Myeloid differentiation factor 88-deficient bone

doi:10.1093/brain/awq325

Brain 2011: 134; 278–292

| 278

BRAIN A JOURNAL OF NEUROLOGY

Myeloid differentiation factor 88-deficient bone marrow cells improve Alzheimer’s disease-related symptoms and pathology Wenlin Hao,1,2,* Yang Liu,1,2,* Shirong Liu,1,2 Silke Walter,1,2 Marcus O. Grimm,1,2 Amanda J. Kiliaan,3 Botond Penke,4 Tobias Hartmann,1,2 Claudia E. Ru¨be,5 Michael D. Menger6 and Klaus Fassbender1,2

*These authors contributed equally to this work. Correspondence to: Dr Yang Liu, Department of Neurology, University of the Saarland, Kirrberger Straße, 66421 Homburg/Saar, Germany E-mail: [email protected]

Alzheimer’s disease is characterized by extracellular deposits of amyloid b peptide in the brain. Increasing evidence suggests that amyloid b peptide injures neurons both directly and indirectly by triggering neurotoxic innate immune responses. Myeloid differentiation factor 88 is the key signalling molecule downstream to most innate immune receptors crucial in inflammatory activation. For this reason, we investigated the effects of myeloid differentiation factor 88-deficient bone marrow cells on Alzheimer’s disease-related symptoms and pathology by establishing bone marrow chimeric amyloid b peptide precursor transgenic mice, in which bone marrow cells differentiate into microglia and are recruited to amyloid b peptide deposits. We observed that myeloid differentiation factor 88-deficient bone marrow reconstruction reduced both inflammatory activation and amyloid b peptide burden in the brain. In addition, synaptophysin, a marker of neuronal integrity, was preserved and the expression of neuronal plasticity-related genes, ARC and NMDA-R1, was increased. Thus, myeloid differentiation factor 88-deficient microglia significantly improved the cognitive function of amyloid b peptide precursor protein transgenic mice. Myeloid differentiation factor 88-deficiency enhanced amyloid b peptide phagocytosis by microglia/macrophages and blunted toxic inflammatory activation. Both the expression of amyloid b peptide precursor protein and amyloid b peptide degrading enzymes and also the efflux of amyloid b peptide from brain parenchyma were unaffected by myeloid differentiation factor 88-deficient microglia. By contrast, the activity of b-secretase was increased. b-Secretase is expressed primarily in neurons, with relatively little expression in astrocytes and microglia. Therefore, microglial replenishment with myeloid differentiation factor 88-deficient bone marrow cells might improve cognitive functions in Alzheimer’s disease mouse models by enhancing amyloid b peptide phagocytosis and reducing inflammatory activation. These results could offer a new therapeutic option that might delay the progression of Alzheimer’s disease.

Received April 30, 2010. Revised September 8, 2010. Accepted September 19, 2010. Advance Access publication November 28, 2010 ß The Author (2010). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: [email protected]

Downloaded from brain.oxfordjournals.org by guest on December 31, 2010

1 Department of Neurology, University of the Saarland, 66421 Homburg/Saar, Germany 2 German Institute for Dementia Prevention (DIDP), University of the Saarland, 66421 Homburg/Saar, Germany 3 Department of Anatomy and Department of Cognitive Neuroscience, Donders Institute for Brain, Cognition, and Behavior, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands 4 Department of Medical Chemistry, Albert Szent Gyorgyi Medical University, Szeged, Hungary 5 Department of Radiation Therapy and Radiation Oncology, University of the Saarland, 66421 Homburg/Saar, Germany 6 Institute for Clinical and Experimental Surgery, University of the Saarland, 66421 Homburg/Saar, Germany

MyD88-deficiency improves AD pathology

Brain 2011: 134; 278–292

| 279

Keywords: Alzheimer’s disease; MyD88; microglia; inflammation; phagocytosis Abbreviations: Ab = amyloid b peptide; Arc = activity-regulated cytoskeleton-associated protein; MyD88 = myeloid differentiation factor 88; TNF = tumour necrosis factor

Introduction

Materials and methods Mice TgCRND8 Ab precursor protein transgenic mice in a C57BL6/129 background (continuously interbred for over three generations) were provided by D. Westaway (University of Toronto; Chishti et al., 2001). APPswe/PS1dE9 mice were obtained from D. Borchelt (University of Florida; Jankowsky et al., 2001) and kept in a C57BL/6 background (Hooijmans et al., 2009). MyD88 knockout (MyD88 / ) mice in a C57BL/6 background were provided by S. Akira (Osaka University). Green fluorescent protein (GFP) transgenic (C57BL/6-Tg UBC-GFP 30Scha/J) and C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, USA). All animal experiments were approved by the regional ethical committee in Saarland, Germany.

Cell isolation and culture Bone marrow cells were isolated from medullar cavities of the tibia and femur of 8-week-old female MyD88–/–, C57BL/6 or green fluorescent protein-transgenic mice. For bone marrow transplantation, cells were washed and suspended in ice-cold phosphate buffered saline. Bone marrow-derived macrophages were cultured in Dulbecco’s modified Eagle medium (Invitrogen, Karlsruhe, Germany) supplemented with 10% foetal calf serum (PAN Biotech, Aidenbach, Germany) and 20% L929 cell-conditioned medium (DSMZ, Braunschweig, Germany) in 75 cm2 flasks (BD, Heidelberg, Germany) according to a published protocol (Takahashi et al., 2007).

Preparation of Ab peptides The 42-amino acid form of human Ab (Ab42) was provided by B. Penke. Fluorescent Ab was prepared by mixing fluorescein isothiocyanate-conjugated Ab42 (Bachem, Heidelberg, Germany) and unlabelled Ab42 at a ratio of 1:4. Aggregated Ab was prepared using an established method (Dahlgren et al., 2002). Endotoxin concentrations of peptide samples were 50.01 EU/ml as determined by the Limulus amoebocyte lysate (LAL) assay (Lonza, Basel, Switzerland). The oligomeric conformations of fluorescein isothiocyanate-conjugated and non-conjugated forms of Ab42 peptides were the same as confirmed by western blot and electron microscopy.


Growing evidence suggests that inflammatory processes driven by microglia/macrophages contribute to the pathogenesis of Alzheimer’s disease (Akiyama et al., 2000; Wyss-Coray, 2006). Increased expression of inflammatory mediators has been observed in post-mortem Alzheimer’s disease brains (Eikelenboom and Stam, 1982; McGeer et al., 1988) and positron emission tomography studies have shown an association between microglial activation and Alzheimer’s disease progression (Cagnin et al., 2001; Edison et al., 2008; Okello et al., 2009). Moreover, epidemiological studies consistently link the use of non-steroidal anti-inflammatory drugs with reduced risk for later Alzheimer’s disease (McGeer et al., 1996; in t’ Veld et al., 2001; Szekely et al., 2004) and there is evidence that the protective benefit of non-steroidal anti-inflammatory drugs arises from their anti-inflammatory effects (Szekely et al., 2008), apart from their possible inhibitory effects on the amyloid b peptide (Ab) processing (Weggen et al., 2001; Eriksen et al., 2003). In the animal model of Alzheimer’s disease that over-expresses Ab precursor protein, activation and recruitment of microglia can be visualized around Ab deposits even before neuronal damage occurs (Benzing et al., 1999; McGowan et al., 1999; Bolmont et al., 2008; Meyer-Luehmann et al., 2008), further arguing for a role of neuroinflammation in progressive neurodegeneration in Alzheimer’s disease. Recent studies, however, challenge the view that microglia play a solely detrimental role in Alzheimer’s disease. Paradoxically, microglia also exert beneficial effects in Alzheimer’s disease pathophysiology. Microglia can clear Ab plaques via phagocytosis (Liu et al., 2005; Simard et al., 2006; Town et al., 2008), especially when therapeutically enhanced by vaccination against Ab (Bard et al., 2000, McLaurin et al., 2002; Koenigsknecht-Talboo et al., 2008). It has even been argued that defective Ab plaque clearance by mononuclear phagocytes might be a mechanism for the most common, ‘sporadic’ form of Alzheimer’s disease (Fiala et al., 2007). As key, innate immune-effector cells of the central nervous system, microglia are armed with an extensive repertoire of pattern recognition receptors, including CD14 and toll-like receptors that are essential for detection and rapid elimination of invading microorganisms (Akira et al., 2001; Bsibsi et al., 2002). Upon ligand binding, CD14/toll-like receptor signals are transduced via a common adaptor protein, myeloid differentiation factor 88 (MyD88), resulting in activation of AP-1 and NF-B transcription factors that regulate the expression of numerous important immune response genes (Akira et al., 2001). We and other researchers, have shown that CD14, toll-like receptor 2 and toll-like receptor 4 on microglia interact with aggregated Ab, thereby triggering microglial inflammatory activation and stimulating microglia-mediated clearance of Ab (Fassbender et al., 2004;

Liu et al., 2005; Tahara et al., 2006; Jana et al., 2008; Richard et al., 2008; Reed-Geaghan et al., 2009). Thus MyD88, the molecule that controls signalling for these important innate immune receptors, could be a key determinant for the detrimental and beneficial effects of microglia in Alzheimer’s disease pathophysiology. In this study, we created MyD88-deficient bone marrow chimeric Ab precursor protein transgenic mice to examine the role of the key, innate immune protein, MyD88, on symptoms and pathological features of Alzheimer’s disease as a potentially relevant therapeutic target in this disease.

280

| Brain 2011: 134; 278–292

Flow cytometric analysis of fluorescein isothiocyanate-Ab42 internalization in primary macrophages Primary bone marrow-derived macrophages were cultured in a 24-well plate (BD) at a density of 2 105 cells per well and treated with fluorescein isothiocyanate-conjugated Ab42 (0.5, 5 or 10 mM) for 0, 1, 3, 6 and 24 h. Macrophages were washed and detached. The mean fluorescence intensity of fluorescein isothiocyanate-labelled Ab42 was immediately determined by BD FACSCantoTM II flow cytometry (Franklin Lakes, USA). In order to monitor the inflammatory activation during Ab internalization, supernatants were collected for ELISA detection of tumour necrosis factor (TNF-) using a kit from R&D Systems (Wiesbaden, Germany).

Bone marrow transplantation

Barnes maze test The Barnes circular maze is a brightly lit open disk with 40 holes in the perimeter. Mice are motivated using distal cues to escape from the maze by finding the one hole in 40 that leads to a small dark recessed chamber (Barnes, 1979; Mayford et al., 1996). The test involved five days of acquisition training with two trials per day and was performed according to the published protocol (O’Leary and Brown, 2009). In order to reduce the stress for mice, no additional aversive stimuli were given. For each trial, latency to enter the escape hole and distance travelled were recorded by EthoVisionÕ XT (V6.1) tracking software (Noldus Information Technology, Wageningen, The Netherlands). The experimenter, blind to the genotype of bone marrow cells, recorded errors, which were defined as searches of any hole that did not have the chamber beneath it. Searches included nose pokes and head deflections over the hole. Finally, the latency, total distance and number of errors were averaged from the two trials per day.

Tissue collection Animals were sacrificed three months after bone marrow transplantation by inhalation of isofluorane (Abbott, Wiesbaden, Germany). Whole blood was collected via intracardial puncture and kept in ethylenediaminetetraacetic acid-containing Eppendorf tubes. Mice were then rapidly perfused transcardially with ice-cold phosphate buffered saline and the brain removed and divided. The left hemi-brain was immediately fixed in 4% paraformaldehyde (Sigma) for immunohistochemistry. A 0.5 mm thick piece of tissue was sagittally cut from the right hemi-brain and homogenized in Trizol (Invitrogen) for RNA isolation. The remainder of the right hemi-brain was used to prepare a

single cell suspension or snap frozen in liquid nitrogen for biochemical analysis.

Flow cytometric analysis of blood cell lineages after bone marrow reconstruction To examine the degree of chimerism of bone marrow cells, green fluorescent protein-positive white blood cells were directly detected by flow cytometry after green fluorescent protein-transgenic bone marrow reconstruction. To identify cell types after MyD88-deficient and wild-type bone marrow reconstruction, the blood was co-stained with fluorescein isothiocyanate-conjugated rat anti-mouse CD45 (30-F11) and Alexa647-conjugated rat anti-mouse CD11b (M1/70), PE-conjugated rat anti-mouse Ly-6G (1A8) or rat anti-T cell antigen receptor b chain (H57-597) antibodies, all of which were from Pharmingen, Heidelberg, Germany. After 1 h incubation on ice, the blood cells were treated with BD FACS Lysing Solution (Pharmingen) and measured with flow cytometry.

Flow cytometric analysis of CD45-positive microglia in the brain Hippocampal and cortical tissues including the grey and white matter were carefully dissected to prepare a single cell suspension using a 24 gauge needle, 2 ml syringe and 70 mm mesh (BD) according to the published protocol (Babcock et al., 2008; Hagihara et al., 2009). After pelleting cells by centrifugation, 200 ml of blocking buffer containing 25 mg/ml rat anti-mouse CD16/CD32 antibody (2.4G2, Pharmingen), 10% foetal calf serum and 0.1% sodium azide was added to prevent non-specific binding. Cells stained with Alexa647-conjugated CD11b antibody or isotype controls were used to identify the CD11b-positive cell population. After side scatter and CD11b gating, positive staining with fluorescein isothiocyanateconjugated CD45 antibody was determined based on fluorescence levels compared with the relevant isotype control.

Brain homogenization and Ab/Ab precursor protein ELISA The brain was homogenized as previously described (Maier et al., 2008). Briefly, frozen hemispheres were bounce-homogenized in a carbonate buffer (sodium carbonate buffer: 100 mM Na2CO3, 50 mM NaCl, pH 11.5) containing protease inhibitor cocktail (Roche Applied Science, Mannheim, Germany) and centrifuged at 16 000g for 30 min at 4 C. The supernatant (sodium carbonate buffer-soluble fraction) was collected and stored at 80 C. The pellets were resuspended in the sodium carbonate buffer plus 1% Triton-X (sodium carbonate buffer plus Triton), sonicated for 5 min in a 4 C water bath and centrifuged at 16 000g for another 30 min at 4 C. The supernatant was collected and stored at 80 C as the sodium carbonate buffer plus Triton-soluble fraction. The pellets were extracted for a third time using an ice-cold guanidine buffer (5 M guanidine-HCl/ 50 mM Tris, pH 8.0) to produce the guanidine-soluble fraction. Protein concentrations of all samples were measured using the Bio-Rad Protein Assay (Bio-Rad Laboratories, Munich, Germany). Ab42/40 and Ab precursor protein ELISA kits (Invitrogen) were used to determine Ab concentrations in three different fractions of brain homogenates and in plasma and Ab precursor protein concentrations in the sodium carbonate buffer plus Triton-soluble fraction. In each


On the day of transplantation, 3-month-old TgCRND8 mice or 6- or 9-month-old APPswe/PS1dE9 mice (recipients) were exposed to a 10 Gy total-body irradiation given as split doses of 2 5 Gy with a 4 h interval using a linear accelerator ( -source). Donor bone marrow cells (1 107 per mouse) derived from MyD88 / , C57BL6 or green fluorescent protein mice were then administered via the tail vein to each recipient. Transplanted mice were housed in autoclaved cages and treated with antibiotics in drinking water (0.2 mg/ml trimethoprim and 1 mg/ml sulphamethoxazole, both from Sigma, Schnelldorf, Germany) for three weeks.

W. Hao et al.

MyD88-deficiency improves AD pathology case except plasma, the result was normalized to the protein concentration of the sample.

b-Secretase 1 activity assay The activity of b-secretase 1 was quantified in sodium carbonate buffer plus Triton soluble brain homogenate using SensiZyme b-secretase 1 Activity Assay Kit (Sigma). In this assay, b-secretase 1 in the tissue extract is caught by a b-secretase 1-specific antibody coated on a 96-well plate. The b-secretase 1 activity was expressed as nanogram per millilitre relative to the pure b-secretase 1 protein standard.

Evaluation of synaptophysin protein levels

| 281

serial sections from each mouse were co-stained with anti-green fluorescent protein (Abcam), anti-Iba-1 or anti-Ab, and then Alexa488-, Cy3- or Cy5-conjugated second antibodies. Under confocal microscopy (QLC-100, Visitron Systems, Puchheim, Germany) with 20 objective, Ab deposits labelled with Cy5 were imaged after excitation with a 635 nm laser. Thereafter, Z-stack serial scanning from 5 to + 5 mm was performed under the excitation of 488, 561 and 635 nm lasers. From each section, five randomly chosen areas were analysed. The images of Alexa488 (green fluorescent protein), Cy3 (Iba-1) and Cy5 (Ab) from the same lay were colour-coded and merged with ImageJ (http://rsb.info.nih .gov/ij/) and Adobe Photoshop (V9.0.2, Adobe Systems Incorporated, San Jose, USA). The total number of Iba-1-positive cells, the number of Iba-1-positive cells co-localizing with Ab deposits and the number of green fluorescent protein-positive cells were counted. To identify the cellular origin of b-secretase 1 protein, the MyD88-deficient bone marrow chimeric Ab precursor protein transgenic mouse brain was co-stained with rabbit anti-b-secretase 1 antibody (anti-B690, Merck KGaA, Darmstadt, Germany; Yan et al., 2001), and with antibodies against different cellular markers: mouse anti-NeuN (for neurons, clone A60, Millipore), mouse anti-MHC II (for microglia, clone OX6, Abcam) and mouse anti-GFAP (for astrocyte, clone 2A5, Abcam). Relevant Alexa488 or Cy3-conjugated second antibodies were used. The b-secretase 1 expression was analysed under confocal microscopy.

Immunohistochemistry Paraformaldehyde-fixed brains were embedded in paraffin and serial 2 mm thick sagittal sections were mounted on glass slides. Immunohistochemical staining was performed on these sections with the VectaStain Elite ABC kit (Vector Laboratories, Burlingame, USA) to demonstrate the neuropathological changes. The following primary antibodies were used: rabbit anti-Iba-1 (1:500, Wako Chemicals, Neuss, Germany) and mouse anti-Ab antibody (1:50, clone 6F/3D, Dako, High Wycombe, UK). Diaminobenzadine (Sigma) was used as the chromogen. All images were acquired by Olympus AH-2 light microscopy. To quantify Ab deposits by morphometry, we used computerassisted image analysis capable of colour segmentation and automation via programmable macros (AnalySIS, Soft Imaging System, Mu¨nster, Germany). Briefly, for each animal, the percentage area occupied by immunoreactivity of Ab was calculated for four equidistant sections of total hippocampus area and one visual field of the whole cortex dorsal to the hippocampus (mid-frontal cortex, 1.5 mm2), with each section separated by a 100 mm interval. The threshold in individual red-green-blue channels for detection of immunoreactivity was kept constant for the analysis of an entire series of sections. For semiquantitative assessment of microglial cell density, only Iba-1 positive cells with clear haematoxylin nucleus staining were counted. Data were reported as the number of Iba-1-labelled cells in the total hippocampus divided by the full area (mm2). All immunohistochemical analyses described above were performed by an experimenter blinded to the genotype of reconstructed bone marrow cells.

Confocal microscopy To investigate the relationship between bone marrow cells, resident microglia and Ab deposits, four equidistant (100 mm interval)

Quantitative reverse transcriptionpolymerase chain reaction for analysis of gene transcripts Total RNA was isolated from the brain homogenate in Trizol. Reverse transcription and quantitative polymerase chain reaction on the 7500 Fast real-time polymerase chain reaction system (Applied Biosystems, Foster City, USA) were performed using our established protocol (Liu et al., 2009). TaqmanÕ gene expression assays obtained from Applied Biosystems were used to measure 18s RNA and transcripts of the following genes: TNF-, interleukin-1 (IL-1 ), inducible nitric oxide synthase (iNOS), chemokine (C-C motif) ligand 2 (CCL2), interleukin-4 (IL-4), interleukin-10 (IL-10), neprilysin (NEP), insulin-degrading enzyme (IDE), scavenger receptor A (SR-A), CD36, receptor for advanced glycation end Products (RAGE), low-density lipoprotein receptor-related protein 1 (LRP1), activity-regulated cytoskeleton-associated protein (Arc), NMDA receptor subunits NR1 (NMDA-R1), and -secretase 1.

Statistics Data shown in the figures are presented as mean SEM. For multiple comparisons, we used one-way or two-way ANOVA followed by Bonferroni’s, Tukey’s Honestly Significant Difference or Tamhane’s T2 post hoc test (dependent on the result of Levene’s test to determine the equality of variances). Two-independent-samples t-test was used to compare means for two groups of cases. All statistical analysis was performed on Statistical Package for the Social Sciences 15.0 for Windows (SPSS, Chicago, USA). Statistical significance was set at P 5 0.05.


Synaptophysin levels in sodium carbonate buffer plus triton-soluble brain homogenates were quantified with a capture ELISA as described previously (Oakley et al., 2006). A SY38 monoclonal synaptophysin antibody (1.75 mg/ml; Millipore, Schwalbach/Ts, Germany) was used to coat plates and a rabbit polyclonal synaptophysin antibody (0.125 mg/ml, Abcam, Cambridge, USA) was used to detect synaptophysin. Because purified synaptophysin was not available, the results are shown in optical density (OD) values adjusted by protein concentration in the same sample.

Brain 2011: 134; 278–292

282

| Brain 2011: 134; 278–292

W. Hao et al.

Results Bone marrow-derived cells are recruited to cerebral Ab deposits in bone marrow chimeric Ab precursor protein transgenic mice

Figure 1 MyD88-deficiency does not impair bone marrow reconstruction and bone marrow-derived cells are recruited to Ab deposits and differentiate into microglia. (A) The health status of recipient Ab precursor protein transgenic mice was monitored for 14 weeks after bone marrow transplantation. The body weight (BW) of mice receiving MyD88-deficient and wild-type bone marrow cells increased with the same tendency (n 5 12 per group). (B) Percentages of CD11b + , T cell antigen receptor + and Ly6G + cells in the CD45 + population as measured by flow cytometry were not different between these two mouse groups (t-test, P 4 0.05, n 5 12 per group). In order to track bone marrow-derived cells, some TgCRND8 Ab precursor protein transgenic mice were transplanted with donor cells derived from green fluorescent protein-transgenic mice. The location relevant to Ab and differentiation of bone marrowderived cells in the brain parenchyma was investigated by immunofluorescent staining with antibodies against green fluorescent protein, Iba-1 and Ab (C). Brain sections were analysed by confocal microscopy. Co-localization of green fluorescent protein and Iba-1 was shown by yellow fluorescence (arrow) and co-localization of all three molecules by white fluorescence (arrow head), superimposing fluorescent images of green fluorescent protein (green), Iba-1 (red) and Ab (blue). wt = wild-type; ko = MyD88-deficient.

bone marrow-reconstructed Ab precursor protein transgenic mice, we did not detect IgG leakage into the brain tissue, which indicated the intactness of blood–brain barrier (data not shown).


Bone marrow-derived cells are able to migrate to the brain and specifically differentiate into microglia (Hickey and Kimura 1988; Priller et al., 2001; Simard et al., 2006), especially under the conditions of diseases or direct irradiation (Mildner et al., 2007). In order to investigate the pathogenic effect of MyD88 in microglia, we reconstructed MyD88-deficient bone marrow in Alzheimer’s mice. After bone marrow transplantation, we observed that both groups of TgCRND8 Ab precursor protein transgenic mice receiving MyD88-deficient or wild-type cells showed the same tendency to increase in body weight (Fig. 1A). Similarly, the percentages of different cell lineages [CD11b + , T-cell antigen receptor (TCR) + and Ly6G + cells] in the white blood cells were not altered by MyD88-deficient bone marrow reconstruction (Fig. 1B). To examine the degree of chimerism of bone marrow cells, we transplanted some of our TgCRND8 Ab precursor protein transgenic mice with donor cells derived from green fluorescent protein transgenic mice. Three months after bone marrow transplantation, 72.45 3.51% (n = 5) of circulating monocytes expressed green fluorescent protein as measured by flow cytometry, which was not different from the green fluorescent protein-expression level in pooled monocytes (69.78%) from three green fluorescent protein-transgenic mice serving as bone marrow donors. As shown in Fig. 1C, bone marrow-derived cells (green fluorescent protein-positive) were observed to cross the blood–brain barrier and to be recruited to Ab deposits in the brain parenchyma. Most bone marrow-derived cells in the perivascular area and parenchyma expressed Iba-1, a marker for microglia/ macrophages, confirming the differentiation of bone marrow cells into microglial cells (Simard et al., 2006). Bone marrow-derived cells within the blood vessel remained Iba-1 negative (data not shown). We further observed 27.9 5.8% (n = 5) of whole microglial population around Ab deposits was bone marrow derived. All five TgCRND8 Ab precursor protein transgenic mice, which received phosphate buffered saline instead of donor bone marrow cells after whole-body irradiation with 10 Gy, died within two weeks. To exclude the possibility that MyD88-deficient bone marrow reconstruction has a general effect on mouse survival, we reconstructed MyD88-deficient and wild-type bone marrows in wild-type littermates of Ab precursor protein transgenic mice. We did not observe any differences between the two mouse groups in body weight or longevity, nor did we observe clear signs of graft versus host rejection in any of our experiments. To exclude the possibility that whole-body irradiation severely damages the blood–brain barrier, we performed immunohistochemical staining of mouse IgG in the brain. In both MyD88-deficient and wild-type


Brain 2011: 134; 278–292

| 283

MyD88-deficient bone marrow reconstruction improves cognitive function of Ab precursor protein transgenic mice

MyD88-deficient bone marrow reconstruction attenuates neuroinflammation in Ab precursor protein transgenic mice As inflammatory activation contributes to Alzheimer’s disease pathogenesis, we investigated whether MyD88-deficient bone marrow reconstruction reduces inflammatory activities in the brain. We measured recruitment of microglial cells in the

Figure 2 MyD88-deficient bone marrow-derived cells improve neuronal functions in Ab precursor protein transgenic mice. The cognitive function was determined with Barnes maze. TgCRND8 Ab precursor protein transgenic mice [three months post-transplantation with wild-type (wt) bone marrow] spent more time and travelled longer distances to reach the escaping hole than did their wild-type littermates (non-APP) after a similar bone marrow reconstruction. MyD88-deficient (ko) bone marrow reconstruction significantly reduced the cognitive deficits in both TgCRND8 and APPswe/PS1dE9 mice compared with the wild-type bone marrow transplantation [(A and B) two-way ANOVA, P 5 0.05, n = 11 per group for TgCRND8 mice and n = 8 per group for APPswe/PS1dE9 mice]. The amount of synaptophysin in the brain homogenate derived from TgCRND8 mice was measured with ELISA (C). MyD88-ko bone marrow reconstruction was associated with a higher level of synaptophysin in the Ab precursor protein transgenic mouse than the wild-type bone marrow reconstruction (t-test, P 5 0.05, n = 11 per group, the level in wild-type bone marrow chimeric mice is set at 1). Furthermore, the transcriptional level of Arc or NMDA-R1 gene was significantly higher in MyD88-ko than wild-type bone marrow chimeric Ab precursor protein transgenic mice [(D) t-test, P 5 0.05, n = 11 per group].

hippocampus by counting Iba-1 positive cells in TgCRND8 mice. The two groups differed significantly in microglial recruitment, with 122.37 4.01 cells/mm2 in wild-type bone marrowreconstructed Ab precursor protein transgenic mice compared to 111.17 3.38 cells/mm2 in MyD88-deficient bone marrowreconstructed Ab precursor protein mice (P = 0.043, Fig. 3A). Microglial morphology was similar in the two groups (Fig. 3A insert). We used flow cytometry to count CD45 + /CD11b + cells, which represent peripherally recruited and resident activated


We examined pathogenic effects of MyD88-deficient bone marrow-derived microglia on neuronal function with the Barnes maze test. TgCRND8 Ab precursor protein transgenic mice (three months post-transplantation with wild-type bone marrow) spent significantly more time (Fig. 2A, P 5 0.001), travelled longer distances (Fig. 2A, P 5 0.001) and made more errors (data not shown, P 5 0.001) before reaching the escape hole than did their wild-type littermates after a similar bone marrow reconstruction. The efficacy of Barnes maze as a cognitive test for APPswe/ PS1dE9 transgenic mice was validated by O’Leary and Brown (2009). After successful bone marrow reconstruction with MyD88-deficient bone marrow cells, both TgCRND8 mice at three months post-transplant and APPswe/PS1dE9 transgenic mice at 12 months post-transplant reached the escape hole in significantly faster times and shorter distances compared with the respective wild-type bone marrow-transplanted Ab precursor protein transgenic mice (Fig. 2A and B, P 5 0.05). Similarly, we observed that MyD88-deficiency in bone marrow cells markedly reduced errors for TgCRND8 mice engaged in searching for the escape hole (data not shown, P 5 0.001). We further measured synaptophysin in the brain homogenate by ELISA to evaluate the neuronal integrity. We observed 0.016 0.001 (OD-value/mg of wet brain) in TgCRND8 Ab precursor protein transgenic mice (n = 16), which was significantly less than 0.019 0.001 in their wild-type littermates (n = 8, t-test versus Ab precursor protein mice, P = 0.004). Interestingly, measurements of synaptophysin levels showed that the protein level in MyD88-deficient bone marrow chimeric TgCRND8 Ab precursor protein mice was 1.39 0.12-fold higher than in the wild-type bone marrow Ab precursor protein control mice (P = 0.003, Fig. 2C). Moreover, transcription of the Arc and NMDA-R1 genes, both of which are involved in neuronal plasticity (Dickey et al., 2003; Bramham et al., 2008; Richard et al., 2008), was significantly upregulated in TgCRND8 Ab precursor protein transgenic mice with MyD88-deficient bone marrow reconstruction compared with Ab precursor protein transgenic mice with wild-type bone marrow reconstruction (P 5 0.05, Fig. 2D). In the APPswe/PS1dE9 mice, we did not observe any effects of MyD88-deficient bone marrow reconstruction on synaptophysin levels or transcription of Arc and NMDA-R1 genes (data not shown).

284

| Brain 2011: 134; 278–292

W. Hao et al.

reconstruction with MyD88-deficient (ko) and wild-type (wt) cells, the brains of TgCRND8 Ab precursor protein (APP) transgenic mice were analysed for inflammatory activation. (A) Microglial recruitment was investigated by performing immunohistochemistry with antibodies against Iba-1. Positive staining appears brown. The number of Iba-1 positive cells was significantly more in the hippocampus of Ab precursor protein transgenic mice reconstructed with wild-type-bone marrow cells than with MyD88-deficient-bone marrow cells (t-test, P 5 0.05, n 5 13 per group). Microglial morphology shown in the inserts did not differ between these two mouse groups. (B) CD45-positive microglia were calculated using flow cytometry. Following wild-type bone marrow reconstruction, the percentage of CD45 + cells among the CD11b + cell population was higher in Ab precursor protein transgenic mice than in their wild-type littermates. Compared with wild-type bone marrow reconstruction, MyD88-ko bone marrow reconstruction significantly reduced CD45 + /CD11b + cells in the Alzheimer’s disease mouse brain (one-way ANOVA followed by post hoc tests, P 5 0.05, n = 4 per group). (C) Inflammatory gene transcripts were measured by quantitative reverse transcription polymerase chain reaction. Transcription of TNF- and CCL-2 in TgCRND8 mice, and TNF- and interleukin-1b in APPswe/PS1dE9 mice was significantly lower in MyD88-deficient-bone marrow-reconstructed mice than in the wild-type-bone marrow controls (t-test, P 5 0.05, n 5 13 per group). FITC = fluorescein isothiocyanate; IL = interleukin; iNOS = inducible nitric oxide synthase.

microglia (Gate et al., 2010). We observed that more CD45 + microglia were recruited in Ab precursor protein transgenic mice than in their wild-type littermates after bone marrow reconstruction (Fig. 3B, P = 0.002). In bone marrow-reconstructed Ab precursor protein mice, MyD88-deficiency in bone marrow cells significantly reduced the number of CD45 + microglia compared to the wild-type control (Fig. 3B, P = 0.027). In additional control experiments, we counted Iba-1 positive cells in the hippocampus of C57BL6 mice three months post wild-type bone marrow

reconstruction and their non-irradiated controls and observed the densities of 62.84 4.64 and 66.87 2.73 cells/mm2, respectively (t-test, P = 0.48, n = 4 per group), which indicates that irradiation has limited effects on microglial recruitment in non-Ab precursor protein transgenic wild-type mice. We further quantified brain transcription of proinflammatory cytokines and chemokines, such as TNF-, interleukin-1b and CCL-2, in both 6-month-old TgCRND8 and 13-month-old APPswe/PS1dE9 transgenic mice three months after bone marrow


Figure 3 MyD88-deficient bone marrow-derived cells reduce inflammatory activities in the brain. Three months after bone marrow

MyD88-deficiency improves AD pathology transplantation. As shown in Fig. 3C, levels of TNF- and CCL-2 transcripts in TgCRND8 Ab precursor protein mice and TNF- and interleukin-1b in APPswe/PS1dE9 mice were significantly reduced by MyD88-deficient bone-marrow reconstruction as compared with wild-type bone marrow reconstruction (P 5 0.05, Fig. 3C). Transcription of anti-inflammatory cytokine interleukin-10 did not differ between the two sets of mice, and interleukin-4 transcription was not detectable in any of the animals.

MyD88-deficient bone marrow reconstruction reduces cerebral Ab in Ab precursor protein transgenic mice To analyse the effects of bone marrow reconstruction on Ab load in the brain, we first determined the Ab profile in MyD88-deficient

Brain 2011: 134; 278–292

| 285

bone marrow chimeric Ab precursor protein transgenic animals by western blots (Supplementary Fig. 1). We separated brain homogenate from MyD88-deficient and control TgCRND8 mice into three fractions (sodium carbonate buffer, sodium carbonate buffer plus Triton and guanidine chloride-soluble). The sodium carbonate buffer fraction consisted primarily of monomers, the sodium carbonate buffer plus Triton fraction consisted of oligomers, and the guanidine fraction was enriched with high molecular-weight aggregates. We measured the amounts of cerebral Ab in TgCRND8 Ab precursor protein transgenic with and without wild-type bone marrow reconstruction. We found the concentration of Ab42 in the guanidine fraction to be markedly reduced following bone marrow transplantation (P 5 0.001, Fig. 4A), consistent with the results recently reported by Keene’s group (2010). In the comparison with wild-type bone marrow


Figure 4 MyD88-deficient bone marrow-derived cells reduce Ab load in the brain. Three months after bone marrow transplantation (BMT) with MyD88-deficient (ko) and wild-type (wt) cells, the brains of TgCRND8 Ab precursor protein (APP) transgenic mice were analysed for Ab load. Six Ab precursor protein transgenic mice without bone marrow transplantation were used as controls. (A) The brain was homogenized and separated into sodium carbonate buffer (CB), sodium carbonate buffer plus Triton (CB-T) and guanidine-soluble fractions. Amounts of Ab40 and Ab42 were measured by ELISA and normalized to the protein concentration. The Ab42 concentration in the guanidine fraction was markedly reduced by bone marrow transplantation. The concentrations of Ab40 or Ab42 in the sodium carbonate buffer plus triton fraction and Ab42 in guanidine fraction were significantly lower in the MyD88-deficient-bone marrow chimeric Ab precursor protein mouse than in the wild-type-bone marrow-chimeric Alzheimer’s disease mouse (one-way ANOVA followed by post hoc tests, P 5 0.05, n 5 14 per group). (B) The Ab load was evaluated by quantifying the positive immunohistochemical staining of human Ab in the whole hippocampus and the cortex dorsal to hippocampus. Ab deposits in MyD88-deficient-bone marrow reconstructed mice were significantly reduced as compared with the wild-type-bone marrow control (t-test, P 5 0.05, n 5 9 per group). ROI = region of interest.

286

| Brain 2011: 134; 278–292

W. Hao et al.

Modulation of Ab metabolism is not the mechanism by which MyD88-deficient bone marrow cells reduce Ab levels After observing that MyD88-deficient bone marrow reconstruction reduced both inflammatory activation and Ab load, we investigated the mechanism responsible for Ab reduction. To determine whether Ab production was altered in our bone marrow chimeric Alzheimer’s disease model, we measured Ab precursor protein level and b-secretase 1 activity. As shown in Fig. 5A, Ab precursor protein amount in sodium carbonate buffer plus Triton-soluble brain homogenates were not significantly different between the two groups of mice reconstructed with MyD88-deficient and wild-type bone marrows (MyD88-deficient bone marrow: 0.31 0.01 and wild-type bone marrow: 0.34 0.03 mg/mg brain, P = 0.467). Interestingly, b-secretase 1 activity (cleaving Ab precursor protein to produce Ab) was significantly increased in MyD88-deficient bone marrow chimeric mice compared with wild-type control mice (P = 0.013, Fig. 5B). To confirm this result, we performed b-secretase 1 activity assays in brains from 6-month-old MyD88 or toll-like receptor 2-deficient and wild-type mice without Ab precursor protein transgenic expression and observed that b-secretase 1 activity was significantly elevated in MyD88 or toll-like receptor 2-deficient mice compared with C57BL6 controls (Supplementary Fig. 2). In order to clarify the biological significance of increased b-secretase 1 activity, we investigated the cellular origin of b-secretase 1. As shown in Fig. 5C, we observed that b-secretase 1 was mainly expressed in neurons, but was also seen in astrocytes and microglia, especially around Ab

Figure 5 MyD88-deficient bone marrow cells neither reduce Ab production nor increase Ab degradation. Ab precursor protein (APP) level and b-secretase 1 (BACE1) activity in the brain homogenate of TgCRND8 Ab precursor protein transgenic mice were analysed three months after bone marrow reconstruction. MyD88-deficient (ko) bone marrow reconstruction did not alter Ab precursor protein amount [(A) t-test, P 4 0.05, n 5 9 per group] but significantly increased b-secretase 1 activity [(B) t-test, P 5 0.05, n 5 9 per group]. The cellular origin of b-secretase 1 is shown in C by the co-localization (in yellow) of b-secretase 1 (in green) and different cell markers (NeuN, GFAP and MHC II, in red) under confocal microscopy. For comparison with these co-stained cells (marked with filled arrows), cells that express MHC II but not b-secretase 1 are marked with an empty arrow. The transcription of neprilysin (NEP) or insulin-degrading enzyme (IDE) was not significantly changed [(D) t-test, P 4 0.05, n 5 9 per group]. wt = wild-type.

deposits. We detected b-secretase 1 transcription at low levels in cultured macrophages. Significantly higher levels of b-secretase 1 transcript were seen in MyD88-deficient macrophages compared with the wild-type control; however, this transcriptional difference disappeared after treatment with 10 mM aggregated Ab42 (data not shown). Levels of b-secretase 1 transcription were similar in


chimeric TgCRND8 mice, we observed that the concentrations of both Ab40 and Ab42 in the sodium carbonate buffer plus Triton fraction and Ab42 in guanidine fraction were significantly lower following MyD88-deficient bone marrow reconstruction (P 5 0.05, Fig. 4A). In the guanidine fraction, the concentration of Ab40 was also reduced in MyD88-deficient bone marrow chimeric mice, compared with control animals, although the difference did not reach statistical significance (P 4 0.05, Fig. 4A). In the sodium carbonate buffer fraction, the concentrations of Ab40 and Ab42 were nearly the same between the two bone marrow chimeric mouse groups (P 4 0.05, Fig. 4A). Immunohistochemical staining showed that the levels of Ab deposits in the TgCRND8 Ab precursor protein transgenic mouse brain receiving MyD88-deficient bone marrow cells (0.35 0.04% of region of interest) were significantly decreased as compared with levels in the control brain (0.68 0.10% of region of interest; P = 0.047, Fig. 4B). We further confirmed the Ab-reducing effect of MyD88deficient bone marrow cells by measuring Ab concentrations in the brain of 13-month-old bone marrow chimeric APPswe/PS1dE9 mice. The concentrations of Ab42 in both sodium carbonate buffer plus Triton and guanidine fractions were significantly reduced by MyD88-deficient bone marrow cells (sodium carbonate buffer plus triton: 0.368 0.016 ! 0.243 0.059 and guanidine: 84.911 9.362 ! 66.653 8.396 ng/mg wet brain; t-test, P = 0.043 and 0.037, respectively; n = 7 per group).

MyD88-deficiency improves AD pathology microglia isolated from brains of 6-month-old MyD88-deficient and wild-type mice (isolation according to Cardona et al., 2006; data not shown). Thus, elevated b-secretase 1 activity could result from cross-talk between bone marrow-derived and resident microglia or from the talk between microglia and neurons. MyD88deficient bone marrow reconstruction did not inhibit Ab precursor protein processing. We further used quantitative reverse transcription polymerase chain reaction to measure expression of Ab-degrading enzymes, neprilysin and insulin-degrading enzyme (Leissring et al., 2003; Miners et al., 2008). There was no increase in transcription of NEP and IDE genes in Ab precursor protein transgenic mice reconstructed with MyD88-deficient bone marrow (Fig. 5D), suggesting that MyD88 deficiency has no effect on Ab catabolism.

Deane and colleagues (2004) reported that altered communication of Ab between the brain interstitial fluid and peripheral blood stream affects cerebral Ab load. Thus, we measured plasma levels of Ab and expression of RAGE and LRP1 in the brain, two receptors that act as shuttles to transport Ab across blood–brain barrier (Shibata et al., 2000; Deane et al., 2003, 2004). As shown in Fig. 6, neither plasma Ab40/42 concentrations (Fig. 6A), nor receptor expression in the brain (Fig. 6B) was affected by MyD88-deficient bone marrow reconstruction (P 4 0.05).

MyD88-deficient bone marrow reconstruction enhances microglial recruitment towards Ab deposits and cerebral expression of Ab-interacting receptors We hypothesized that microglial clearance of Ab could be the mechanism to reduce cerebral Ab. To test this hypothesis, we used confocal microscopy to investigate the relationship between Ab deposits and microglia. We could only count microglial cells co-localizing with Ab deposits, as technical difficulties prevented reliable identification of intracellular Ab in the whole brain. However, we clearly observed the co-localization of Ab and microglia (Fig. 7A). Recruitment of microglial cells to Ab deposits appeared to be enhanced in MyD88-deficient bone marrow chimeric mice compared with wild-type control mice (P = 0.069, Fig. 7B). Furthermore, we observed that MyD88-deficient bone marrow reconstruction significantly increased the transcription in the brain of the receptors, scavenger receptor A and CD36 (P 5 0.05, Fig. 7C), which are known to be involved in Ab phagocytosis (El Khoury et al., 1996, 2003).

| 287

Figure 6 MyD88-deficient bone marrow cells do not alter transportation of Ab across the blood–brain barrier in TgCRND8 Ab precursor protein transgenic mice. Ab concentration in the plasma of Ab precursor protein transgenic mice was measured 3 months after bone marrow reconstruction with MyD88-deficient (ko) and wild-type (wt) cells. Transcripts of RAGE and LRP1 in the brain were detected with quantitative reverse transcription polymerase chain reaction. Neither plasma Ab concentrations (A) nor gene transcription (B) were significantly changed (t-test, P 4 0.05, n 5 9 per group).

MyD88-deficiency significantly enhances phagocytosis of aggregated Ab42 by macrophages after a long-term incubation As it is difficult to show the effects of MyD88 on microglial uptake of Ab in vivo, we performed Ab internalization assays in cultured MyD88-deficient and wild-type bone marrow-derived macrophages. Macrophage internalization of Ab was detectable within 1 h after administration of aggregated fluorescein isothiocyanatelabelled Ab42, and levels of internalized Ab increased as the incubation time increased (Fig. 8A and B). Treatment with cytochalasin D, a known inhibitor of phagocytosis (Liu et al., 2005), blocked the internalization of aggregated Ab within 2 h, indicating the phagocytotic pattern of Ab42 internalization (data not shown). In comparing the phagocytotic capacities of MyD88-deficient and wild-type macrophages, we observed that MyD88-deficiency did not reduce Ab phagocytosis following treatment with Ab42 aggregates at concentrations of 0.5 and 5 mM (data not shown). Notably, following treatment with aggregated Ab42 at 10 mM, we observed that Ab42 uptake over a 6 h interval was significantly increased in MyD88-deficient macrophages compared with wild-type macrophages (P = 0.033, Fig. 8A and B). This difference became even greater when Ab42 treatment was prolonged (24 h, P = 0.044, Fig. 8B). No differences in autofluorescence were observed between MyD88-deficient and wild-type macrophages. To investigate the relationship between Ab phagocytosis and inflammatory activation, we monitored TNF- secretion from Ab42-phagocytosing macrophages. TNF- release from MyD88wild-type macrophages treated with 10 mM Ab42 started to


MyD88-deficient bone marrow reconstruction does not affect Ab transportation across the blood–brain barrier

Brain 2011: 134; 278–292

288

| Brain 2011: 134; 278–292

W. Hao et al.

towards Ab deposits and cerebral expression of Ab-interacting receptors. Three months after bone marrow reconstruction with MyD88-deficient (ko) and wild-type (wt) cells, the brain of TgCRND8 Ab precursor protein transgenic mouse was analysed for the interaction between Ab and microglia under confocal microscopy after the immunofluorescent staining with antibodies against Ab and Iba-1 [(A) Ab in red and Iba-1 in green, the co-localization was shown by arrows]. The ratio of microglia co-localizing with Ab was counted and tended to be higher in MyD88-deficient bone marrow chimeric mice as compared with the wild-type bone marrow control mice [(B) t-test, P = 0.069, n = 8 per group]. The transcription of scavenger receptor A (SR-A) and CD36 was measured with quantitative reverse transcription polymerase chain reaction and was upregulated after MyD88-deficient bone marrow reconstruction [(C) t-test, P 5 0.05, n 5 9 per group, the level in wild-type bone marrow chimeric mice is set at 1].

increase significantly after 6 h, whereas no significantly elevated secretion of TNF- was detected from MyD88-deficient macrophages (Fig. 8C). Ab42 at 5 mM triggered TNF- secretion at a lower level (data not shown). Internalization of Ab42 aggregates at 0.5 mM did not cause TNF- secretion within the time frame studied (1–24 h) in either cell type, in corroboration with earlier observations (Liu et al., 2005). To analyse cell damage from high concentrations of Ab, we used 10 mM aggregated Ab42 for cell treatment in experiments to measure caspase-3/7 activity and leakage of lactate dehydrogenase into the extracellular medium as indicators of apoptosis and necrosis of macrophages. We did not observe significant cytotoxicity of Ab42 on macrophages within 24 h (Supplementary Fig.3). To confirm the effects of MyD88-deficiency on macrophage phagocytosis of Ab, we examined the effect of toll-like receptor 2 deficiency on Ab phagocytosis by macrophages. Toll-like receptor 2 is a receptor upstream of MyD88 and we found that toll-like receptor 2-deficiency similarly increased Ab internalization (Supplementary Fig. 4).

Figure 8 MyD88-deficiency increases Ab42 phagocytosis by macrophages. Bone marrow-derived macrophages from MyD88-deficient (ko) or wild-type (wt) mice were incubated with 10 mM aggregated fluorescein isothiocyanate (FITC)-conjugated Ab42 for 0, 1, 3, 6 and 24 h. The cells were detached for flow cytometric analysis of fluorescence (A and B) and the culture medium was collected for ELISA measurement of TNF- (C). (A) A dot-plot shows gating of analysed cells and histograms demonstrate the different Ab-fluorescence in MyD88-deficient and wild-type macrophages at 6 h time point. a: t-test versus wild-type control in phagocytosis assay, P = 0.033 and 0.044 at 6- and 24-h time points, respectively, n 5 11 per group; *: one-way ANOVA followed by post hoc test versus TNF- secretion at time point of 0 h, P = 0.043 and 50.001, at 6 and 24 h time points, respectively, n = 8 per group. mFI = mean fluorescence intensity.

Discussion In the pathophysiology of Alzheimer’s disease, microglia are key players, although they behave like a double-edged sword; on one hand, they promote neuronal degeneration by releasing neurotoxic inflammatory mediators; on the other hand, they can clear Ab plaques by phagocytosis (Akiyama et al., 2000; Bard et al., 2000; Selkoe, 2002; Liu et al., 2005; Wyss-Coray, 2006). Here, we show that depleting MyD88 in bone marrow cells of Ab precursor protein transgenic mice simultaneously shifts both phagocytotic clearance of Ab and neuroinflammation in the beneficial direction. Furthermore, MyD88 deficiency improves spatial learning and contextual memory deficits in Ab precursor protein transgenic mice. The establishment of our MyD88-deficient bone marrowchimeric Alzheimer’s disease mice was based on the concept that behaviour of brain mononuclear phagocytes can be modified


Figure 7 MyD88 deficiency enhances microglial recruitment


| 289

but also is seen in astrocytes and microglia. We further demonstrated that MyD88-deficiency does not significantly alter b-secretase 1 expression in adult brain-derived microglia or cultured Ab-activated macrophages. Thus, bone marrow-derived microglia themselves are unlikely to contribute to the elevated activity of cerebral b-secretase 1; however, it remains a possibility that the interaction between bone marrow-derived and resident microglia and the cross-talk between microglia and neurons could modify Ab precursor protein processing in the brain. Furthermore, neprilysin and insulin-degrading enzyme, considered as Ab-degrading enzymes (Leissring et al., 2003; Miners et al., 2008), were not upregulated in the brain of the MyD88-deficient animal. Therefore, neither decreased Ab production nor increased Ab degradation can explain the reduced cerebral Ab load in our experiments. Another possibility is that altered transport of Ab between peripheral circulation and brain parenchyma could affect cerebral Ab load. However, compared with control animals, Ab precursor protein transgenic mice with MyD88-deficient bone marrow reconstruction showed changes neither in plasma Ab levels nor in cerebral expression of RAGE and LRP1, involved in transport of Ab across the blood–brain barrier (Shibata et al., 2000; Deane et al., 2003, 2004). Previous studies suggest that inflammatory activation inhibits phagocytosis in mononuclear phagocytes (Koenigsknecht-Talboo and Landreth, 2005; Townsend et al., 2005; Zelcer et al., 2007; Hickman et al., 2008). In both patients with Alzheimer’s disease and aged Ab precursor protein transgenic mice, decreased Ab clearance correlates with higher cytokine production (Fiala et al., 2005, 2007; Hickman et al., 2008). Therefore, the reduction of Ab load in our Ab precursor protein transgenic mice reconstructed with MyD88-deficient bone marrow could be caused by enhanced microglial phagocytosis of Ab due to lack of suppression by inflammatory activation. Activation of toll-like receptors 2, 4 and 9, the upstream receptors of MyD88, has been shown to increase Ab phagocytosis (Iribarren et al., 2005; Chen et al., 2006; Tahara et al., 2006; Scholtzova et al., 2009). However, in those experiments, the increased Ab phagocytosis occurred after prestimulation by toll-like receptor ligands. Upon Ab challenge, phagocytosis, indeed, occurs earlier than inflammatory activation (Liu et al., 2005), with phagocytosis even serving to trigger inflammatory activation (Halle et al., 2008). Moreover, toll-like receptortriggered phagocytosis was observed not to be strictly dependent on MyD88 (Kong and Ge, 2008). It is still controversial which cells (resident microglia or bone marrow-derived microglia) are more important for Ab removal. Resident microglia have limited effects on formation and maintenance of Ab deposits (Grathwohl et al., 2009). Bone marrowderived microglia have been reported to be more efficient than their resident counterparts in eliminating Ab deposits (Simard et al., 2006). However, in our bone marrow chimeric Alzheimer’s disease model, we observed that the number of recruited bone marrow cells was much lower than that of their resident counterparts. It is even possible that recruited bone marrow cells cross-talk with resident microglia, especially around the Ab deposits, thereby enhancing their phagocytotic capacity.


by manipulating bone marrow cells. Bone marrow-derived cells have been observed to migrate to the brain and differentiate into microglia (Priller et al., 2001; Simard et al., 2006). These peripherally recruited cells serve immune functions similar to those of their resident counterparts (Turrin et al., 2007). This concept potentially has considerable clinical relevance and has already been translated into clinical practice in patients with amyotrophic lateral sclerosis (Appel et al., 2008). As a therapeutic approach, a genetically engineered MyD88 gene inserted into bone marrow cells is preferable to a pharmacological inhibitor of MyD88, as the latter non-specifically targets MyD88-expressing cells and could have toxic effects on neuronal and endothelial cells. Here we used 3-month-old TgCRND8 and 6- to 9-month-old APPswe/ PS1dE9 mice with already marked plaque pathology (Chishti et al., 2001; Kim et al., 2009) as animal models of symptomatic Alzheimer’s disease, in order to better model the clinical intervention. The results of this study raise the possibility that depleting MyD88 in bone marrow cells could become a therapeutic strategy for patients with Alzheimer’s disease. The inhibition of neuroinflammation by recruited MyD88deficient bone marrow cells corroborates findings from earlier studies on the role of innate immune receptors, e.g. CD14, toll-like receptor 2 and toll-like receptor 4, in Ab-triggered microglial activation (Fassbender et al., 2004; Walter et al., 2007; Jana et al., 2008; Reed-Geaghan et al., 2009). Notably, these innate immune receptors all depend on MyD88 in transduction of inflammatory signals (Akira et al., 2001). Our study showed that bone marrow reconstruction with wild-type cells following whole body irradiation reduces cerebral Ab in Ab precursor protein transgenic mice, which was also recently reported by Keene’s group (2010). This Ab-reducing effect is unlikely to be due to the irradiation, as head irradiation alone without bone marrow manipulation does not change Ab amount in the brain (Keene et al., 2010). Interestingly, we demonstrated that MyD88-deficiency augments this bone marrow cells-mediated Ab reduction and probably enhances the microglial phagocytosis of Ab. In Ab precursor protein transgenic mice reconstructed with MyD88-deficient bone marrow, the total number of microglia was decreased, but recruitment of microglial cells at sites of Ab deposits was not impaired, and Ab phagocytosis-related receptors, such as scavenger receptor A and CD36 (El Khoury et al., 1996, 2003) were even upregulated in the brain. In principle, this upregulated scavenger receptor A and CD36 could even facilitate microglia to adhere to Ab, although this hypothesis needs to be tested further. Direct evidence of phagocytotic clearance as the mechanism for reducing Ab load comes from the observation that Ab phagocytosis was markedly enhanced in cultured MyD88-deficient macrophages, while the inflammatory responsiveness to Ab was nearly completely shut down. Alternatively, inflammation could have modulated expression of Ab precursor protein and b-secretase 1 as previously reported (Grilli et al., 1995; He et al., 2007; Buggia-Prevot et al., 2008). However, we observed that MyD88-deficient microglia did not decrease but instead increased b-secretase 1 activity, consistent with our observations that systemic knockout of MyD88 markedly increases both b-secretase 1 and -secretase activity in the brain. We observed that b-secretase 1 is mainly expressed in neurons,

Brain 2011: 134; 278–292

290

| Brain 2011: 134; 278–292

In conclusion, this study demonstrates that replenishment of microglia with MyD88-deficient bone marrow-derived precursor cells improves cognitive function of the animal model of Alzheimer’s disease, potentially by suppressing detrimental neuroinflammation while, in parallel, enhancing beneficial Ab clearance. Engineering the MyD88 gene in host bone marrow cells could feasibly offer a therapeutic option to prevent progression of Alzheimer’s disease.

Acknowledgements

Funding Alzheimer Forschung Initiative e.V. (to Y.L.); Fritz Thyssen Stiftung (to Y.L. and K.F.); Hermann und Dr Charlotte Deutsch Stiftung (to K.F. and Y.L.); University of the Saarland (to Y.L.).

Supplementary material Supplementary material is available at Brain online.

References Akira S, Takeda K, Kaisho T. Toll-like receptors: critical proteins linking innate and acquired immunity. Nat Immunol 2001; 2: 675–80. Akiyama H, Barger S, Barnum S, Bradt B, Bauer J, Cole GM, et al. Inflammation and Alzheimer’s disease. Neurobiol Aging 2000; 21: 383–421. Appel SH, Engelhardt JI, Henkel JS, Siklos L, Beers DR, Yen AA, et al. Hematopoietic stem cell transplantation in patients with sporadic amyotrophic lateral sclerosis. Neurology 2008; 71: 1326–34. Babcock AA, Toft-Hansen H, Owens T. Signaling through MyD88 regulates leukocyte recruitment after brain injury. J Immunol 2008; 181: 6481–90. Bard F, Cannon C, Barbour R, Burke RL, Games D, Grajeda H, et al. Peripherally administered antibodies against amyloid beta-peptide enter the central nervous system and reduce pathology in a mouse model of Alzheimer disease. Nat Med 2000; 6: 916–19. Barnes CA. Memory deficits associated with senescence: a neurophysiological and behavioral study in the rat. J Comp Physiol Psychol 1979; 93: 74–104. Benzing WC, Wujek JR, Ward EK, Shaffer D, Ashe KH, Younkin SG, et al. Evidence for glial-mediated inflammation in aged APP(SW) transgenic mice. Neurobiol Aging 1999; 20: 581–9. Bsibsi M, Ravid R, Gveric D, van Noort JM. Broad expression of Toll-like receptors in the human central nervous system. J Neuropathol Exp Neurol 2002; 61: 1013–21.

Bolmont T, Haiss F, Eicke D, Radde R, Mathis CA, Klunk WE, et al. Dynamics of the microglial/amyloid interaction indicate a role in plaque maintenance. J Neurosci 2008; 28: 4283–92. Bramham CR, Worley PF, Moore MJ, Guzowski JF. The immediate early gene arc/arg3.1: regulation, mechanisms, and function. J Neurosci 2008; 28: 11760–7. Buggia-Prevot V, Sevalle J, Rossner S, Checler F. NFkappaB-dependent control of BACE1 promoter transactivation by Abeta42. J Biol Chem 2008; 283: 10037–47. Cardona AE, Huang D, Sasse ME, Ransohoff RM. Isolation of murine microglial cells for RNA analysis or flow cytometry. Nat Protoc 2006; 1: 1947–51. Cagnin A, Brooks DJ, Kennedy AM, Gunn RN, Myers R, Turkheimer FE, et al. In-vivo measurement of activated microglia in dementia. Lancet 2001; 358: 461–7. Chen K, Iribarren P, Hu J, Chen J, Gong W, Cho EH, et al. Activation of Toll-like receptor 2 on microglia promotes cell uptake of Alzheimer disease-associated amyloid beta peptide. J Biol Chem 2006; 281: 3651–9. Chishti MA, Yang DS, Janus C, Phinney AL, Horne P, Pearson J, et al. Early-onset amyloid deposition and cognitive deficits in transgenic mice expressing a double mutant form of amyloid precursor protein 695. J Biol Chem 2001; 276: 21562–70. Dahlgren KN, Manelli AM, Stine WB Jr, Baker LK, Krafft GA, LaDu MJ. Oligomeric and fibrillar species of amyloid-beta peptides differentially affect neuronal viability. J Biol Chem 2002; 277: 32046–53. Deane R, Du Yan S, Submamaryan RK, LaRue B, Jovanovic S, Hogg E, et al. RAGE mediates amyloid-beta peptide transport across the blood– brain barrier and accumulation in brain. Nat Med 2003; 9: 907–13. Deane R, Wu Z, Zlokovic BV. RAGE (yin) versus LRP (yang) balance regulates alzheimer amyloid beta-peptide clearance through transport across the blood-brain barrier. Stroke 2004; 35: 2628–31. Dickey CA, Loring JF, Montgomery J, Gordon MN, Eastman PS, Morgan D. Selectively reduced expression of synaptic plasticity-related genes in amyloid precursor protein + presenilin-1 transgenic mice. J Neurosci 2003; 23: 5219–26. Edison P, Archer HA, Gerhard A, Hinz R, Pavese N, Turkheimer FE, et al. Microglia, amyloid, and cognition in Alzheimer’s disease: an [11C](R)PK11195-PET and [11C]PIB-PET study. Neurobiol Dis 2008; 32: 412–19. Eikelenboom P, Stam FC. Immunoglobulins and complement factors in senile plaques. An immunoperoxidase study. Acta Neuropathol 1982; 57: 239–42. El Khoury J, Hickman SE, Thomas CA, Cao L, Silverstein SC, Loike JD. Scavenger receptor-mediated adhesion of microglia to beta-amyloid fibrils. Nature 1996; 382: 716–19. El Khoury JB, Moore KJ, Means TK, Leung J, Terada K, Toft M, et al. CD36 mediates the innate host response to beta-amyloid. J Exp Med 2003; 197: 1657–66. Eriksen JL, Sagi SA, Smith TE, Weggen S, Das P, McLendon DC, et al. NSAIDs and enantiomers of flurbiprofen target gamma-secretase and lower Abeta 42 in vivo. J Clin Invest 2003; 112: 440–9. Fassbender K, Walter S, Ku¨hl S, Landmann R, Ishii K, Bertsch T, et al. The LPS receptor (CD14) links innate immunity with Alzheimer’s disease. FASEB J 2004; 18: 203–5. Fiala M, Lin J, Ringman J, Kermani-Arab V, Tsao G, Patel A, et al. Ineffective phagocytosis of amyloid-beta by macrophages of Alzheimer’s disease patients. J Alzheimers Dis 2005; 7: 221–32. Fiala M, Liu PT, Espinosa-Jeffrey A, Rosenthal MJ, Bernard G, Ringman JM, et al. Innate immunity and transcription of MGAT-III and Toll-like receptors in Alzheimer’s disease patients are improved by bisdemethoxycurcumin. Proc Natl Acad Sci USA 2007; 104: 12849–54. Gate D, Rezai-Zadeh K, Jodry D, Rentsendorj A, Town T. Macrophages in Alzheimer’s disease: the blood-borne identity. J Neural Transm 2010; 117: 961–70. Grathwohl SA, Ka¨lin RE, Bolmont T, Prokop S, Winkelmann G, Kaeser SA, et al. Formation and maintenance of Alzheimer’s disease


We are grateful to Prof. Harald Neumann (University of Bonn, Germany) for providing GFP transgenic mice and Prof. Heikki Tanila (University of Kuopio, Finland) for providing APPswe/ PS1dE9 mice for preliminary experiments. We also thank M. Ku¨hne, E. Gluding, A. Schottek and N. Commercon for their excellent technical assistance. We also appreciate the help from Prof. G. Wennemuth, Prof. F. Schmitz, Prof. P. Lipp, Dr K. Lars and S. Natarajan.

W. Hao et al.


| 291

Maier M, Peng Y, Jiang L, Seabrook TJ, Carroll MC, Lemere CA. Complement C3 deficiency leads to accelerated amyloid beta plaque deposition and neurodegeneration and modulation of the microglia/ macrophage phenotype in amyloid precursor protein transgenic mice. J Neurosci 2008; 28: 6333–41. Mayford M, Bach ME, Huang YY, Wang L, Hawkins RD, Kandel ER. Control of memory formation through regulated expression of a CaMKII transgene. Science 1996; 274: 1678–83. McGeer PL, Itagaki S, Boyes BE, McGeer EG. Reactive microglia are positive for HLA-DR in the substantia nigra of Parkinson’s and Alzheimer’s disease brains. Neurology 1988; 38: 1285–91. McGeer PL, Schulzer M, McGeer EG. Arthritis and anti-inflammatory agents as possible protective factors for Alzheimer’s disease: a review of 17 epidemiologic studies. Neurology 1996; 47: 425–32. McGowan E, Sanders S, Iwatsubo T, Takeuchi A, Saido T, Zehr C, et al. Amyloid phenotype characterization of transgenic mice overexpressing both mutant amyloid precursor protein and mutant presenilin 1 transgenes. Neurobiol Dis 1999; 6: 231–44. McLaurin J, Cecal R, Kierstead ME, Tian X, Phinney AL, Manea M, et al. Therapeutically effective antibodies against amyloid-beta peptide target amyloid-beta residues 4–10 and inhibit cytotoxicity and fibrillogenesis. Nat Med 2002; 8: 1263–9. Meyer-Luehmann M, Spires-Jones TL, Prada C, Garcia-Alloza M, de Calignon A, Rozkalne A. Rapid appearance and local toxicity of amyloid-beta plaques in a mouse model of Alzheimer’s disease. Nature 2008; 451: 720–4. Mildner A, Schmidt H, Nitsche M, Merkler D, Hanisch UK, Mack M, et al. Microglia in the adult brain arise from Ly-6ChiCCR2 + monocytes only under defined host conditions. Nat Neurosci 2007; 10: 1544–53. Miners JS, Baig S, Palmer J, Palmer LE, Kehoe PG, Love S. Abeta-degrading enzymes in Alzheimer’s disease. Brain Pathol 2008; 18: 240–52. Oakley H, Cole SL, Logan S, Maus E, Shao P, Craft J, et al. Intraneuronal beta-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: potential factors in amyloid plaque formation. J Neurosci 2006; 26: 10129–40. Okello A, Edison P, Archer HA, Turkheimer FE, Kennedy J, Bullock R, et al. Microglial activation and amyloid deposition in mild cognitive impairment: a PET study. Neurology 2009; 72: 56–62. O’Leary TP, Brown RE. Visuo-spatial learning and memory deficits on the Barnes maze in the 16-month-old APPswe/PS1dE9 mouse model of Alzheimer’s disease. Behav Brain Res 2009; 201: 120–7. Priller J, Flu¨gel A, Wehner T, Boentert M, Haas CA, Prinz M, et al. Targeting gene-modified hematopoietic cells to the central nervous system: use of green fluorescent protein uncovers microglial engraftment. Nat Med 2001; 7: 1356–61. Reed-Geaghan EG, Savage JC, Hise AG, Landreth GE. CD14 and toll-like receptors 2 and 4 are required for fibrillar A{beta}-stimulated microglial activation. J Neurosci 2009; 29: 11982–92. Richard KL, Filali M, Pre´fontaine P, Rivest S. Toll-like receptor 2 acts as a natural innate immune receptor to clear amyloid beta 1-42 and delay the cognitive decline in a mouse model of Alzheimer’s disease. J Neurosci 2008; 28: 5784–93. Scholtzova H, Kascsak RJ, Bates KA, Boutajangout A, Kerr DJ, Meeker HC, et al. Induction of toll-like receptor 9 signaling as a method for ameliorating Alzheimer’s disease-related pathology. J Neurosci 2009; 29: 1846–54. Selkoe DJ. Alzheimer’s disease is a synaptic failure. Science 2002; 298: 789–91. Shibata M, Yamada S, Kumar SR, Calero M, Bading J, Frangione B, et al. Clearance of Alzheimer’s amyloid-ss(1-40) peptide from brain by LDL receptor-related protein-1 at the blood-brain barrier. J Clin Invest 2000; 106: 1489–99. Simard AR, Soulet D, Gowing G, Julien JP, Rivest S. Bone marrow-derived microglia play a critical role in restricting senile plaque formation in Alzheimer’s disease. Neuron 2006; 49: 489–502.


beta-amyloid plaques in the absence of microglia. Nat Neurosci 2009; 12: 1361–3. Grilli M, Ribola M, Alberici A, Valerio A, Memo M, Spano P. Identification and characterization of a kappa B/Rel binding site in the regulatory region of the amyloid precursor protein gene. J Biol Chem 1995; 270: 26774–7. Hagihara H, Toyama K, Yamasaki N, Miyakawa T. Dissection of hippocampal dentate gyrus from adult mouse. J Vis Exp 2009; 33. pii: 1543. Halle A, Hornung V, Petzold GC, Stewart CR, Monks BG, Reinheckel T, et al. The NALP3 inflammasome is involved in the innate immune response to amyloid-beta. Nat Immunol 2008; 9: 857–65. He P, Zhong Z, Lindholm K, Berning L, Lee W, Lemere C, et al. Deletion of tumor necrosis factor death receptor inhibits amyloid beta generation and prevents learning and memory deficits in Alzheimer’s mice. J Cell Biol 2007; 178: 829–41. Hickey WF, Kimura H. Perivascular microglial cells of the CNS are bone marrow-derived and present antigen in vivo. Science 1988; 239: 290–2. Hickman SE, Allison EK, El Khoury J. Microglial dysfunction and defective beta-amyloid clearance pathways in aging Alzheimer’s disease mice. J Neurosci 2008; 28: 8354–60. Hooijmans CR, Van der Zee CE, Dederen PJ, Brouwer KM, Reijmer YD, van Groen T, et al. DHA and cholesterol containing diets influence Alzheimer-like pathology, cognition and cerebral vasculature in APPswe/PS1dE9 mice. Neurobiol Dis 2009; 33: 482–98. in t’ Veld BA, Ruitenberg A, Hofman A, Launer LJ, van Duijn CM, Stijnen T, et al. Nonsteroidal antiinflammatory drugs and the risk of Alzheimer’s disease. N Engl J Med 2001; 345: 1515–21. Iribarren P, Chen K, Hu J, Gong W, Cho EH, Lockett S, Uranchimeg B, Wang JM. CpG-containing oligodeoxynucleotide promotes microglial cell uptake of amyloid beta 1-42 peptide by up-regulating the expression of the G-protein-coupled receptor mFPR2. FASEB J 2005; 19: 2032–4. Jana M, Palencia CA, Pahan K. Fibrillar amyloid-beta peptides activate microglia via TLR2: implications for Alzheimer’s disease. J Immunol 2008; 181: 7254–62. Jankowsky JL, Slunt HH, Ratovitski T, Jenkins NA, Copeland NG, Borchelt DR. Co-expression of multiple transgenes in mouse CNS: a comparison of strategies. Biomol Eng 2001; 17: 157–65. Keene CD, Chang RC, Lopez-Yglesias AH, Shalloway BR, Sokal I, Li X, et al. Suppressed accumulation of cerebral amyloid {beta} peptides in aged transgenic Alzheimer’s disease mice by transplantation with wild-type or prostaglandin E2 receptor subtype 2-null bone marrow. Am J Pathol 2010; 177: 346–54. Kim J, Castellano JM, Jiang H, Basak JM, Parsadanian M, Pham V, et al. Overexpression of low-density lipoprotein receptor in the brain markedly inhibits amyloid deposition and increases extracellular Ab clearance. Neuron 2009; 64: 632–44. Koenigsknecht-Talboo J, Landreth GE. Microglial phagocytosis induced by fibrillar beta-amyloid and IgGs are differentially regulated by proinflammatory cytokines. J Neurosci 2005; 25: 8240–9. Koenigsknecht-Talboo J, Meyer-Luehmann M, Parsadanian M, GarciaAlloza M, Finn MB, Hyman BT, et al. Rapid microglial response around amyloid pathology after systemic anti-Abeta antibody administration in PDAPP mice. J Neurosci 2008; 28: 14156–64. Kong L, Ge BX. MyD88-independent activation of a novel actin-Cdc42/ Rac pathway is required for Toll-like receptor-stimulated phagocytosis. Cell Res 2008; 18: 745–55. Leissring MA, Farris W, Chang AY, Walsh DM, Wu X, Sun X, et al. Enhanced proteolysis of beta-amyloid in APP transgenic mice prevents plaque formation, secondary pathology, and premature death. Neuron 2003; 40: 1087–93. Liu Y, Hao W, Dawson A, Liu S, Fassbender K. Expression of amyotrophic lateral sclerosis-linked SOD1 mutant increases the neurotoxic potential of microglia via TLR2. J Biol Chem 2009; 284: 3691–9. Liu Y, Walter S, Stagi M, Cherny D, Letiembre M, Schulz-Schaeffer W, et al. LPS receptor (CD14): a receptor for phagocytosis of Alzheimer’s amyloid peptide. Brain 2005; 128: 1778–89.

Brain 2011: 134; 278–292

292

| Brain 2011: 134; 278–292

Turrin NP, Plante MM, Lessard M, Rivest S. Irradiation does not compromise or exacerbate the innate immune response in the brains of mice that were transplanted with bone marrow stem cells. Stem Cells 2007; 25: 3165–72. Walter S, Letiembre M, Liu Y, Heine H, Penke B, Hao W, et al. Role of the toll-like receptor 4 in neuroinflammation in Alzheimer’s disease. Cell Physiol Biochem 2007; 20: 947–56. Weggen S, Eriksen JL, Das P, Sagi SA, Wang R, Pietrzik CU, et al. A subset of NSAIDs lower amyloidogenic Abeta42 independently of cyclooxygenase activity. Nature 2001; 414: 212–16. Wyss-Coray T. Inflammation in Alzheimer disease: driving force, bystander or beneficial response? Nat Med 2006; 12: 1005–15. Yan R, Han P, Miao H, Greengard P, Xu H. The transmembrane domain of the Alzheimer’s beta-secretase (BACE1) determines its late Golgi localization and access to beta-amyloid precursor protein (APP) substrate. J Biol Chem 2001; 276: 36788–96. Zelcer N, Khanlou N, Clare R, Jiang Q, Reed-Geaghan EG, Landreth GE, et al. Attenuation of neuroinflammation and Alzheimer’s disease pathology by liver x receptors. Proc Natl Acad Sci USA 2007; 104: 10601–6.


Szekely CA, Green RC, Breitner JC, Østbye T, Beiser AS, Corrada MM, et al. No advantage of A beta 42-lowering NSAIDs for prevention of Alzheimer dementia in six pooled cohort studie. Neurology 2008; 70: 2291–8. Szekely CA, Thorne JE, Zandi PP, Ek M, Messias E, Breitner JC, et al. Nonsteroidal anti-inflammatory drugs for the prevention of Alzheimer’s disease: a systematic review. Neuroepidemiology 2004; 23: 159–69. Tahara K, Kim HD, Jin JJ, Maxwell JA, Li L, Fukuchi K. Role of toll-like receptor signalling in Abeta uptake and clearance. Brain 2006; 129: 3006–19. Takahashi K, Prinz M, Stagi M, Chechneva O, Neumann H. TREM2-transduced myeloid precursors mediate nervous tissue debris clearance and facilitate recovery in an animal model of multiple sclerosis. PLoS Med 2007; 4: e124. Town T, Laouar Y, Pittenger C, Mori T, Szekely CA, Tan J, et al. Blocking TGF-beta-Smad2/3 innate immune signaling mitigates Alzheimer-like pathology. Nat Med 2008; 14: 681–7. Townsend KP, Town T, Mori T, Lue LF, Shytle D, Sanberg PR, et al. CD40 signaling regulates innate and adaptive activation of microglia in response to amyloid beta-peptide. Eur J Immunol 2005; 35: 901–10.

W. Hao et al.