Journal of Alzheimer’s Disease 21 (2010) 507–518 507DOI 10.3233/JAD-2010-091471IOS Press

GM-CSF Upregulated in RheumatoidArthritis Reverses Cognitive Impairment andAmyloidosis in Alzheimer Mice

Tim D. Boyda,b,c,1, Steven P. Bennetta,b,c,1, Takashi Morid, Nicholas Governatorie, Melissa Runfeldte,Michelle Nordena, Jaya Padmanabhana,b, Peter Neamea,b, Inge Wefesb,c, Juan Sanchez-Ramosa,f,Gary W. Arendashe,g and Huntington Pottera,b,c,g,∗

aUSF Health Byrd Alzheimer’s Center and Research Institute,Tampa, FL, USAbDepartment of Molecular Medicine, College of Medicine, University of South Florida, Tampa, FL, USAcEric Pfeiffer Suncoast Gerontology Center, University of South Florida, Tampa, FL, USAdDepartments of Biomedical Sciences and Pathology, SaitamaMedical Center and Saitama Medical University,Kamoda, Kawagoe, Saitama, JapaneDepartment of Cell Biology, Microbiology, and Molecular Biology, University of South Florida, Tampa, FL, USAfDepartment of Neurology, College of Medicine, University of South Florida, Tampa, FL, USAgFlorida Alzheimer’s Disease Research Center, University of South Florida, Tampa, FL, USA

Handling Associate Editor: Luciano D’Adamio

Accepted 18 March 2010

Abstract. Rheumatoid arthritis (RA) is a negative risk factor for thedevelopment of Alzheimer’s disease (AD). While it hasbeen commonly assumed that RA patients’ usage of non-steroidal anti-inflammatory drugs (NSAIDs) helped prevent onsetand progression of AD, NSAID clinical trials have proven unsuccessful in AD patients. To determine whether intrinsic factorswithin RA pathogenesis itself may underlie RA’s protectiveeffect, we investigated the activity of colony-stimulating factors,upregulated in RA, on the pathology and behavior of transgenic AD mice. 5µg bolus injections of macrophage, granulocyte,and granulocyte-macrophage colony-stimulating factors (M-CSF, G-CSF, or GM-CSF) were administered unilaterally into thehippocampus of aged cognitively-impaired AD mice and the resulting amyloid load reductions determined one week later,using the artificial cerebrospinal fluid-injected contralateral sides as controls. G-CSF and more significantly, GM-CSF reducedamyloidosis throughout the treated brain hemisphere one week following bolus administration to AD mice. 20 daily subcutaneousinjections of 5µg of GM-CSF (the most amyloid-reducing CSF in the bolus experiment) were administered to balanced cohorts ofAD mice after assessment in a battery of cognitive tests. Reductions in amyloid load and improvements in cognitive function wereassessed. Subcutaneous GM-CSF administration significantly reduced brain amyloidosis and completely reversed the cognitiveimpairment, while increasing hippocampal synaptic area and microglial density. These findings, along with two decadesofaccrued safety data using Leukine, recombinant human GM-CSF, in elderly leukopenic patients, suggest that Leukine should betested as a treatment to reverse cerebral amyloid pathologyand cognitive impairment in AD.

Keywords: Alzheimer’s disease, amyloid-β, cognitive interference task, granulocyte-macrophage colony-stimulating factor, in-trahippocampal, radial arm water maze, rheumatoid arthritis, subcutaneous, transgenic mice

Supplementary data available online at http://www.j-alz.com/issues/21/vol21-2.html#supplementarydata03

1Both authors contributed equally to this work.∗Correspondence to: Huntington Potter, PhD, USF Health Byrd,

Alzheimer’s Center and Research Institute, 4001 E. Fletcher Avenue,Tampa, FL 33613, USA. Tel.: +1 813 396 0660; Fax: +1 813 9710373; E-mail: hpotter@health.usf.edu.

ISSN 1387-2877/10/$27.50 2010 – IOS Press and the authors. All rights reserved

508 T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice

INTRODUCTION

Although numerous studies have reported that rhe-umatoid arthritis (RA) reduces the risk of Alzheimer’sdisease (AD), the mechanisms for RA’s protectiveeffect are still unknown [1]. It was proposed, andcommonly assumed, that RA patients’ usage of non-steroidal anti-inflammatory drugs (NSAIDs) may helpprevent the onset and progression of AD [1]. However,the largest NSAID clinical trials have not demonstrat-ed efficacy in reducing the incidence of dementia, andrecently Naproxen was reported to be detrimental, withincreased risk of cardiovascular and cerebrovascularevents [2]. These results suggested to us that intrin-sic, probably immunological, factors within RA patho-genesis itself may underlie the protective effect of RAagainst AD. We surmised that upregulated local cellu-lar populations in RA would have the highest potentialto enter into the brain and inhibit the development ofAD pathology and/or neuronal dysfunction.

AD is an age-related, progressive neurodegenerativedisorder that presents as increasing decline in cogni-tive and executive function. Alzheimer dementia is as-sociated with cerebrovascular dysfunction [3], extra-cellular accumulation of amyloid-β (Aβ) peptides inthe brain parenchyma and vasculature walls [4,5] (pre-dominantly Aβ1−42 and Aβ1−40), and intraneuronalaccumulation of neurofibrillary tangles consisting ofhyperphosphorylated tau proteins [6]. Associated neu-roinflammationmay contribute to AD pathogenesis [7],as the inflammatory proteins apolipoprotein E (ApoE)andα1antichymotrypsin (ACT) catalyze the polymer-ization of Aβ peptides into amyloid filamentsin vivoand in vitro [8–11], and ACT has been shown to in-duce the phosphorylation of tau [12]. Conversely, ithas also been shown that amyloid plaques form rapid-ly and then become decorated by microglia [13,14],both resident and bone marrow-derived, suggesting anability and intention to remove amyloid [15–17]. Thusit is unclear whether neuroinflammation is deleteriousor beneficial in the AD brain, and indeed the role ofmicroglia in AD is complex and may involve differentstates of activation with different activities.

RA is an autoimmune disease in which inflamed syn-ovial tissue and highly vascularized pannus form, ir-reparably damaging the cartilage and bone. In this in-flammatory pannus, leukocyte populations are greatlyexpanded, perhaps as an endogenous, but ineffectiveattempt to remove the inflammatory insult. As a result,many proinflammatory factors are produced that worktogether in feedforward mechanisms, further increas-

ing leukocytosis, cytokine/chemokine release, osteo-clastogenesis, angiogenesis, and autoantibody produc-tion (rheumatoid factors and anticitrullinated proteinantibodies) [18,19]. Additionally, the adaptive immunesystem presents a Th17 phenotype within CD4+ lym-phocytes, with ultimate production of interleukin 17(IL-17), which is then responsible for inducing much ofthe proinflammatory effects [20,21]. Further enhance-ments of leukocyte populations come from increasedexpression of structurally-unrelatedcolony-stimulatingfactors: MCSF (macrophage), GCSF (granulocyte),and GM-CSF (granulocyte-macrophage) [22–25].

Although upregulated leukocytes in response to RAcould potentially enter into the brain and inhibit de-velopment of AD pathology and/or neuronal dysfunc-tion, lymphocytic infiltrates into AD patient brains havenot been reported. The lack of infiltration suggests thatRA-induced proliferation and activation of the innateimmune system by the CSFs described above mightbe responsible for preventing AD pathology in RA pa-tients. Evidence supporting the innate immune system’srole in AD pathogenesis show that complement pro-teins are upregulated in AD brain, and that inhibition ofC3 convertase significantly increases amyloid patholo-gy in AD mice [26]. Bone marrow-derived microgliaalso play a critical role in restricting amyloid depo-sition, and indeed, microglia activation and many as-sociated receptors and enzymes, such as CD36, scav-enger receptor A, and receptor for advanced glycationend products, neprilysin, insulysin, and matrix metal-loproteinases, decline with age as risk of AD pathologyincreases [17,27,28].

To investigate the interplay of the innate immunesystem and AD, we studied the effects on AD pathol-ogy of the three colony-stimulating factors (M-CSF,G-CSF, and GM-CSF), which are upregulated dur-ing RA pathogenesis [22–25]. These CSFs enhancethe survival and function of their respective leuko-cytes and drive their proliferation and differentiationfrom myeloid lineage precursors. GM-CSF inducesdendritic cells, macrophages, and granulocytes (neu-trophils, basophils, and eosinophils), while MCSF andGCSF, respectively, induce the macrophage and gran-ulocyte subsets of the innate immune system. Theseinnate cells have the ability to diapedese from the cir-culatory system and to differentiate further into vari-ous specialized immune cells within organs (microglia,Langerhan’s cells, etc.). GM-CSF and G-CSF are alsoknown to be involved in erythropoiesis, and GM-CSFand erythropoietin act synergistically in the matura-tion and proliferation of the burst-forming and colony-

T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice 509

forming erythroid units to the normoblast stage of ery-thropoiesis [29,30]. Circulating Aβ binds to comple-ment opsonin C3b in an antibody-independent fashion,and C3-opsonized particles bind to the complement re-ceptor, CR1, on erythrocytes and to CR1g on liver-resident kupffer macrophages [31,32]. Thus GM-CSFcould function in both the peripheral clearance of Aβ

and in bone marrow-derived microglial activity, sinceit is involved in the proliferation, differentiation, andmaintenance of most innate leukocytes.

Here, we report on experiments that investigated theeffect of CSF administration on amyloid plaque depo-sition, microglial activation, synaptic function, and as-sociated cognitive decline in a mouse model of AD.Our results, particularly with GM-CSF, provide a com-pelling explanation for RA’s inverse relationship withAD. Moreover, the reduction of amyloidosis and en-hancement of cognition by GM-CSF warrant clini-cal investigation of Leukine for the treatment of mildcognitive impairment (MCI) and AD patients, espe-cially with Leukine’s long-standing safety history inleukopenic patients.

MATERIALS AND METHODS

All procedures involving experimentation on ani-mals were performed in accordance with the guidelinesset forth by the University of South Florida AnimalCare and Use Committee. Transgene detections wereperformed using QPCR (Bio-Rad iCycler, Hercules,CA).

Transgenic mouse studies involving intracerebraladministration of CSFs

PS/amyloid-β protein precursor (AβPP) mice inthis study, which begin accumulating robust amyloidplaques at 6–8 months, were generated by crossing het-erozygous PDGF-hAβPP (V717F) mice with PDGF-hPS1 (M146L) on both Swiss Webster and C57BL/6backgrounds.

Bilateral intracerebroventricular infusion of M-CSF

M-CSF was bilaterally infused directly into thelateral ventricles (5µg/day) for 14 days using anovel intracranial catheter infusion system (patentpending: PCT/US08/73974) [33]. This completelysubcutaneously-contained system allows bilateral in-tracerebral infusion of test substances ipsilaterally

and vehicle contralaterally, and overcomes the prob-lem of amyloidosis variance between animals (Sup-plementary Fig. 1; available online: http://www.j-alz.com/issues/21/vol21-2.html#supplementarydata03),ef-fectively making each animal its own control. Briefly,animals (PS/AβPP, all 8.8–9.6 months, numbered se-quentially according to date of birth, 25–35 g, both gen-ders) were anesthetized with 1–2% isoflurane, shavedand scrubbed with 10% Betadine solution at the siteof incision, and placed into a stereotaxic frame (KopfInstruments, Tujunga, Ca.). A small (3 cm) incisionwas made, exposing the skull, and curved Strabysmussurgical scissors were used to form a subcutaneouspocket along the animal’s back into which 2 osmoticminipumps (Alzet model 1004, average flow rate of0.12µL/h, Durect Corp., Cupertino, CA) were inserted.Two holes were drilled into the skull (from Bregma±0.1 mm anterior-posterior,± 0.9 mm medial-lateral),and 30 gauge catheters were inserted at a depth of3.0 mm, corresponding to the lateral ventricles. Lead-ing from the Alzet pump was a proprietary cathetersystem with the delivery tips fashioned to the con-tours of the skull rather than the commercially-availablepedestal cannula. The cannulae are affixed to the skullusing Locktite 454 adhesive (Plastics One, Roanoke,VA) and secured with 1 cm diameter nitrile, followedby silk sutures to close the scalp.

After 2 weeks of M-CSF infusions, mice were per-fused, brain tissues were fixed in 10% neutral bufferedformalin, and cryosectioned at 14µm.

Intrahippocampal injections of CSFs

All three CSFs were stereotaxically-injected (5µg/injection) into the (ipsilateral) hippocampus, with arti-ficial cerebrospinal fluid vehicle injected contralateral-ly into four PS/AβPP mice each (all 10–12 months old,25–35 g, both genders). Two holes were drilled intothe skull (from bregma± 2.5 mm anterior-posterior,±2.5 mm medial-lateral, and the 30 gauge needle insert-ed to a depth of 2.5 mm). Mice were perfused with0.9% cold saline 7 days later and their brains placedin 10% neutral buffered formalin. Recombinant mouseGM-CSF (rmGM-CSF), recombinant murine G-CSF(rmG-CSF), and recombinant mouse M-CSF (rmM-CSF) (R&D Systems, Minneapolis, MN) will be re-ferred to as GM-CSF, G-CSF, and M-CSF throughoutthis publication.

510 T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice

Immunohistochemistry and image analysis ofintrahippocampal-injected mice

Formalin-fixed brains were either coronally cryosec-tioned at 14µm, or paraffin-embedded and sectionedat 5µm, with standard deparaffination and antigen re-trieval steps (boiled in 10 mM sodium citrate buffer for20 min) performed before immunohistochemical stain-ing. To significantly reduce cost of reagents and an-tibodies with paraffin-embedded slides, a novel mag-netic immunohistochemical staining device was devel-oped (patent pending, Tech ID# 09A015). Standard flu-orescent immunohistochemical techniques used prima-ry anti-Aβ antibodies 6E10 (Covance, Emeryville, CA,1:1000), and MabTech’s 3740-5 (MabTech, Cincin-nati, OH, 1:5000) to immunolabel amyloid deposi-tion coupled with Alexa fluorophore-labeledsecondaryantibodies (Molecular Probes, Eugene, OR, 1:1000,1:4000), and Hoechst (Sigma) nuclear staining. Im-munofluorescence was detected and all pictures per sec-tion were taken at the same exposure on a Zeiss ImagerZ1 microscope (Oberkochen, Germany) using Axiovi-sion 4.7 software. Digital images were quantified usingImageJ (method described in Supplementary Fig. 6).Briefly, each analyzed picture per coronal section wasthresholded equally to the same standard deviation fromthe histogram mean, and analyzed for area, perime-ter, feret diameter, and integrated density parametersof each plaque. Area and Perimeter data were calcu-lated from the total and average number of plaque val-ues in each hemisphere per section, and Feret Diameterand Integrated Density were calculated from the aver-age values of the plaques measured in each hemisphereper section. For GM-CSF-injected mice, each sectionquantified contained analysis of 15–25 individual 10Xpictures of each hemisphere, with fewer pictures quan-tified in the anterior brain and more in posterior brain.For G-CSF- and M-CSF-injected mice, each sectionquantified contained analysis of 7–9 individual 5X pic-tures of each hemisphere Statistical significance wasobtained from comparing parameter values of ipsilat-eral CSF-injected hemispheres versus contralateral ar-tificial cerebrospinal fluid-injected hemispheres. Sig-nificance was determined by paired student’s ttest withp values< 0.5 considered significant.

Behavioral transgenic mouse study involvingGM-CSF treatment

Mice in this study were derived from the FloridaAlzheimer’s Disease Research Center mouse colony,

wherein heterozygous mice carrying the mutantAβPPK670N, M671L gene (AβPPsw) are routinelycrossed with heterozygous PS1 (line 6.2) mice to ob-tain AβPPsw/PS1, AβPPsw, PS1, and non-transgenic(NT) genotypeoffspring with a mixed C57/B6/SW/SJLbackground. Eleven AβPPsw (Tg) and 17 NT mice, all12-months old, were selected and evaluated for 8 daysin the radial arm water maze (RAWM) task of workingmemory (as previously described [34] (SupplementaryFig. 7). Briefly, an aluminum insert was placed intoa 100 cm circular pool to create 6 radially distribut-ed swim arms emanating from a central circular swimarea. The number of errors prior to locating whichone of the 6 swim arms contained a submerged escapeplatform (9 cm diameter) was determined for 5 trialsper day. The platform location was changed daily toa different arm, with different start arms for each ofthe 5 trials semi-randomly selected from the remain-ing 5 swim arms. The numbers of errors during trials4 and 5 are both considered indices of working mem-ory and are temporally similar to the standard regis-tration/recall testing of specific items used clinicallyin evaluating AD patients. Following 8 days of pre-treatment RAWM testing, the 11 Tg mice were dividedinto two groups balanced in RAWM performance. The17 NT mice were also divided into two groups balancedin RAWM performance.

Two weeks following pretreatment testing, onegroup of Tg mice (n = 5) and one group of NT mice(n = 9) were started on a 10-day treatment protocolwith GM-CSF (5µg/day given subcutaneously), whileanimals in the control Tg group (n = 6) and controlNT group (n = 8) concurrently received daily vehicle(saline) treatment subcutaneously. On the 11th day ofinjections, all mice began four days of RAWM evalua-tion, were given 2 days of rest, then evaluated for 4 daysin Cognitive Interference task testing as previously de-scribed [35,36]. We designed this task measure-for-measure from a Cognitive Interference task used to dis-criminate normal aged, MCI, and AD patients fromone another [36]. Our analogous interference task formice RAWM was set-up in two different rooms, eachwith different sets of visual cues. The task requires an-imals to remember a set of visual cues, so that follow-ing interference with a different set of cues, the ini-tial set of cues can be recalled to successfully solvethe RAWM task. A set of four behavioral measureswas examined. Behavioral measures were: “A1–A3”(Composite three-trial recall score from first 3 trialsperformed in RAWM “A”), “B” (proactive interferencemeasure attained from a single trial in RAWM “B”),

T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice 511

“A4” (retroactive interference measure attained duringa single trial in RAWM “A”), and “A5” (delayed-recallmeasure attained from a single trial in RAWM “A” fol-lowing a 20 minute delay between “A4” and “A5”). Asa distracter between trials, animals are placed in a Y-maze and allowed to explore for 60 s between succes-sive trials of the three-trial recall task, as well as duringthe proactive interference task. As with the standardRAWM task, this interference task involves the plat-form location being changed daily to a different arm forboth of the RAWM set-ups utilized, and different startarms for each day of testing for both RAWM set-ups.For A1 and B trials, the animal was initially allowed1 min to find the platform on their own before theywere guided to the platform. Then the actual trial wasperformed in each case. As with the standard RAWMtask, animals were given 60 s to find the escape plat-form for each trial, with the number of errors recordedfor each trial. Animals were tested for cognitive in-terference performance on four successive days, withstatistical analysis performed for the two resultant 2-day blocks. For both RAWM (combined T4 and T5overall) and cognitive interference testing (each of thefour measures overall), swim speed was analyzed bydividing error numbers by latency and statistical signif-icance was determined by one-way ANOVA followedby post hoc Fisher’s LSD (least significant difference)test to determine significant group differences atp <

0.05.Daily GM-CSF and saline injections were continued

throughout the behavioral testing period. After com-pletion of behavioral testing at 20 days into treatment,all mice were euthanatized, brains fixed as describedabove, and paraffin-embedded. Careful visual exam-ination of all tissues upon necropsy revealed no mor-phological abnormalities, and the mice tolerated dailysubcutaneous injections well. Each analysis was doneby a single examiner blinded to sample identities, andstatistical analyses were performed by a single examin-er blinded to treatment group identities. The code wasnot broken until analyses were completed.

Immunohistochemistry and image analysis ofsubcutaneous GM-CSF-treated mice

Five 5 µm sections (150µm apart) were made offormalin-fixed, paraffin-embeddedsections throughoutthe hippocampus of each mouse and immunoreactiv-ity was developed using the Vectastain ABCElite kit(Vector Laboratories, Burlingame, CA) coupled withthe diaminobenzidine reaction, according to the manu-

facturer’s protocol. Immunostaining used biotinylatedanti-Aβ clone 4G8 (Covance, Emeryville, CA, 1:200),synaptophysin (DAKO, Carpinteria, CA, undiluted),and Iba1 (Wako, Richmond, VA, 1:1000) as prima-ry antibodies. Since the 4G8 antibody was obtainedwith biotin label, the secondary step of the ABC proto-col was omitted. However, treatment with 70% formicacid prior to the pre-blocking step was necessary. For4G8 immunohistochemistry, phosphate-bufferedsaline(0.1 mM, pH 7.4) was used instead of primary anti-body or ABC reagent as a negative control. For Iba 1and synaptophysin immunohistochemistry,normal rab-bit serum was used instead of primary antibody or ABCreagent as a negative control. Images were acquiredusing an Olympus BX60 microscope and digital im-ages were quantified using SimplePCI software (Com-pix Inc., Imaging Systems, Cranberry Township, PA),according to previous methods [37]. Briefly, a thresholdoptical density was obtained that discriminated stainingfrom background, and each region of interest was man-ually edited to eliminate artifacts. Data are reported aspercentage of immunolabeled area captured (positivepixels) relative to the full area captured (total pixels).To evaluate synaptophysin immunoreactivity, after themode of all images was converted to gray scale, the av-erage intensity of positive signals from each image wasquantified in the CA1 and CA3 regions of hippocampusas a relative number from zero (white) to 255 (black).

Statistical significance between GM-CSF-treatedversus saline-treated groups was determined by two-tailed homoscedastic Student’s t-test with a p value of< 0.05 considered significant. Each analysis was doneby a single investigator blinded to sample identities andgenotype.

RESULTS

Intracerebral administration of CSFs

Following bilateral intracerebroventricular infusionof M-CSF for two weeks into PS/AβPP mice, immuno-histochemical analysis of both experimental and con-trol mice showed considerable variances of amyloid de-position between mice of similar age (SupplementaryFig. 1), significantly compromising our ability to deter-mine M-CSF’s effect in a limited mouse cohort. Whileimprovingour drug delivery system by developingnov-el bilateral brain infusion catheters [33], we found thatparenchymally-infused recombinant peptides remainedlocalized to the infused hemisphere. These findings led

512 T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice

us to administer the CSFs as a unilateral intrahippocam-pal bolus with a contralateral injection of vehicle ascontrol, thus obviating the need for large numbers oftransgenic mice and age-matched littermate controls toobtain statistical significance. Each CSF was stereo-taxically injected into the hippocampus of 4 mice, withartificial cerebrospinal fluid vehicle injected contralat-erally. The mice were sacrificed 7 days post-injection.

M-CSF intrahippocampal injections into mice result-ed in swelling of the entire hemisphere as compared tothe control side, and in one mouse, an apparent hyper-plasia had formed at the injection site (SupplementaryFig. 2). Quantification of amyloid plaque loads fromanterior to posterior of each mouse showed similar de-position in the M-CSF-injected hemispheres as com-pared to the control sides (data not shown). In contrastto M-CSF, G-CSF intrahippocampal injections did notinduce swelling and showed some modest reductionsof amyloid deposition (Supplementary Fig. 3). Sincethe analyses were conducted on low magnification (5X)photomicrographs, only the area and integrated densityvalues achieved significance (p < 0.05). Reduction inamyloidosis was subsequently corroborated by inde-pendent observations from fellow investigators follow-ing peripheral G-CSF administration [37].

GM-CSF-injections, however, demonstrated pro-nounced decreases in amyloid deposition, as comparedto control hemispheres, in visual observations of coro-nal tissue sections (Fig. 1a; Supplementary Fig. 4).High magnification (10X) quantification of amyloidplaques anterior to posterior revealed significant reduc-tions within individual mice and significant overall re-ductions for all plaque parameters measured (Fig. 1band Supplementary Fig. 5).

Daily subcutaneous injection of GM-CSF

Based upon the positive results from intrahippocam-pal injections, we investigated the effect of subcuta-neous GM-CSF injection on AD pathology and cog-nitive function. Prior to GM-CSF treatment, AβPPswmice were first confirmed by RAWM testing to becognitively-impaired for working memory (Fig. 2a).Both the NT control mice and the Tg mice were thensub-divided into two cognitively-balanced groups, foreither GM-CSF or saline treatment. RAWM testingpost-injection re-confirmed that Tg control mice weresubstantially impaired compared to NT control mice.This impairment was not only evident in individualblocks of testing, but also over all 4 days of testing(Fig. 2b). In sharp contrast, GM-CSF-treated Tg mice

performed equally well or better than NT control miceduring individual blocks and overall. GM-CSF-treatedNT mice performed as well as or slightly better thanNT controls (Fig. 2b).

Before evaluation in the Cognitive Interference Task,the mice rested two days. This task mimics human in-terference testing, which discriminates between normalaged, MCI, and AD patients [36]. In three of four cog-nitive interference measures assessed over 4 days oftesting (Fig. 2c), Tg control mice were clearly impairedcompared to NT mice, and Tg mice treated with GM-CSF exhibited significantly better three-trial recall anddelayed recall compared to Tg controls. Indeed, for allfour cognitive measures, GM-CSF-treated transgenicAD mice performed similarly to NT mice. A partic-ularly strong effect of GM-CSF treatment in Tg micewas evident for the proactive interference measure dur-ing the first half of testing (Fig. 2d), wherein GM-CSF-treated Tg mice performed substantially better than Tgcontrols and identically to both groups of NT mice. In-cluding this strong effect on proactive interference test-ing, GM-CSF treatment resulted in significantly bet-ter performance of Tg mice for all four measures ofcognitive interference testing. Proactive interferencesusceptibility has been reported to be a more sensitivemarker for differentiating MCI and AD patients fromaged controls than traditional measures of delayed re-call and rate of forgetting [36]. Parenthetically, eventhe GM-CSF-treated NT mice showed a trend towardsimprovedcognition in behavioral studies, albeit not sta-tistically significant. Analysis of swim speed for boththe RAWM and cognitive interference tasks revealedthat Tg control mice were significantly faster than theother three groups in the RAWM task and for two offour measures in the Cognitive Interference task (3-trialrecall and delayed recall). However, since error num-bers were utilized for statistical analysis of both tasks,this difference in swim speed was negated since it isimportant only if latency measures had been used.

Following completion of all behavioral evaluations,subsequent analysis of brains from Tg mice of thisstudy revealed that GM-CSF treatment induced largereductions in amyloid burdens within entorhinal cortex(↓55%) and hippocampal (↓57%) compared to controlTg mice (Fig. 3). The improved cognitive function andreduced cortical amyloidosis of GM-CSF-treated Tgmice were paralleled by increased microglial density ascompared to saline-treated Tg mice (Fig. 4), implyingan augmentedability to bind and removeamyloid depo-sition [27,28]. The GM-CSF-treated Tg mice similar-ly demonstrated increased synaptophysin immunore-

T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice 513

Fig. 1. Intrahippocampal injection of GM-CSF (left) and artificial cerebrospinal fluid (aCSF) (right). a) Representative coronal tissuecryo-sectioned at 14µm and stained with MabTechα-Aβ/Alexa 546. Image is a montage of about 145 pictures taken at 10X. White spotsindicate amyloid plaque immunolabeling (see Supplementary Fig. 4 online for representative montaged sections of all 4mice). b) Significantoverall plaque reductions seen in all 4 plaque parameters measured from 5 quantified sections per mouse (n = 4 mice). Error bars are± StandardError of the Mean. (Area:p < 1.11E-07; Perimeter:p < 1.41E-06; Feret Diameter:p < 2.36E-09; Integrated Density:p < 1.11E-07).

activity in both CA1 and CA3 regions (Fig. 5), indi-cating increased synaptic area in these hippocampalregions. Prior work has shown that adult neural stemcells in hippocampal dentate gyrus (DG) express GM-CSF receptors, and GM-CSF increases neuronal dif-ferentiation of these cells in a dose-dependent fash-ion [38]. Thus, one mechanism for the observed GM-CSF-induced cognitive improvement is enhanced re-moval of deposited Aβ in hippocampus, with ensuingneuronal growth/synaptic differentiation of DG mossyfiber innervation to CA3, resulting in increased inner-vation/synaptogenesis of Schaffer collaterals into CA1.Removal of deposited Aβ from entorhinal cortex mayalso increase perforant pathway viability to hippocam-pal projection fields in DG and CA1.

Thus GM-CSF-induced reduction of amyloidosisand enhancementof hippocampal/entorhinalcortex cir-cuitry, critical for working (short-term) memory, mayunderlie GM-CSF’s reversal of working memory im-pairment in Alzheimer’s Tg mice.

DISCUSSION

Since peripheral leukocyte populations are increasedin RA and possess the ability to infiltrate into the brain,we initially investigated M-CSF, G-CSF, and GM-CSFto determine which CSF might affect amyloidosis afterdirect injection into the brain. In the vasculature, allthree CSFs work to drive the proliferation, differentia-

514 T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice

Fig. 2. Behavioral analysis following daily subcutaneous GM-CSF injections. a) Standard RAWM errors prior to treatment. The final block andoverall performance of the Tg and NT mice during 8 days of consecutive daily pretreatment testing in the RAWM maze. The data was analyzed infour 2-day blocks and overall (Blocks 1-4). (∗p < 0.02 or higher significance). b) Standard RAWM errors after treatment. Tg control mice (n =

6) show substantial impairment on working memory trials T4 and T5 compared to NT control mice (n = 8) in individual blocks of testing (upper),and over all 4 days of testing (lower). GM-CSF-treated Tg mice (n = 5) performed as well as or better than NT control mice on working memorytrials T4 and T5 during individual blocks and over all. GM-CSF-treated NT mice (n = 9) performed similarly to or slightly better than NTcontrols (Note significantly better performance of NT+GM-CSF group versus NT group for T4 of Block 1), although this effect was not significantoverall. (∗∗p < 0.02 or higher significance versus all other groups;†p < 0.02 or higher significance versus Tg+GM-CSF and NT+GM-CSF).c) Cognitive Interference Task. Overall (4 Days) Tg controlmice are impaired compared to NT mice on all four cognitive measures assessed.GM-CSF-treated Tg mice exhibited significantly better 3-trial recall (A1-A3) and delayed recall (A5) compared to Tg controls and performedsimilarly to NT mice in all four cognitive measures. GM-CSF treatment of NT mice did not result in significantly better performance comparedto NT controls, although trends for a beneficial GM-CSF effect in NT mice were evident overall. (∗Tg significantly different from NT+GM-CSF,∗∗Tg significantly different from all other groups). d) Cognitive Interference Task. Proactive Interference testing (First 2 days). GM-CSF-treatedTg mice performed significantly better than Tg controls and equally to NT and GM-CSF-treated NT mice. (Colours are visible in the electronicversion of the article at www.iospress.nl.)

tion, and survival of their respective innate leukocytesfrom monocytic precursors and would be expected tohave similar effects on microglial precursors.

In our study, we found different functional effectsfor each intrahippocampal-injected CSF. In the MCSFinjected mice, there was no effect on amyloidosis, butpathological changes were noticed, such as swellingand hyperplasia (Supplementary Fig. 2). Parenchymal

overexpression of MCSF in any organ is probably notadvisable as overexpression of MCSF and/or its recep-tor within mammary glands has similarly resulted intumor formation and hyperplasia [39]. In a study byBoissonneault et al. (2009), the authors published thatchronic intraperitoneal injection of MCSF prevents andreverses amyloid deposition and cognitive impairmentand induces a large accumulation of bone marrow-

T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice 515

Fig. 3. Amyloid deposition in subcutaneous GM-CSF-injected mice.a-d) Photomicrographs of coronal 5µm paraffin-embedded sectionsimmunolabelled with anti-Aβ antibody (clone 4G8) in entorhinalcortex (EC) and hippocampus (H). Pictures are representative ofamyloid load closest to the mean of the GM-CSF- or saline-treated Tggroups. Scale bar= 50 µm. e) Percent of amyloid burden from theaverage of five 5µm sections (150µm apart) through both anatomicregions of interest (hippocampus and entorhinal cortex) per mouse ofGM-CSF-treated (n = 5) versus saline-treated (n = 6). Entorhinalcortex (∗p < 0.026), and hippocampus (p = 0.12).

derived microglia in the brain [40]. The authors alsoconfirmed previous research [17] that bone marrow-derived microglia efficiently phagocytose and internal-ize Aβ. Although the authors did not relate their find-ings to RA’s inverse relationship with AD, their dataprovide evidence that upregulated MCSF in RA patho-genesis and systemic administration of M-CSF in ADpatients may impart protection against AD onset orprogression. However, M-CSF’s activation of matureosteoclasts, macrophages, and other innate leukocytes,induced thrombocytopenia, and autocrine signaling insome tumors may limit its usage in the clinical setting.

Peripheral administration of G-CSF has also beenfound to ameliorate amyloid pathology and reduce cog-

Fig. 4. Microglial immunostaining in subcutaneous GM-CSF-inject-ed mice. a-d) Photomicrographs of coronal 5µm paraffin-embeddedsections immunolabeled with Iba-1 antibody in entorhinal cortex(EC) and hippocampus (H). Pictures are representative of Iba-1 im-munolabeling closest to the mean of the GM-CSF- or saline con-trol-treated groups. Scale bar= 50 µm. e) Percent of Iba1 burdenfrom the average of five 5µm sections (150µm apart) through bothanatomic regions of interest per mouse of GM-CSF-treated (n = 5)versus saline-treated (n = 6). Hippocampus (p < 0.02), entorhinalcortex (p < 0.05).

nitive defects in AD models [37,41], which corrobo-rates our observations of modest amyloid reduction byintrahippocampal injection. In the study by Sanchez-Ramos et al. [37], the authors also show a large increasein bone marrow-derived microglia with correspondingamyloid reductions, increased synaptic area, and partialreversal of cognitive impairment.

Although the M-CSF and G-CSF intrahippocampalfindings are encouraging, our GM-CSF intrahippocam-pal injections into an AD mouse model demonstrat-ed a much more pronounced reduction of amyloidosis.These data led us to further examine GM-CSF on thebehavior of aged cognitively-impaired AD mice. Sub-cutaneous administration of GM-CSF resulted in al-

516 T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice

Fig. 5. Synaptophysin immunostaining in subcutaneous GM-CSF-injected mice. a-d) Photomicrographs of coronal 5µm paraffin-embedded sections immunolabeled with antisynaptophysin antibody.Pictures are representative of synaptophysin immunolabeling closestto the mean of the GM-CSF- or saline control-treated groups.Scalebar= 50 µm. e) Percent of synaptophysin immunoreactivity fromthe average of 5 sections per mouse of GM-CSF-treated (n = 5)versus saline control-treated (n = 6). CA1 (p < 0.0013), CA3 (p <

0.0023). (Colours are visible in the electronic version of the articleat www.iospress.nl.)

most complete reversal of cognitive impairment, result-ing in function similar to that of wild-type mice, with acorresponding average of about 50% reduction of amy-loidosis in entorhinal cortex and hippocampus. In con-trast, M-CSF and G-CSF have been shown to only par-tially reverse cognitive impairment in aged AD mice,although the preventative chronic study of M-CSF ad-ministration into young AD mice showed equal cogni-tive function over time compared to wild-type controls.This cognitive advantage of peripherally-administeredGM-CSF into aged AD mice could result from a com-binatorial effect of increased macrophage and granulo-cyte populations from bone marrow-derived monocyticprecursors in the periphery, as well as induction of oth-

er innate leukocyte subsets in both periphery and brain.Indeed, GM-CSF has been shown to pass the bloodbrain barrier [42], and a recent study showed that GM-CSF injected into the brains of normal mice activatesmicroglia [43]. Further explanations for the very ro-bust cognitive benefits of GM-CSF in our experimentsare multiple, including the aforementioned augmen-tation of peripheral erythropoietic amyloid-clearancemechanisms [31,32,44], as well as increased neurogen-esis [38], increased cerebral angiogenesis [45], neuro-protection from apoptosis [46], reduction in amyloido-sis (Figs 1,3, and Supplementary Fig. 5), and increasedneuronal plasticity (Fig. 5).

Our results mirror those reported for G-CSF [37],and supports the current practice of interchangeableprescription of either recombinant human GM-CSF(Leukine) or GCSF (Granocyte, Neupogen, and Neu-lasta) into patients with depressed bone marrow func-tion. G-CSF primarily treats neutropenia while GM-CSF treats all leukopenia, and both have long recordsof safety data from two decades of FDA-approved us-age. Rare adverse events are usually mild febrile inci-dents that quickly subside upon cessation of adminis-tration. G-CSF is currently in clinical trial for strokeand was recently approved for an AD Phase II clini-cal trial. However, GM-CSF/Leukine is more effec-tive in the AD mouse model and while neutrophils areshort-lived leukocytes, the fact that GM-CSF inducesthe upregulation of all innate cells means that it couldpotentially impart prolonged protective effects againstAD.

The failure of NSAID clinical trials in AD andthe multiple studies that show that a defective in-nate immune system propagates AD pathology en-couraged us to develop and test our hypothesis thatintrinsic pathogenic properties of RA are protectiveagainst AD. Indeed the beneficial effects of all threeRA-upregulated CSFs, especially GM-CSF, in mousemodels of AD point to a potential new approach toAD therapy and indicate that age-linked depressedhematopoiesis may be etiological for AD pathogenesis.

ACKNOWLEDGMENTS

Support for this work was provided for by the USFHealth Byrd Alzheimer’s Center and Research Institute,the Eric Pfeiffer Chair for research on Alzheimer’s dis-ease, the Florida Alzheimer’s Disease Research Center(P50AG25711),and the James H. and Martha M. PorterAlzheimer’s Fund.

Authors’ disclosures available online (http://www.j-alz.com/disclosures/view.php?id=380).

T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice 517

REFERENCES

[1] McGeer PL, Rogers J, McGeer EG (2006) Inflammation, anti-inflammatory agents and Alzheimer disease: the last 12 years.J Alzheimers Dis9, 271-276.

[2] Martin BK, Szekely C, Brandt J, Piantadosi S, Breitner JC,Craft S, Evans D, Green R, Mullan M (2008) Cognitive func-tion over time in the Alzheimer’s Disease Anti-inflammatoryPrevention Trial (ADAPT): results of a randomized, controlledtrial of naproxen and celecoxib.Arch Neurol65, 896-905.

[3] Okazaki H, Reagan TJ, Campbell RJ (1979) Clinicopathologicstudies of primary cerebral amyloid angiopathy.Mayo ClinProc 54, 22-31.

[4] Hardy J, Selkoe DJ (2002) The amyloid hypothesis ofAlzheimer’s disease: progress and problems on the road totherapeutics.Science297, 353-356.

[5] Glenner GG, Wong CW (1984) Alzheimer’s disease andDown’s syndrome: sharing of a unique cerebrovascular amy-loid fibril protein. Biochem Biophys Res Commun122, 1131-1135.

[6] Lee VM (1996) Regulation of tau phosphorylation inAlzheimer’s disease.Ann N Y Acad Sci777, 107-113.

[7] Griffin WS, Stanley LC, Ling C, White L, MacLeod V, PerrotLJ, White CL, 3rd, Araoz C (1989) Brain interleukin 1 andS-100 immunoreactivity are elevated in Down syndrome andAlzheimer disease.Proc Natl Acad Sci U S A86, 7611-7615.

[8] Wisniewski T, Castano EM, Golabek A, Vogel T, FrangioneB (1994) Acceleration of Alzheimer’s fibril formation byapolipoprotein Ein vitro. Am J Pathol145, 1030-1035.

[9] Ma J, Yee A, Brewer HB, Das S, Potter H (1994) TheAlzheimer amyloid-associated proteins al-antichymotrypsinand apolipoprotein E promote the assembly of the Alzheimerβ-protein into filaments.Nature372, 92-94.

[10] Potter H, Wefes IM, Nilsson LN (2001) The inflammation-induced pathological chaperones ACT and apo-E are neces-sary catalysts of Alzheimer amyloid formation.Neurobiol Ag-ing 22, 923-930.

[11] Nilsson LN, Arendash GW, Leighty RE, Costa DA, Low MA,Garcia MF, Cracciolo JR, Rojiani A, Wu X, Bales KR, PaulSM, Potter H (2004) Cognitive impairment in PDAPP mice de-pends on ApoE and ACT-catalyzed amyloid formation.Neu-robiol Aging25, 1153-1167.

[12] Padmanabhan J, Levy M, Dickson DW, Potter H (2006)Alpha1-antichymotrypsin, an inflammatory protein overex-pressed in Alzheimer’s disease brain, induces tau phosphory-lation in neurons.Brain 129, 3020-3034.

[13] Koenigsknecht-Talboo J, Meyer-Luehmann M, ParsadanianM, Garcia-Alloza M, Finn MB, Hyman BT, Bacskai BJ, Holtz-man DM (2008) Rapid microglial response around amyloidpathology after systemic anti-Abeta antibody administrationin PDAPP mice.J Neurosci28, 14156-14164.

[14] Meyer-Luehmann M, Spires-Jones TL, Prada C, Garcia-Alloza M, de Calignon A, Rozkalne A, Koenigsknecht-TalbooJ, Holtzman DM, Bacskai BJ, Hyman BT (2008) Rapid ap-pearance and local toxicity of amyloid-beta plaques in a mousemodel of Alzheimer’s disease.Nature451, 720-724.

[15] Malm TM, Koistinaho M, Parepalo M, Vatanen T, Ooka A,Karlsson S, Koistinaho J (2005) Bone-marrow-derived cellscontribute to the recruitment of microglial cells in responseto beta-amyloid deposition in APP/PS1 double transgenicAlzheimer mice.Neurobiol Dis18, 134-142.

[16] Simard AR, Rivest S (2004) Bone marrow stem cells have theability to populate the entire central nervous system into fullydifferentiated parenchymal microglia.FASEB J18, 998-1000.

[17] Simard AR, Soulet D, Gowing G, Julien JP, Rivest S (2006)Bone marrow-derived microglia play a critical role in restrict-ing senile plaque formation in Alzheimer’s disease.Neuron49, 489-502.

[18] Szekanecz Z, Koch AE (2007) Macrophages and their productsin rheumatoid arthritis.Curr Opin Rheumatol19, 289-295.

[19] Schellekens GA, Visser H, de Jong BA, van den Hoogen FH,Hazes JM, Breedveld FC, van Venrooij WJ (2000) The diag-nostic properties of rheumatoid arthritis antibodies recogniz-ing a cyclic citrullinated peptide.Arthritis Rheum43, 155-163.

[20] Parsonage G, Filer A, Bik M, Hardie D, Lax S, HowlettK, Church LD, Raza K, Wong SH, Trebilcock E, Scheel-Toellner D, Salmon M, Lord JM, Buckley CD (2008) Pro-longed, granulocyte-macrophage colony-stimulating factor-dependent, neutrophil survival following rheumatoid synovialfibroblast activation by IL-17 and TNFalpha.Arthritis ResTher10, R47.

[21] Cox CA, Shi G, Yin H, Vistica BP, Wawrousek EF, ChanCC, Gery I (2008) Both Th1 and Th17 are immunopathogenicbut differ in other key biological activities.J Immunol180,7414-7422.

[22] Xu WD, Firestein GS, Taetle R, Kaushansky K, Zvai-fler NJ (1989) Cytokines in chronic inflammatory arthri-tis. II. Granulocyte-macrophage colony-stimulating factor inrheumatoid synovial effusions.J Clin Invest83, 876-882.

[23] Nakamura H, Ueki Y, Sakito S, Matsumoto K, Yano M,Miyake S, Tominaga T, Tominaga M, Eguchi K (2000) Highserum and synovial fluid granulocyte colony stimulating factor(G-CSF) concentrations in patients with rheumatoid arthritis.Clin Exp Rheumatol18, 713-718.

[24] Olszewski WL, Pazdur J, Kubasiewicz E, Zaleska M, CookeCJ, Miller NE (2001) Lymph draining from foot joints inrheumatoid arthritis provides insight into local cytokineandchemokine production and transport to lymph nodes.ArthritisRheum44, 541-549.

[25] Kawaji H, Yokomura K, Kikuchi K, Somoto Y, Shirai Y(1995) [Macrophage colony-stimulating factor in patientswithrheumatoid arthritis].Nippon Ika Daigaku Zasshi62, 260-270.

[26] Wyss-Coray T, Yan F, Lin AH, Lambris JD, AlexanderJJ, Quigg RJ, Masliah E (2002) Prominent neurodegenera-tion and increased plaque formation in complement-inhibitedAlzheimer’s mice.Proc Natl Acad Sci U S A99, 10837-10842.

[27] Hickman SE, Allison EK, El Khoury J (2008) Microglial dys-function and defective beta-amyloid clearance pathways inaging Alzheimer’s disease mice.J Neurosci28, 8354-8360.

[28] El Khoury J, Toft M, Hickman SE, Means TK, Terada K,Geula C, Luster AD (2007) Ccr2 deficiency impairs microglialaccumulation and accelerates progression of Alzheimer-likedisease.Nat Med13, 432-438.

[29] Emerson SG, Sieff CA, Wang EA, Wong GG, Clark SC,Nathan DG (1985) Purification of fetal hematopoietic progen-itors and demonstration of recombinant multipotential colony-stimulating activity.J Clin Invest76, 1286-1290.

[30] Wu H, Liu X, Jaenisch R, Lodish HF (1995) Generation ofcommitted erythroid BFU-E and CFU-E progenitors does notrequire erythropoietin or the erythropoietin receptor.Cell 83,59-67.

[31] Rogers J, Li R, Mastroeni D, Grover A, Leonard B, Ahern G,Cao P, Kolody H, Vedders L, Kolb WP, Sabbagh M (2006)Peripheral clearance of amyloid beta peptide by complementC3-dependent adherence to erythrocytes.Neurobiol Aging27,1733-1739.

[32] Helmy KY, Katschke KJ, Jr., Gorgani NN, Kljavin NM, ElliottJM, Diehl L, Scales SJ, Ghilardi N, van Lookeren Campagne

518 T.D. Boyd et al. / Arthritis CSF Reverses AD in Mice

M (2006) CRIg: a macrophage complement receptor requiredfor phagocytosis of circulating pathogens.Cell 124, 915-927.

[33] Bennett SP, Boyd TD, Norden M, Padmanabhan J, Neame P,Wefes I, Potter H (2009) A novel technique for simultaneousbilateral brain infusions in a mouse model of neurodegenera-tive disease.J Neurosci Methods184, 320-326.

[34] Arendash GW, King DL, Gordon MN, Morgan D, Hatcher JM,Hope CE, Diamond DM (2001) Progressive, age-related be-havioral impairments in transgenic mice carrying both mutantamyloid precursor protein and presenilin-1 transgenes.BrainRes891, 42-53.

[35] Echeverria V, Burgess S, Gamble-George J, Zeitlin R, LinX, Cao C, Arendash GW (2009) Sorafenib inhibits nuclearfactor kappa B, decreases inducible nitric oxide synthase andcyclooxygenase-2 expression, and restores working memoryin APPswe mice.Neuroscience162, 1220-1231.

[36] Loewenstein DA, Acevedo A, Luis C, Crum T, Barker WW,Duara R (2004) Semantic interference deficits and the detec-tion of mild Alzheimer’s disease and mild cognitive impair-ment without dementia.J Int Neuropsychol Soc10, 91-100.

[37] Sanchez-Ramos J, Song S, Sava V, Catlow B, Lin X, Mori T,Cao C, Arendash GW (2009) Granulocyte colony stimulatingfactor (G-CSF) decreases brain amyloid burden and reversescognitive impairment in Alzheimer’s mice.Neuroscience163,55-72.

[38] Kruger C, Laage R, Pitzer C, Schabitz WR, SchneiderA (2007) The hematopoietic factor GM-CSF (granulocyte-macrophage colony-stimulating factor) promotes neuronaldif-ferentiation of adult neural stem cellsin vitro. BMC Neurosci8, 88.

[39] Kirma N, Luthra R, Jones J, Liu YG, Nair HB, Mandava U,Tekmal RR (2004) Overexpression of the colony-stimulating

factor (CSF-1) and/or its receptor c-fms in mammary glandsof transgenic mice results in hyperplasia and tumor formation.Cancer Res64, 4162-4170.

[40] Boissonneault V, Filali M, Lessard M, Relton J, Wong G,Rivest S (2009) Powerful beneficial effects of macrophagecolony-stimulating factor on beta-amyloid deposition andcog-nitive impairment in Alzheimer’s disease.Brain 132, 1078-1092.

[41] Tsai KJ, Tsai YC, Shen CK (2007) G-CSF rescues the memoryimpairment of animal models of Alzheimer’s disease.J ExpMed204, 1273-1280.

[42] McLay RN, Kimura M, Banks WA, Kastin AJ (1997)Granulocyte-macrophage colony-stimulating factor crossesthe blood–brain and blood–spinal cord barriers.Brain 120(Pt11), 2083-2091.

[43] Reddy PH, Manczak M, Zhao W, Nakamura K, Bebbington C,Yarranton G, Mao P (2009) Granulocyte-macrophage colony-stimulating factor antibody suppresses microglial activity: im-plications for anti-inflammatory effects in Alzheimer’s diseaseand multiple sclerosis.J Neurochem111, 1514-1528.

[44] Fisher JW (2003) Erythropoietin: physiology and pharmacol-ogy update.Exp Biol Med(Maywood) 228, 1-14.

[45] Schneider UC, Schilling L, Schroeck H, Nebe CT, Vajkoczy P,Woitzik J (2007) Granulocyte-macrophage colony-stimulatingfactor-induced vessel growth restores cerebral blood supplyafter bilateral carotid artery occlusion.Stroke38, 1320-1328.

[46] Schabitz WR, Kruger C, Pitzer C, Weber D, Laage R, GasslerN, Aronowski J, Mier W, Kirsch F, Dittgen T, Bach A, Som-mer C, Schneider A (2008) A neuroprotective function for thehematopoietic protein granulocyte-macrophage colony stimu-lating factor (GM-CSF).J Cereb Blood Flow Metab28, 29-43.