mild/moderate AD dementia and non-demented cases was compared using high-density oligonucleotide microarrays. Genes that were differentially expressed between the two groups were identified using DChip and CyberT statistics. Genes flagged as significantly altered by both methods (DChip and CyberT) were of primary focus for further analysis since they increase statistical confidence in the data.

Regional differences in mRNA levels with mild/moderate dementia

More sequences were significantly altered in the pre-frontal cortex compared to the hippocampus in AD cases with mild to moderate dementia (MMD) relative to non-demented controls (NDC) (Figure 1A). Specifically, four times as many changes in gene expression were observed in the pre-frontal cortex compared to the hippocampus, 217 and 53 sequences respectively. A complete list of significantly altered mRNAs can be found in the appendix. In cases with mild to moderate dementia, trend analysis revealed that the majority of sequences in the pre-frontal cortex were up-regulated while most in the hippocampus were down-regulated (Figure 1B). Thus, the number of differentially expressed sequences and the direction of the change were region-specific during the early stages of dementia in AD.

Figure 1

Statistical analysis of GeneChip data in the hippocampus and pre-frontal cortex comparing mild/moderate dementia cases (MMD) to non-dementia control (NDC). (A) DChip and CyberT statistics generated lists of significantly altered sequences from the 12,000 (more ...)

Inflammatory markers are up-regulated in the early stages of AD dementia

After examining the pattern of changes at the expression level, the identity of these sequences was determined. Most sequences were functionally different but included numerous mRNAs associated with inflammation. Moreover, markers of inflammation were higher in both the pre-frontal cortex and hippocampus in cases with mild/moderate dementia compared to controls. It is important to note that while most sequences were lower in the hippocampus, those associated with inflammation were increased.

In the pre-frontal cortex of AD patients with mild to moderate dementia, MHC class I and II sequences were increased, in addition to a number of innate immune response markers including the LPS-receptor (CD14), toll-like receptor 2 and mac-1β compared to controls (Table 2). Also, anti-inflammatory sequences including transforming growth factor (TGF) β and the IL-10 receptor were higher in cases with mild/moderate dementia. In comparison, inflammation in the hippocampus was very specific, with 86% of the inflammatory markers being sequences of MHC class II haplotypes (Table 3A) compared to 14% for the pre-frontal cortex. None of the innate immune response markers or MHC class I, which were increased in the pre-frontal cortex, were differentially expressed in the hippocampus. Next, the nearly exclusive increase in MHC class II mRNA expression in the hippocampus was further characterized by examining the gene lists that were found to be significant with either CyberT- or DChip-only. The CyberT-only list showed that additional MHC class II probe sets were greater in mild/moderate dementia cases compared to controls (Table 3B).

Table 2

Inflammatory Markers in the Frontal Cortex

Table 3

Inflammatory Markers in the Hippocampus

3.2. Changes in Protein Levels

Increased MHC class II mRNA is paralleled by increased MHC class II protein in AD cases with mild/moderate dementia

Altered levels of MHC class II expression in the hippocampus were further investigated since the hippocampus is considered integral for learning and memory. To determine if the increase in MHC class II mRNA corresponded to protein changes, MHC class II protein levels were studied by Western blot. Normalized to the loading control (β-actin), MHC class II protein levels were significantly greater (P<0.01) in cases with mild to moderate dementia compared to controls including four high plaque and/or tangle-pathology controls (Figure 2A, B). The 1.8-fold greater mRNA expression corresponded to a 2.6-fold greater protein load. The regulation of protein translation or degradation may be affected in cases with AD dementia, which may account for this difference in mRNA versus protein expression levels. Furthermore, MHC class II protein levels were inversely correlated with cognitive function as measured by MMSE (R²=0.648, P<0.01, Figure 2C).

Figure 2

MHC class II protein levels in the hippocampus as assessed via Western blot in seven controls including 4 high plaque and/or tangle pathology controls and seven mild/moderate dementia cases. (A) Western Blot for CR3/43 and β-actin. (B) Bar graph (more ...)

MHC class II co-localizes with reactive microglia in the hippocampus

A highly specific up-regulation of MHC class II was noted in the hippocampus with mild to moderate dementia in contrast to a broader inflammatory response in the pre-frontal cortex. Since MHC class II expression is restricted to antigen presenting cells (APCs) including dendritic cells, B cells, and monocytes/microglia, the relation between MHC class II and microglia was examined in the hippocampus of control and early AD dementia cases. As shown in Figure 3, microglia (red) and MHC class II (green), as assessed with immunostaining for IBA1 and CR3/43 respectively, were frequently co-localized as visualized using high-resolution confocal microscopy in mild/moderate AD cases (Figure 3 A-C). These microglia appeared reactive as they were highly ramified (Figure 3D). MHC II-positive cells that were IBA1-negative were absent, suggesting that MHC class II is restricted to microglia in the hippocampus during the early stages of AD dementia (Figure 3). Control non-demented cases exhibited minimal MHC class II labeling, supportive of the Western data, as well as less intense IBA1 labeling, suggestive of diminished microglia activation (Figure 3 E-H).

Figure 3

Double immunofluorescence for microglia (red) and MHC class II (green) in the hippocampus with mild/moderate dementia (A-D) and control (E-H). The majority of microglia were double labeled in the mild/moderate AD case (C, D). The control case exhibited (more ...)

MHC class II and the adaptive immune response in AD

MHC class II molecules, which are located on the surface of APCs, function to present portions of an ingested antigen to specific T cells, thereby initiating an adaptive immune response. To determine if adaptive immunity has been initiated, it is important to demonstrate that an antigen, such as Aβ, is presented in the context of MHC class II and that T cells are present. Therefore hippocampal sections were immunostained for both MHC II (CR4/3) and Aβ_1-42 or for CD3, which is a marker of both T cell subtypes: CD8⁺ (cytotoxic) and CD4⁺ (helper). As shown in Figure 4, numerous clusters of cells which were immunopositive for both MHC II (green) and Aβ_1-42 (red) were observed in mild/moderate AD (Figure 4A-D) and high pathology control (Figure 4 E-H) cases as visualized with confocal imaging. MHC II immunostaining appeared localized to Aβ_1-42 regions in the high-pathology control whereas mild/moderate dementia cases exhibited a more widespread MHC II staining. Higher magnification imaging showed co-localization of MHC II and Aβ (Figure 4 I-L).

Figure 4

Aβ (red) and MHC class II (green) immunofluorescence in the hippocampus of mild/moderate AD (A-D) and high pathology control (E-H). Low power images show Aβ-positive and MHC class II-positive clusters. High power imaging reveals co-localization (more ...)

CD3-positive T cells were present in the hippocampus of both non-demented controls and mild/moderate dementia cases (Figure 5A, B). The T cells were distributed in a non-homogeneous fashion. Some regions showed occasional and scattered T cells while other fields showed clusters of cells, as illustrated in Figure 5. In order to obtain a measure of total T cells, we counted the number of CD3-positive cells in sections of the hippocampus. Quantification of these T cell numbers revealed that T cells are significantly decreased in cases with mild/moderate AD dementia compared to controls (P=0.04, Figure 5C). Further quantification of CD3 steady state levels in hippocampal brain homogenates of control and demented cases was assessed by dot blot (Figure 5E). Likewise to CD3-positive T cell counts, demented cases showed a significant decrease in CD3 steady state levels (P=0.03, Figure 5F). The presence of T cells in the hippocampus was further confirmed by staining for an alternative T cell marker, CD8, as shown in Figure 5D. Analyses of frontal cortex from control and mild/moderate dementia cases paralleled the hippocampal findings with decreased T cell numbers with dementia (P=0.04, Figure 5G) as counted from a 1 cm square area from comparable cortical areas. Together these results are suggestive of a dysfunctional adaptive immune response induced by Aβ or another mechanism.

Figure 5

Light level immunohistochemistry for CD3⁺T cells depicting a cluster in the hippocampus of control (A) and mild/moderate dementia cases (B). Scale bar equals 100 μm. T cell quantification in the hippocampus of 7 MMD and 12 non-demented controls (more ...)

4. Discussion

This study suggests that immune responses in the CNS may be involved in the transition to AD dementia. Our goal was to identify factors which correlate with cognitive decline by comparing non-demented controls, including high-pathology controls, to cases with mild/moderate clinical grades of AD dementia. We demonstrated that inflammatory molecules, most prominently MHC II, a marker of microglia activation, were increased in AD patients with early stage dementia. One prominent consequence of MHC II up-regulation is the potential for antigen presentation to activated T cells thereby inducing an adaptive immune response. However, mild/moderate AD dementia cases had reduced numbers of T cells compared to non-demented controls, suggesting that adaptive immune responses may be dysfunctional during the early stages of AD dementia and may contribute to the cognitive decline observed in these individuals. This work complements a number of prior studies which have also demonstrated immune responses in AD, both in the CNS [2, 7, 12, 32] and also from peripheral lymphocytes [27, 46, 51]. However as we have focused on the dementia component of AD, rather than the pathology, we highlight that this immune response may be directly involved in cognitive decline.

In the hippocampus, a primary area implicated in learning and memory, several members of the MHC class II family are significantly increased, in both mRNA and protein levels, at early stages of AD dementia and this increase is inversely correlated with MMSE scores of cognitive function. The comparison with control cases that were cognitively intact but had high AD pathology supports the notion that increased MHC class II levels are correlated to the dementia component of the disease. Previous studies have highlighted abnormal MHC II expression in AD [30, 31, 39, 64] as well as in transgenic mouse models of AD [18, 42, 60], which suggests that these effects on MHC II are mediated by increased APP fragments including Aβ. In addition, severe late stage AD dementia has been correlated with increased MHC class II levels in the brain [4, 35, 67]. The current study provides evidence showing that the increase in MHC class II occurs earlier than previously described, is already present in cases with mild to moderate AD dementia and correlates inversely with MMSE scores.

In addition to increased MHC class II molecules, two other lines of evidence presented here indicate that T cell mediated adaptive immunity may be involved in the transition to AD dementia. First, MHC class II expression was restricted to microglia cells and these cells co-localized with Aβ, suggesting that Aβ may be an antigen being presented. MHC class II expression in activated microglia associated with plaques has been previously established [53] [38, 48]. Second, the presence of T cells in the hippocampus of mild to moderate AD dementia cases, as well as controls, suggests a potential for a T cell response. Although the central nervous system (CNS) has been thought to be immunologically privileged, T cells have been shown to routinely penetrate the blood brain barrier (BBB) and enter the CNS [23, 66] and they are present in the aged human hippocampus [63] as well as the AD brain [25, 53]. We hypothesize that since in AD patients Aβ can cross the BBB and enter into the bloodstream, it is likely that the immune system can recognize this self-antigen and induce specific immune responses. In support, Monsonego and colleagues demonstrated the presence of Aβ-reactive T cells restricted to MHC class II in the brains of AD patients with moderate to severe grades of dementia [43]. These Aβ-reactive T cells may have been recruited by microglia, since microglia have been shown to serve as antigen presenting cells that stimulate T cells in vitro [6]. However, the exact role of T cells in the AD brain remains ambiguous. This ambiguity stems from evidence that T cells may actually die during inflammatory cascades [17, 47]. Monsonego and colleagues documented that T cell apoptosis, caused by nitric oxide (NO) production at sites of Aβ plaques [21, 55], occurs prior to T cell activation by microglia [43]. In the current study, we show that early stage AD dementia is associated with a decreased number of T cells in the hippocampus as compared to non-demented controls, suggesting that the T cell response in AD is dysfunctional. In addition to T cell death, lack of co-stimulatory molecules is known to induce T cell anergy. In the current study, CD80 and CD86 co-stimulatory molecules are not found to be altered at the mRNA level between non-demented control and mild/moderate dementia cases (data not shown), suggesting that proper activation of cell-mediated immunity is not taking place, although this may be due to the small sample size that T cells constitute in the brain homogenates. Therefore, it is unclear whether the antigen presentation capacity of MHC II-positive microglia cells to T cells in the AD brain is a functional process leading to Aβ removal, or if removal is inhibited by the death and/or anergy of T cells.

Importantly, there is also an established positive link between T cells and cognition. More specifically, T cell deficiency has been associated with deficits in spatial learning and memory [28]. In addition, a recent study demonstrated that CNS-specific T cells maintain neurogenesis and are required for spatial learning and memory in adulthood [68]. Immune deficient SCID mice show memory deficits and the addition of CNS-specific T cells restored neurogenesis and improved cognition [68]. Our studies show a decrease in T cells in AD. Therefore, it is interesting to speculate that decreased T cells contribute to the cognitive deficits observed in transitional AD cases as compared to non-demented controls including individuals with high plaque and/or tangle pathology.

In conclusion, increased microglia activation, paralleled by a decrease in T cell number, correlates with the initial stages of AD dementia. In particular, the progression into AD-type dementia coincides with a specific MHC class II increase in the hippocampus, the primary brain area related to learning and memory. Thus, MHC class II-positive microglia-mediated adaptive immunity may be fundamental to the cognitive decline that occurs in AD. But the findings of decreased T cells in the same brain region suggest that cognitive decline in AD is associated with immune dysfunction.

Supplementary Material

Click here to view.^{(50K, xls)}

Click here to view.^{(22K, xls)}

Acknowledgments

We would like to thank Dr. Kim Green and Dr. Danielle Simmons for the critical review of the manuscript. Also thanks to Dr. Kathryn Nichol for cell counting. The brain tissue was obtained from the Tissue Repository at the Institute for Brain Aging and Dementia and the Brain Donation Program at Sun Heath Research Institute. The Sun Health Research Institute Brain Donation Program is supported by the National Institute on Aging (P30 AG19610 Arizona Alzheimer's Disease Core Center, the Arizona Department of Health Services (contract 211002, Arizona Alzheimer's Research Center) and the Arizona Biomedical Research Commission (contract 0011, Arizona Parkinson's Disease Center). This work was supported by AG16573, AG00538, AG 023173, AG 020241 and NS 050895.

Footnotes

Disclosure Statement There is no actual or potential conflict of interest.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

1. Aisen PS, Schafer KA, Grundman M, Pfeiffer E, Sano M, Davis KL, Farlow MR, Jin S, Thomas RG, Thal LJ. Effects of rofecoxib or naproxen vs placebo on Alzheimer disease progression: a randomized controlled trial. Jama. 2003;289(21):2819–26. [PubMed]

2. Ait-Ghezala G, Mathura VS, Laporte V, Quadros A, Paris D, Patel N, Volmar CH, Kolippakkam D, Crawford F, Mullan M. Genomic regulation after CD40 stimulation in microglia: relevance to Alzheimer's disease. Brain Res Mol Brain Res. 2005;140(12):73–85. [PubMed]

3. Akiyama H, Barger S, Barnum S, Bradt B, Bauer J, Cole GM, Cooper NR, Eikelenboom P, Emmerling M, Fiebich BL, Finch CE, Frautschy S, Griffin WS, Hampel H, Hull M, Landreth G, Lue L, Mrak R, Mackenzie IR, McGeer PL, O'Banion MK, Pachter J, Pasinetti G, Plata-Salaman C, Rogers J, Rydel R, Shen Y, Streit W, Strohmeyer R, Tooyoma I, Van Muiswinkel FL, Veerhuis R, Walker D, Webster S, Wegrzyniak B, Wenk G, Wyss-Coray T. Inflammation and Alzheimer's disease. Neurobiol Aging. 2000;21(3):383–421. [PubMed]

4. Bajetto A, Bonavia R, Barbero S, Florio T, Schettini G. Chemokines and their receptors in the central nervous system. Front Neuroendocrinol. 2001;22(3):147–84. [PubMed]

5. Baldi P, Long AD. A Bayesian framework for the analysis of microarray expression data: regularized t -test and statistical inferences of gene changes. Bioinformatics. 2001;17(6):509–19. [PubMed]

6. Becher B, Prat A, Antel JP. Brain-immune connection: immuno-regulatory properties of CNS-resident cells. Glia. 2000;29(4):293–304. [PubMed]

7. Blalock EM, Geddes JW, Chen KC, Porter NM, Markesbery WR, Landfield PW. Incipient Alzheimer's disease: microarray correlation analyses reveal major transcriptional and tumor suppressor responses. Proc Natl Acad Sci U S A. 2004;101(7):2173–8. [PubMed]

8. Blasko I, Stampfer-Kountchev M, Robatscher P, Veerhuis R, Eikelenboom P, Grubeck-Loebenstein B. How chronic inflammation can affect the brain and support the development of Alzheimer's disease in old age: the role of microglia and astrocytes. Aging Cell. 2004;3(4):169–76. [PubMed]

9. Braak H, Braak E. Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol (Berl) 1991;82(4):239–59. [PubMed]

10. Chapman PF, Falinska AM, Knevett SG, Ramsay MF. Genes, models and Alzheimer's disease. Trends Genet. 2001;17(5):254–61. [PubMed]

11. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem. 1987;162(1):156–9. [PubMed]

12. Colangelo V, Schurr J, Ball MJ, Pelaez RP, Bazan NG, Lukiw WJ. Gene expression profiling of 12633 genes in Alzheimer hippocampal CA1: transcription and neurotrophic factor down-regulation and up-regulation of apoptotic and pro-inflammatory signaling. J Neurosci Res. 2002;70(3):462–73. [PubMed]

13. Cooper NR, Bradt BM, O'Barr S, Yu JX. Focal inflammation in the brain: role in Alzheimer's disease. Immunol Res. 2000;21(23):159–65. [PubMed]

14. Crystal H, Dickson D, Fuld P, Masur D, Scott R, Mehler M, Masdeu J, Kawas C, Aronson M, Wolfson L. Clinico-pathologic studies in dementia: nondemented subjects with pathologically confirmed Alzheimer's disease. Neurology. 1988;38(11):1682–7. [PubMed]

15. Cummings BJ, Cotman CW. Image analysis of beta-amyloid load in Alzheimer's disease and relation to dementia severity. Lancet. 1995;346(8989):1524–8. [PubMed]

16. DeKosky ST, Scheff SW. Synapse loss in frontal cortex biopsies in Alzheimer's disease: correlation with cognitive severity. Ann Neurol. 1990;27(5):457–64. [PubMed]

17. Ford AL, Foulcher E, Lemckert FA, Sedgwick JD. Microglia induce CD4 T lymphocyte final effector function and death. J Exp Med. 1996;184(5):1737–45. [PMC free article] [PubMed]

18. Gordon MN, Holcomb LA, Jantzen PT, DiCarlo G, Wilcock D, Boyett KW, Connor K, Melachrino J, O'Callaghan JP, Morgan D. Time course of the development of Alzheimer-like pathology in the doubly transgenic PS1+APP mouse. Exp Neurol. 2002;173(2):183–95. [PubMed]

19. Hardy J, Selkoe DJ. The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science. 2002;297(5580):353–6. [PubMed]

20. Hardy JA, Higgins GA. Alzheimer's disease: the amyloid cascade hypothesis. Science. 1992;256(5054):184–5. [PubMed]

21. Heneka MT, Wiesinger H, Dumitrescu-Ozimek L, Riederer P, Feinstein DL, Klockgether T. Neuronal and glial coexpression of argininosuccinate synthetase and inducible nitric oxide synthase in Alzheimer disease. J Neuropathol Exp Neurol. 2001;60(9):906–16. [PubMed]

22. Heneka MT, Sastre M, Dumitrescu-Ozimek L, Hanke A, Dewachter I, Kuiperi C, O'Banion K, Klockgether T, Van Leuven F, Landreth GE. Acute treatment with the PPAR{gamma} agonist pioglitazone and ibuprofen reduces glial inflammation and A{beta}1-42 levels in APPV717I transgenic mice. Brain. 2005

23. Hickey WF. Basic principles of immunological surveillance of the normal central nervous system. Glia. 2001;36(2):118–24. [PubMed]

24. Imai Y, Ibata I, Ito D, Ohsawa K, Kohsaka S. A novel gene iba1 in the major histocompatibility complex class III region encoding an EF hand protein expressed in a monocytic lineage. Biochem Biophys Res Commun. 1996;224(3):855–62. [PubMed]

25. Itagaki S, McGeer PL, Akiyama H. Presence of T-cytotoxic suppressor and leucocyte common antigen positive cells in Alzheimer's disease brain tissue. Neurosci Lett. 1988;91(3):259–64. [PubMed]

26. Ito D, Imai Y, Ohsawa K, Nakajima K, Fukuuchi Y, Kohsaka S. Microglia-specific localisation of a novel calcium binding protein, Iba1. Brain Res Mol Brain Res. 1998;57(1):1–9. [PubMed]

27. Kalman J, Kitajka K, Pakaski M, Zvara A, Juhasz A, Vincze G, Janka Z, Puskas LG. Gene expression profile analysis of lymphocytes from Alzheimer's patients. Psychiatr Genet. 2005;15(1):1–6. [PubMed]

28. Kipnis J, Cohen H, Cardon M, Ziv Y, Schwartz M. T cell deficiency leads to cognitive dysfunction: implications for therapeutic vaccination for schizophrenia and other psychiatric conditions. Proc Natl Acad Sci U S A. 2004;101(21):8180–5. [PubMed]

29. Launer L. Nonsteroidal anti-inflammatory drug use and the risk for Alzheimer's disease: dissecting the epidemiological evidence. Drugs. 2003;63(8):731–9. [PubMed]

30. Lehmann DJ, Wiebusch H, Marshall SE, Johnston C, Warden DR, Morgan K, Schappert K, Poirier J, Xuereb J, Kalsheker N, Welsh KI, Smith AD. HLA class I, II & III genes in confirmed late-onset Alzheimer's disease. Neurobiol Aging. 2001;22(1):71–7. [PubMed]

31. Liew SC, Penfold PL, Provis JM, Madigan MC, Billson FA. Modulation of MHC class II expression in the absence of lymphocytic infiltrates in Alzheimer's retinae. J Neuropathol Exp Neurol. 1994;53(2):150–7. [PubMed]

32. Loring JF, Wen X, Lee JM, Seilhamer J, Somogyi R. A gene expression profile of Alzheimer's disease. DNA Cell Biol. 2001;20(11):683–95. [PubMed]

33. Lue LF, Brachova L, Civin WH, Rogers J. Inflammation, A beta deposition, and neurofibrillary tangle formation as correlates of Alzheimer's disease neurodegeneration. J Neuropathol Exp Neurol. 1996;55(10):1083–8. [PubMed]

34. Lue LF, Kuo YM, Roher AE, Brachova L, Shen Y, Sue L, Beach T, Kurth JH, Rydel RE, Rogers J. Soluble amyloid beta peptide concentration as a predictor of synaptic change in Alzheimer's disease. Am J Pathol. 1999;155(3):853–62. [PubMed]

35. Luterman JD, Haroutunian V, Yemul S, Ho L, Purohit D, Aisen PS, Mohs R, Pasinetti GM. Cytokine gene expression as a function of the clinical progression of Alzheimer disease dementia. Arch Neurol. 2000;57(8):1153–60. [PubMed]

36. Mackenzie IR, Munoz DG. Nonsteroidal anti-inflammatory drug use and Alzheimer-type pathology in aging. Neurology. 1998;50(4):986–90. [PubMed]

37. McGeer EG, McGeer PL. The importance of inflammatory mechanisms in Alzheimer disease. Exp Gerontol. 1998;33(5):371–8. [PubMed]

38. McGeer PL, Itagaki S, Tago H, McGeer EG. Occurrence of HLA-DR reactive microglia in Alzheimer's disease. Ann N Y Acad Sci. 1988;540:319–23. [PubMed]

39. McGeer PL, Akiyama H, Itagaki S, McGeer EG. Immune system response in Alzheimer's disease. Can J Neurol Sci. 1989;16 4:516–27. [PubMed]

40. McGeer PL, McGeer EG. Anti-inflammatory drugs in the fight against Alzheimer's disease. Ann N Y Acad Sci. 1996;777:213–20. [PubMed]

41. Meda L, Cassatella MA, Szendrei GI, Otvos L, Jr, Baron P, Villalba M, Ferrari D, Rossi F. Activation of microglial cells by beta-amyloid protein and interferon-gamma. Nature. 1995;374(6523):647–50. [PubMed]

42. Middleton D, Mawhinney H, Curran MD, Edwardson JA, Perry R, McKeith I, Morris C, Ince PG, Neill D. Frequency of HLA-A and B alleles in early and late-onset Alzheimer's disease. Neurosci Lett. 1999;262(2):140–2. [PubMed]

43. Monsonego A, Imitola J, Zota V, Oida T, Weiner HL. Microglia-mediated nitric oxide cytotoxicity of T cells following amyloid beta-peptide presentation to Th1 cells. J Immunol. 2003;171(5):2216–24. [PubMed]

44. Mukaetova-Ladinska EB, Garcia-Siera F, Hurt J, Gertz HJ, Xuereb JH, Hills R, Brayne C, Huppert FA, Paykel ES, McGee M, Jakes R, Honer WG, Harrington CR, Wischik CM. Staging of cytoskeletal and beta-amyloid changes in human isocortex reveals biphasic synaptic protein response during progression of Alzheimer's disease. Am J Pathol. 2000;157(2):623–36. [PubMed]

45. Oddo S, Caccamo A, Shepherd JD, Murphy MP, Golde TE, Kayed R, Metherate R, Mattson MP, Akbari Y, LaFerla FM. Triple-transgenic model of Alzheimer's disease with plaques and tangles: intracellular Abeta and synaptic dysfunction. Neuron. 2003;39(3):409–21. [PubMed]

46. Palotas A, Puskas LG, Kitajka K, Palotas M, Molnar J, Pakaski M, Janka Z, Penke B, Kalman J. Altered response to mirtazapine on gene expression profile of lymphocytes from Alzheimer's patients. Eur J Pharmacol. 2004;497(3):247–54. [PubMed]

47. Pender MP, Rist MJ. Apoptosis of inflammatory cells in immune control of the nervous system: role of glia. Glia. 2001;36(2):137–44. [PubMed]

48. Perlmutter LS, Scott SA, Barron E, Chui HC. MHC class II-positive microglia in human brain: association with Alzheimer lesions. J Neurosci Res. 1992;33(4):549–58. [PubMed]

49. Polvikoski T, Sulkava R, Myllykangas L, Notkola IL, Niinisto L, Verkkoniemi A, Kainulainen K, Kontula K, Perez-Tur J, Hardy J, Haltia M. Prevalence of Alzheimer's disease in very elderly people: a prospective neuropathological study. Neurology. 2001;56(12):1690–6. [PubMed]

50. Reines SA, Block GA, Morris JC, Liu G, Nessly ML, Lines CR, Norman BA, Baranak CC. Rofecoxib: no effect on Alzheimer's disease in a 1-year, randomized, blinded, controlled study. Neurology. 2004;62(1):66–71. [PubMed]

51. Rensink AA, Gellekink H, Otte-Holler I, ten Donkelaar HJ, de Waal RM, Verbeek MM, Kremer B. Expression of the cytokine leukemia inhibitory factor and pro-apoptotic insulin-like growth factor binding protein-3 in Alzheimer's disease. Acta Neuropathol (Berl) 2002;104(5):525–33. [PubMed]

52. Rich JB, Rasmusson DX, Folstein MF, Carson KA, Kawas C, Brandt J. Nonsteroidal anti-inflammatory drugs in Alzheimer's disease. Neurology. 1995;45(1):51–5. [PubMed]

53. Rogers J, Luber-Narod J, Styren SD, Civin WH. Expression of immune system-associated antigens by cells of the human central nervous system: relationship to the pathology of Alzheimer's disease. Neurobiol Aging. 1988;9(4):339–49. [PubMed]

54. Scheff SW, Price DA. Synaptic pathology in Alzheimer's disease: a review of ultrastructural studies. Neurobiol Aging. 2003;24(8):1029–46. [PubMed]

55. Smits HA, de Vos NM, Wat JW, van der Bruggen T, Verhoef J, Nottet HS. Intracellular pathways involved in TNF-alpha and superoxide anion release by Abeta(1-42)-stimulated primary human macrophages. J Neuroimmunol. 2001;115(12):144–51. [PubMed]

56. Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. Brain infarction and the clinical expression of Alzheimer disease. The Nun Study. Jama. 1997;277(10):813–7. [PubMed]

57. Sparks DL, Liu H, Scheff SW, Coyne CM, Hunsaker JC., 3rd Temporal sequence of plaque formation in the cerebral cortex of non-demented individuals. J Neuropathol Exp Neurol. 1993;52(2):135–42. [PubMed]

58. Stewart WF, Kawas C, Corrada M, Metter EJ. Risk of Alzheimer's disease and duration of NSAID use. Neurology. 1997;48(3):626–32. [PubMed]

59. Sze CI, Troncoso JC, Kawas C, Mouton P, Price DL, Martin LJ. Loss of the presynaptic vesicle protein synaptophysin in hippocampus correlates with cognitive decline in Alzheimer disease. J Neuropathol Exp Neurol. 1997;56(8):933–44. [PubMed]

60. Szpak GM, Lechowicz W, Lewandowska E, Bertrand E, Wierzba-Bobrowicz T, Gwiazda E, Schmidt-Sidor B, Dymecki J. Neurones and microglia in central nervous system immune response to degenerative processes. Part 1: Alzheimer's disease and Lewy body variant of Alzheimer's disease. Quantitative study. Folia Neuropathol. 2001;39(3):181–92. [PubMed]

61. Tan SC, Patel BK, Jackson SH, Swift CG, Hutt AJ. Ibuprofen stereochemistry: double-the-trouble? Enantiomer. 1999;4(34):195–203. [PubMed]

62. Terry RD, Masliah E, Salmon DP, Butters N, DeTeresa R, Hill R, Hansen LA, Katzman R. Physical basis of cognitive alterations in Alzheimer's disease: synapse loss is the major correlate of cognitive impairment. Ann Neurol. 1991;30(4):572–80. [PubMed]

63. Togo T, Akiyama H, Iseki E, Kondo H, Ikeda K, Kato M, Oda T, Tsuchiya K, Kosaka K. Occurrence of T cells in the brain of Alzheimer's disease and other neurological diseases. J Neuroimmunol. 2002;124(12):83–92. [PubMed]

64. Tooyama I, Kimura H, Akiyama H, McGeer PL. Reactive microglia express class I and class II major histocompatibility complex antigens in Alzheimer's disease. Brain Res. 1990;523(2):273–80. [PubMed]

65. Vallat-Decouvelaere AV, Chretien F, Gras G, Le Pavec G, Dormont D, Gray F. Expression of excitatory amino acid transporter-1 in brain macrophages and microglia of HIV-infected patients. A neuroprotective role for activated microglia? J Neuropathol Exp Neurol. 2003;62(5):475–85. [PubMed]

66. Wekerle H. T-cell autoimmunity in the central nervous system. Intervirology. 1993;35(14):95–100. [PubMed]

67. Xiang Z, Haroutunian V, Ho L, Purohit D, Pasinetti GM. Microglia activation in the brain as inflammatory biomarker of Alzheimer's disease neuropathology and clinical dementia. Dis Markers. 2006;22(12):95–102. [PubMed]

68. Ziv Y, Ron N, Butovsky O, Landa G, Sudai E, Greenberg N, Cohen H, Kipnis J, Schwartz M. Immune cells contribute to the maintenance of neurogenesis and spatial learning abilities in adulthood. Nat Neurosci. 2006;9(2):268–75. [PubMed]