A

OMa(bTap1gbCa©

K

1

c

anitt

0d

Neurobiology of Aging 28 (2007) 336–342

Determinants of platelet activation in Alzheimer’s disease�

Giovanni Ciabattoni a,c, Ettore Porreca a,b, Concetta Di Febbo a,b, Angelo Di Iorio a,b,Roberto Paganelli a,b, Tonino Bucciarelli a,b, Lea Pescara a,b, Letizia Del Re d,

Cinzia Giusti e, Angela Falco a,b, Antonella Sau a,b, Carlo Patrono f,∗, Giovanni Davı a,b

a Center of Excellence on Aging, “G. d’Annunzio” University Foundation, Italyb Department of Medicine, University of Chieti “G. d’Annunzio”, School of Medicine, Italy

c Department of Drug Sciences, University of Chieti “G. d’Annunzio”, School of Pharmacy, Italyd Department of Geriatrics, Civil Hospital Pescara, Italy

e Department of Laboratory Medicine, Civil Hospital Pescara, Italyf Department of Pharmacology, University of Rome “La Sapienza”, II Facolta di Medicina e Chirurgia, Via di Grottarossa 1035, 00189 Rome, Italy

Received 3 August 2005; received in revised form 7 December 2005; accepted 20 December 2005Available online 24 January 2006

bstract

bjectives: To investigate the rate of platelet thromboxane (TX) biosynthesis and its determinants in Alzheimer’s disease.ethods and results: A cross-sectional comparison of urinary 11-dehydro-TXB2 and 8-iso-prostaglandin (PG)F2� (markers of in vivo platelet

ctivation and lipid peroxidation, respectively), plasma Vitamin E, C-reactive protein (CRP), tumor necrosis factor (TNF)-� and interleukinIL)-6, was carried-out in 44 Alzheimer patients and 44 matched controls. To investigate the cyclooxygenase (COX)-isoform involved in TXA2

iosynthesis, nine Alzheimer patients were treated with low-dose aspirin (100 mg/d) or rofecoxib (25 mg/d) for 4 days. Urinary 11-dehydro-XB2 and 8-iso-PGF2� were significantly higher in Alzheimer patients than in controls (Median: 1983.5 versus 517.5 pg/mg creatininend 938.5 versus 304.0 pg/mg creatinine, p < 0.0001, respectively), with a significant correlation between the two metabolites (ρ = 0.75,< 0.0001). An inverse correlation was observed between Vitamin E and both urinary metabolites (8-iso-PGF2�: Rs = −0.51, p = 0.0004;1-dehydro-TXB2: Rs = −0.44, p = 0.0026) in Alzheimer patients. No difference was found in CRP, TNF-� and IL-6 levels between the tworoups. Urinary 11-dehydro-TXB2 was significantly reduced by aspirin, but not by rofecoxib, consistently with a COX-1-mediated TXA2

iosynthesis. 8-iso-PGF2� excretion was not modified by either COX-inhibitor, consistently with its oxygen radical-catalyzed formation.onclusions: Platelet activation is persistently enhanced in Alzheimer’s disease. This is related, at least in part, to increased lipid peroxidationssociated with inadequate levels of Vitamin E.

2005 Elsevier Inc. All rights reserved.

peroxid

d

eywords: Alzheimer’s disease; Platelet activation; Oxidative stress; Lipid

. Introduction

Alzheimer’s disease is often accompanied by vascularhanges such as small vessel disease due to arteriosclerosis,

� All authors disclose any actual or potential conflicts of interest includingny financial, personal or other relationships with other people or orga-izations within 3 years of beginning the work that could inappropriatelynfluence it. The study was approved by the Medical Ethics Committee ofhe “G. d’Annunzio” University Medical School and conducted accordingo the principles of the Helsinki Declaration.∗ Corresponding author. Tel.: +39 0871 541260; fax: +39 0871 541261.

E-mail address: cpatrono@unich.it (C. Patrono).

ndibsidgifa

197-4580/$ – see front matter © 2005 Elsevier Inc. All rights reserved.oi:10.1016/j.neurobiolaging.2005.12.011

ation; Inflammation

emonstrated to be associated with the development of cog-itive impairment [16]. Stroke is associated with Alzheimer’sisease among elderly individuals. This relation is strongestn the presence of known vascular risk factors, such as dia-etes or hypertension [16]. The observed association betweentroke and Alzheimer’s disease might reflect an underly-ng systemic vascular disease process. In fact, Alzheimer’sisease and atherothrombosis share some common patho-

enetic mechanisms [2,9]. These include the role of chronicnflammation [15], oxidant stress [26] and endothelial dys-unction [1]. In addition, abnormalities in platelet structurend/or function have been reported in both settings [47,17].

iology of Aging 28 (2007) 336–342 337

IptTvpaifi

nliafatamAcsti

2

2

ptbboiotDethimhnMs

mtsmdt

Table 1Demographic and clinical characteristics of Alzheimer’s disease patients andcontrols

Patients(n = 44)

Controls(n = 44)

p

Age (yr)a 73 ± 8 75 ± 7 0.125M/F 19/25 17/27 0.082BMI (kg/m2)a 24.9 ± 1.6 25.5 ± 1.7 0.111MMSEa,b 17 ± 4 29 ± 1 <0.0001Education (yr)a 5 ± 3 4 ± 3 0.122Hypercholesterolemia (n) 8 4 0.214Cigarette smoking (n) 9 9 0.100Cardiovascular disease (n) 18 10 0.109E

hrt

ast

2

81bpptuaiu

ttinfalda

2

r

G. Ciabattoni et al. / Neurob

ncreased thromboxane (TX) biosynthesis in the chronichase after stroke is associated with the presence of, but nothe type of post-stroke dementia [43]. Thus, no difference inX metabolite excretion was found between 44 patients withascular dementia and 18 patients with Alzheimer’s diseaselus cerebrovascular disease. However, the use of low-dosespirin or oral anticoagulants in 80% of the patients enrolledn the study precludes unequivocal interpretation of thesendings [43].

In the present study, we set out to investigate the determi-ants of platelet activation in vivo, as reflected by TX metabo-ite excretion, in a relatively large group of well character-zed patients with probable Alzheimer’s disease not requiringntiplatelet therapy and in a control group of subjects matchedor age, gender, and cardiovascular risk factors potentiallyssociated with persistent platelet activation. In particular, weested the hypothesis that a disturbed oxidant/antioxidant bal-nce and/or abnormal production of inflammatory cytokinesight contribute to a state of persistent platelet activation inlzheimer’s disease, independently of coexisting cardiovas-

ular abnormalities. Furthermore, we performed an ancillaryhort-term cyclooxygenase (COX)-inhibitor study to probehe COX-isoform involved in enhanced TXA2 biosynthesisn this setting.

. Methods

.1. Patients

Healthy elderly individuals and Alzheimer’s diseaseatients were recruited at the Alzheimer’s Disease Centers ofhe “G. d’Annunzio” University and of Pescara Civil Hospitaletween June 2001 and July 2003. The protocol was approvedy the Institutional Review Board and informed consent wasbtained from all participants and their caregivers. The clin-cal diagnosis of probable Alzheimer’s disease was basedn the National Institute of Neurological and Communica-ive Diseases and Stroke-Alzheimer Disease and Relatedisorders Association criteria [25]. The mini-mental state

xamination (MMSE) assessment was performed to evaluatehe clinical severity of the disease [14]. Forty-four patientsad a diagnosis of probable Alzheimer’s disease, and hadmpaired cognitive function as reflected by the MMSE assess-

ent (see Table 1). Forty-four gender- and age-matchedealthy control subjects were recruited from a cohort of cog-itively normal individuals; they scored 28 or higher on theMSE. Urine and blood samples were obtained from all

ubjects.Medical histories including medication, vitamin supple-

ents and physical, neurological, and mental status examina-ions were completed on the day of sample collection. Exten-

ive instrumental examinations were performed, includingagnetic resonance imaging, to exclude other causes of

ementia. Subjects were excluded if they had an acute infec-ious or inflammatory disease, alcohol abuse, cancer, chronic

mic[

ssential hypertension (n) 8 13 0.211a Values are expressed as mean ± S.D.b MMSE: mini-mental state examination.

epatic disease, evidence of other primary psychiatric, neu-ologic or cerebrovascular disorders, or if they were beingreated with antioxidant vitamins.

None of the Alzheimer’s disease patients were takingcethylcholinesterase inhibitors. No subject was taking a non-teroidal antiinflammatory drug or low-dose aspirin at theime of the study.

.2. Design of the studies

A cross-sectional comparison of urinary immunoreactive-iso-PGF2�, a bioactive product of lipid peroxidation, and1-dehydro-TXB2, a major enzymatic metabolite of throm-oxane A2 [4], was performed between Alzheimer’s diseaseatients and controls. All the subjects were studied as out-atients, after a 12-h fast. Blood samples were obtained inhe early morning. Each participant performed an overnightrine collection before blood sampling. Urine samples weredded with the antioxidant 4-hydroxy-Tempo (Sigma Chem-cal Co., St. Louis, MO) (1 mmol/L) and stored at −20 ◦Cntil extraction.

Moreover, an open pilot intervention study was performedo investigate the COX-isoform involved in TXA2 biosyn-hesis and to confirm the non-enzymatic mechanism of F2-soprostane formation in this setting. Thus, we randomizedine patients with Alzheimer’s disease (six males, threeemales; aged 76 ± 6 yr) to a 4-day treatment with low-dosespirin (100 mg/d) or rofecoxib (25 mg/d). These patients col-ected overnight urine samples before dosing and on the lastay of each treatment for measurement of 11-dehydro-TXB2nd 8-iso-PGF2� excretion.

.3. Assays

Urinary 11-dehydro-TXB2 and 8-iso-PGF2� excretionates were measured by previously described radioim-

unoassay methods [5,45]. These methods have been val-

dated using different antisera and by comparison with gashromatography/mass spectrometry, as detailed elsewhere5,45].

3 iology of Aging 28 (2007) 336–342

waamm

2

bWtfaSrsll

sUvDr

septtwv

3

cnhhtc

n(tdawTnrp

Fig. 1. Urinary excretion of 8-iso-prostaglandin (PG)F2� (upper panel) and1Ai

th(bedlpU1sl

lcpllpc

38 G. Ciabattoni et al. / Neurob

Plasma C-reactive protein (CRP) levels were measuredith a highly sensitive immunoassay [20]. Plasma TNF-�

nd IL-6 were measured by enzyme-linked immunosorbentssays (R&D Systems Europe Ltd.). Plasma Vitamin E waseasured by reversed-phase high-performance liquid chro-atography [21].

.4. Statistical analysis

Comparisons among the two groups were performedy nonparametric one-way analysis of variance (Kruskall–allis test). Correlations were examined by the Spearman

est. A multiple-linear regression analysis was performed tourther quantify the relationship between 11-dehydro-TXB2,nd the other variables in the Alzheimer’s disease patients.pecifically, the dependent variable, 11-dehydro-TXB2, wasegressed for gender, age, body mass index (BMI), cigarettemoking, hypercholesterolemia, hypertension, cardiovascu-ar disease, 8-iso-PGF2�, CRP, IL-6, TNF-�, and Vitamin Eevels.

For the odds ratio estimates, multivariate logistic regres-ion analysis was carried out (EGRET by SERC, Seattle, WA,SA). The difference between baseline and post-treatmentalues were analyzed with the Wilcoxon signed-rank test.ata are presented as mean (S.D.) or as median [interquartile

ange (IQR)].With 44 subjects recruited in the cross-sectional compari-

on, the study had at least a 95% power to detect a 50% differ-nce in 11-dehydro-TXB2 urinary levels between Alzheimeratients and control groups with a two-tailed a of 0.05. Statis-ical significance was defined as p < 0.05. All tests were twoailed and analyses were performed using a computer soft-are package (Statistical Package for the Social Sciences,ersion 11.0, SPSS Inc., Chicago, IL).

. Results

The characteristics of the Alzheimer’s disease patients andontrol subjects are detailed in Table 1. No statistically sig-ificant differences were observed between the groups forypercholesterolemia, diabetes, smoking status, and arterialypertension, although some of these cardiovascular risk fac-ors were numerically more represented among patients thanontrols.

Alzheimer’s disease patients had significantly higher uri-ary excretion of 11-dehydro-TXB2 than controls [MedianIQR): 1983.5 (1298.5–3233.5) versus 517.5 pg/mg crea-inine (402.5–653), p < 0.0001] (Fig. 1). All Alzheimer’sisease patients had excretion rates in excess of 2 S.D.bove the control mean. The Alzheimer’s disease patientsithout cardiovascular risk factors had similar 11-dehydro-

XB2 excretion rate [2213 pg/mg creatinine (1341–3439),= 22] as Alzheimer’s disease patients with cardiovascular

isk factors [1910.5 pg/mg creatinine (1277–2898), n = 22,= 0.496].

bupf

1-dehydro-thromboxane (TX)B2 (lower panel) in healthy subjects and inlzheimer’s disease patients. Dots depict individual measurements and hor-

zontal bars indicate mean values.

A similar pattern was observed in the urinary excre-ion of 8-iso-PGF2� (Fig. 1), that was significantlyigher in Alzheimer’s disease patients than controls [938.5665.5–1337.5) versus 304.0 (218.5–348.5), p < 0.0001]. Allut six Alzheimer’s disease patients had excretion rates inxcess of 2 S.D. above the control mean. The Alzheimer’sisease patients without cardiovascular risk factors had simi-ar urinary 8-iso-PGF2� excretion rate as Alzheimer’s diseaseatients with cardiovascular risk factors (data not shown).rinary 8-iso-PGF2� excretion was linearly correlated with1-dehydro-TXB2 excretion (Rs = 0.75, p < 0.0001), thusuggesting a cause-and-effect relationship between enhancedipid peroxidation and platelet activation in this setting.

Plasma Vitamin E levels were significantly (p = 0.0001)ower in Alzheimer’s disease patients (33 ± 15 �mol/L) inomparison with controls (52 ± 13 �mol/L). The ratios oflasma Vitamin E to total cholesterol were also calcu-ated to normalize vitamin concentrations for circulatingipids (5.9 ± 2.9 �mol/mmol versus 8.3 ± 4.8 �mol/mmol,= 0.02). Interestingly, a statistically significant inverseorrelation was observed in Alzheimer’s disease patients

etween plasma Vitamin E/total cholesterol ratio and therinary excretion rates of both metabolites (Rs = −0.320,= 0.038 for 11-dehydro-TXB2 and Rs = −0.445, p = 0.039

or 8-iso-PGF2�).

G. Ciabattoni et al. / Neurobiology o

Table 2Plasma cytokine and CRP levels in Alzheimer’s disease patients and controls

Variable Patients (n = 44) Controls (n = 44) p

Tumor necrosis factor-�(pg/mL)

1.97 (1.20–2.78) 2.60 (1.89–3.80) 0.083

C-reactive protein(mg/L)

0.55 (0.35–1.17) 0.87 (0.60–1.20) 0.089

I

V

nt

Me

ttt

tp

pweow

rcetCtr

4

a

TLw

V

A

G

B

H

H

S

M

C

I

T

8

V

nterleukin-6 (pg/mL) 0.54 (0.20–2.53) 0.95 (0.25–2.85) 0.437

alues are expressed as median (interquartile range).

Circulating levels of CRP, TNF-� and IL-6 were not sig-ificantly different in Alzheimer’s disease patients and con-rols (Table 2).

No statistically significant correlation was found betweenMSE scores and urinary 11-dehydro-TXB2 or 8-iso-PGF2�

xcretion rates.Using multivariate linear regression analysis, we found

hat in Alzheimer’s disease patients 11-dehydro-TXB2 excre-ion was independently correlated with 8-iso-PGF2� excre-

ion (coefficient β = 0.57, S.E. = 0.19, p = 0.007).

Additional analysis of the data was performed using a mul-ivariate logistic regression. High urinary 8-iso-PGF2� (>50thercentile of Alzheimer’s disease patients) and low (<50th

teim

able 3ogistic regression analysis in which high or low values (above and below the 50tere categorical-dependent variable

ariable n 11-D(pg/

<19

gea (yr) <72 18 11≥72 26 11

ender Female 25 11Male 19 11

ody mass indexa <25 15 9≥25 29 13

ypercholesterolemia No 36 18Yes 8 4

ypertension No 36 18Yes 8 4

moking habit No 35 18Yes 9 4

acrovascular complications No 26 15Yes 18 7

RPa (mg/L) <0.55 22 13≥0.55 22 9

L-6a (pg/mL) <0.54 22 10≥0.54 22 12

NF-�a (pg/mL) <1.97 21 9≥1.97 23 13

-iso-PGF2�a (pg/mg creatinine) <938 22 18

≥938 22 4

itamin Ea (�mol/L) ≥32.5 22 16<32.5 22 6

a Above and below the 50th percentile of Alzheimer’s disease patients.

f Aging 28 (2007) 336–342 339

ercentile of Alzheimer’s disease patients) plasma Vitamin E,ere predictors of high (>50th percentile of Alzheimer’s dis-

ase patients) urinary excretion of 11-dehydro-TXB2. Nonef the other examined variables was independently correlatedith urinary excretion of 11-dehydro-TXB2 (Table 3).Urinary 11-dehydro-TXB2 excretion was significantly

educed by low-dose aspirin, but not by rofecoxib (Fig. 2),onsistently with the involvement of platelet COX-1 innhanced TXA2 biosynthesis in Alzheimer’s disease. In con-rast, 8-iso-PGF2� excretion was not modified by eitherOX-inhibitor to any detectable extent, consistently with

he non-enzymatic mechanism of its formation via oxygenadical-catalyzed lipid peroxidation [28].

. Discussion

In the present study, we have characterized a novel mech-nism that may affect both the progression of neurodegenera-

ive disease and cardiovascular morbidity in Alzheimer’s dis-ase patients, i.e. persistent platelet activation. Evidence forn vivo platelet activation was obtained through non-invasiveeasurements of 11-dehydro-TXB2 excretion. The latter rep-

h percentile of Alzheimer’s disease patients) of urinary 11-dehydro-TXB2

ehydro-TXB2

mg creatinine)OR 95% CI p

83 >1983

7 1 –15 0.8 0.4–55.4 0.202

14 1 –8 4.8 0.07–8.3 0.836

6 1 –16 2.6 0.1–54.2 0.533

18 1 –4 0.5 0.01–20.6 0.669

18 1 –4 0.23 0.01–4.2 0.321

17 1 –5 0.5 0.03–9.4 0.721

11 1 –11 1.6 0.2–12.8 0.645

9 1 –13 1.9 0.1–26.8 0.647

12 1 –10 0.2 0.01–2.4 0.204

12 1 –10 0.44 0.05–4.0 0.471

4 1 –18 23.7 2.0–280.0 0.012

6 1 –16 8.8 0.9–84.5 0.059

340 G. Ciabattoni et al. / Neurobiology o

FT(

r[tTibsrrvlmaas2obiTpeadw

ippbh[

ifmp

astbbbpurci

rpe2Twc

uaAtotvtiotpilbdt

of

ig. 2. Urinary excretion of 8-iso-PGF2� (upper panel) and 11-dehydro-XB2 (lower panel), before and after a 4-day treatment with low-dose aspirin

100 mg/d) or rofecoxib (25 mg/d), in nine Alzheimer’s disease patients.

esents a major enzymatic metabolite of TXA2 [3] and TXB219] and provides a time-integrated measure of TXA2 biosyn-hesis in vivo [4]. While platelets represent a major source ofXA2 production in a variety of clinical circumstances [32],

nflammatory cells may also contribute to TXA2 productionecause they are endowed with the specific isomerase, TX-ynthase [29]. Because urinary TX-metabolite excretion mayeflect a heterogeneous cellular source, it is important to useelatively selective pharmacologic tools to probe the plateletersus extra-platelet sources of TXA2 biosynthesis, particu-arly in clinical settings characterized by a chronic inflam-

atory response. Low-dose aspirin given once daily is a rel-tively selective inhibitor of platelet COX-1 activity [30,33]nd reduces TX-metabolite excretion by 70–80% at steadytate [12]. In contrast, rofecoxib is a highly selective COX-

inhibitor [13] and may reduce TX-metabolite excretionnly under conditions where a sizeable proportion of TXA2iosynthesis is contributed by inflammatory cells express-ng COX-2 [24]. Evidence for a platelet origin of enhancedXA2 biosynthesis was confirmed in Alzheimer’s diseaseatients by the profound suppression of 11-dehydro-TXB2

xcretion following a platelet-selective regimen of aspirindministration (Fig. 2). On the contrary, rofecoxib had noetectable effects on TX-metabolite excretion, consistentlyith the expression of COX-1 as the primary COX-isoform

ctod

f Aging 28 (2007) 336–342

n mature platelets [38]. The effects of low-dose aspirin inatients with Alzheimer’s disease were comparable to thosereviously described in other clinical settings characterizedy enhanced platelet activation, such as diabetes mellitus [6],ypercholesterolemia [7], and homozygous homocystinuria10].

The lack of effects of both COX-inhibitors on F2-soprostane excretion supports the hypothesis that enhancedormation of 8-iso-PGF2� in Alzheimer’s disease reflected aechanism of non-enzymatic, oxygen-radical-mediated lipid

eroxidation [28].Deposition of amyloid �-peptide (A�) in senile plaques

nd in the walls of cortical and leptomeningeal blood ves-els is a hallmark of Alzheimer’s disease [42]. One leadingheory is that senile plaques lead to Alzheimer’s diseaseecause of their direct toxicity to adjacent neurons [46]. It haseen postulated that cerebro-vascular amyloid deposits maye derived in part from circulating A� [1]. Indeed, humanlatelets contain high levels of membrane-associated and sol-ble forms of amyloid precursor peptide (APP), which iseleased by platelets, contributing to more than 90% of cir-ulating APP [22]. In vitro, platelets can release APP and A�n response to thrombin, collagen or arachidonic acid [40,41].

Activated platelets that enter into advanced atheroscle-otic plaques through neovessels are phagocitized by macro-hages, with A� release. These A�-positive macrophagesxpress both inducible nitric oxide synthase and COX-, and often exhibit co-localization with platelets [41].hus, platelets might participate in plaque inflammationith a mechanism distinct from the release of inflammatory

ytokines.Furthermore, we investigated whether bioactive prod-

cts of lipid peroxidation and/or inflammatory cytokinesre important determinants of in vivo platelet activation inlzheimer’s disease. Our patients showed increased excre-

ion rates of 8-iso-PGF2�, a non-invasive index of in vivoxidative stress [8], in comparison with age-matched con-rols. Although small amounts of 8-iso-PGF2� can be formedia COX-1 in platelets [35] or COX-2 in monocytes [31],he non-enzymatic mechanism of formation of this F2-soprostane was confirmed by the failure of low-dose aspirinr rofecoxib to modify its urinary excretion to any statis-ically significant extent (Fig. 2). Thromboxane-dependentlatelet activation may represent a functional read-out ofncreased lipid peroxidation, as suggested by the linear corre-ation between excretion rates of the two metabolites. Thus,ioactive iso-eicosanoids may transduce the effects of oxi-ant stress into specialized forms of cellular activation [34],hus amplifying the platelet response to other agonists.

There is a general agreement that markers of increasedxidative damage are present in diseased regions of the brainrom patients with advanced Alzheimer’s disease and in the

erebrospinal fluid of Alzheimer’s disease patients with mildo moderate dementia [36,37,44]. However, the presence ofxidative stress biomarkers in blood and urine of Alzheimer’sisease patients remains controversial [11,27].

iology o

iwspp

aAfMtb8rmtwtE

tdsiapot[tita

apAatc

A

CfoU

R

[

[

[

[

[

[

[

[

[

[

[

[

[

G. Ciabattoni et al. / Neurob

Factors that may contribute to the apparent discrepancynclude the relatively small sample size of these studies, asell as inadequate control for known oxidants (e.g. cigarette

moking and metabolic abnormalities) or antioxidants (e.g.lasma Vitamin E levels), as suggested by the elevated iso-rostane levels in the control group of some of these studies.

At variance with previous studies, we paid particularttention to the presence of cardiovascular risk factors inlzheimer’s disease and controls, because most of these risk

actors are associated with increased lipid peroxidation [8].oreover, we examined whether enhanced lipid peroxida-

ion is related to plasma Vitamin E. The inverse relationetween plasma Vitamin E and urinary excretion rates of-iso-PGF2� in Alzheimer’s disease patients suggests thateduced availability and/or increased consumption of Vita-in E contributes to increased lipid peroxidation in this set-

ing. Moreover, these findings are mechanistically consistentith the results of a clinical trial with Vitamin E supplemen-

ation in Alzheimer’s disease patients, showing that Vitaminis able to slow the progression of the disease [39].A large body of evidence suggests that a chronic inflamma-

ory process is important in the pathogenesis of Alzheimer’sisease, as shown by the presence of activated microglia andeveral proteins associated with immune reactions, includ-ng inflammatory cytokines, such as IL-6 and TNF-�, andcute-phase reactants such as CRP [18]. The inflammatoryrocess seems to be limited to the brain, with limited evidencef systemic involvement, although high serum concentra-ions of acute-phase proteins have sometimes been reported18,23]. Here, we report that circulating levels of CRP andwo cytokines, IL-6 and TNF-�, are not significantly differentn Alzheimer’s disease in comparison with age-matched con-rols. Moreover, their levels were not related to isoprostanend thromboxane biosynthesis in this setting.

We conclude that: (1) a relative deficiency of endogenousntioxidants such as Vitamin E is responsible for increasedroduction of bioactive products of lipid peroxidation inlzheimer’s disease; (2) this results in persistent platelet

ctivation; (3) products of activated platelets may contributeo sustaining amyloid deposits, as well as atherothromboticomplications.

cknowledgments

This research was supported by EC FP6 funding (LSHM-T-2004-0050333). The study was also supported by grants

rom the Italian Ministry of Research and Education (Centerf Excellence on Aging and FIRB) to the “G. d’Annunzio”niversity Foundation.

eferences

[1] Borroni B, Akkawi N, Martini G, Colciaghi F, Prometti P, RozziniL, et al. Microvascular damage and platelet abnormalities in earlyAlzheimer’s disease. J Neurol Sci 2002;204:189–93.

[

f Aging 28 (2007) 336–342 341

[2] Casserly I, Topol E. Convergence of atherosclerosis and Alzheimer’sdisease: inflammation, cholesterol, and misfolded proteins. Lancet2004;363:1139–46.

[3] Catella F, Healy D, Lawson JA, FitzGerald GA. 11-Dehydro-thromboxane B2: a quantitative index of thromboxane A2 formationin the human circulation. Proc Natl Acad Sci USA 1986;83:5861–5.

[4] Ciabattoni G, Pugliese F, Davı G, Pierucci A, Simonetti BM, PatronoC. Fractional conversion of thromboxane B2 to urinary 11-dehydro-thromboxane B2 in man. Biochim Biophys Acta 1989;992:66–70.

[5] Ciabattoni G, Maclouf J, Catella F, FitzGerald GA, Patrono C. Radioim-munoassay of 11-dehydro-thromboxane B2 in human plasma and urine.Biochim Biophys Acta 1987;918:293–7.

[6] Davı G, Catalano I, Averna M, Notarbartolo A, Strano A, Ciabattoni G,et al. Thromboxane biosynthesis and platelet function in type II diabetesmellitus. N Engl J Med 1990;322:1769–74.

[7] Davı G, Averna M, Catalano I, Barbagallo C, Ganci A, NotarbartoloA, et al. Increased thromboxane biosynthesis in type IIa hypercholes-terolemia. Circulation 1992;85:1792–8.

[8] Davı G, Falco A, Patrono C. Determinants of F2-isoprostane biosyn-thesis in man. Chem Phys Lipids 2004;128:149–63.

[9] de la Torre JC. Alzheimer’s disease as a vascular disorder: nosologicalevidence. Stroke 2002;33:1152–62.

10] Di Minno G, Davı G, Margaglione M, Cirillo F, Grandone E, Ciabat-toni G, et al. Abnormally high thromboxane biosynthesis in homozy-gous homocystinuria. Evidence for platelet involvement and probucol-sensitive mechanism. J Clin Invest 1993;92:1400–6.

11] Feillet-Coudray C, Tourtauchaux R, Niculescu M, Rock E, TauveronI, Alexandre-Gouabau MC, et al. Plasma levels of 8-epi-PGF2�, an invivo marker of oxidative stress, are not affected by aging or Alzheimer’sdisease. Free Radic Biol Med 1999;27:463–9.

12] FitzGerald GA, Oates JA, Hawiger J, Maas RL, Roberts II LJ, LawsonJA, et al. Endogenous biosynthesis of prostacyclin and thromboxaneand platelet function during chronic administration of aspirin in man.J Clin Invest 1983;71:676–88.

13] FitzGerald GA, Patrono C. The coxibs, selective inhibitors ofcyclooxygenase-2. N Engl J Med 2001;345:433–42.

14] Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practicalmethod for grading the cognitive state of patients for the clinician. JPsychiatr Res 1975;12:189–98.

15] Hansson GK, Libby P, Schonbeck U, Yan ZQ. Innate and adap-tive immunity in the pathogenesis of atherosclerosis. Circ Res2002;91:281–91.

16] Honig LS, Tang MX, Albert S, Costa R, Luchsinger J, ManlyJ, et al. Stroke and the risk of Alzheimer disease. Arch Neurol2003;60:1707–12.

17] Huo Y, Ley KF. Role of platelets in the development of atherosclerosis.Trends Cardiovasc Med 2004;14:18–22.

18] Jones RW. Inflammation and Alzheimer’s disease. Lancet 2001;358:455–61.

19] Lawson JA, Patrono C, Ciabattoni G, Fitzgerald GA. Long-lived enzy-matic metabolites of thromboxane B2 in the human circulation. AnalBiochem 1986;155:198–205.

20] Ledue TB, Weiner DL, Sipe JD, Poulin SE, Collins MF, Rifai N. Ana-lytical evaluation of particle-enhanced immunonephelometric assaysfor C-reactive protein, serum amyloid A and mannose-binding proteinin human serum. Ann Clin Biochem 1998;35:745–53.

21] Lee BL, Chua SC, Ong HY, Ong CN. High-performance liquid chro-matographic method for routine determination of Vitamins A and E andbeta-carotene in plasma. J Chromatogr 1992;581:41–7.

22] Li QX, Berndt MC, Bush AI, Rumble B, Mackenzie I, FriedhuberA, et al. Membrane-associated forms of the beta A4 amyloid proteinprecursor of Alzheimer’s disease in human platelet and brain: surface

expression on the activated human platelet. Blood 1994;84:133–42.

23] Lombardi VR, Garcia M, Rey L, Cacabelos R. Characterization ofcytokine production, screening of lymphocyte subset patterns and invitro apoptosis in healthy and Alzheimer’s disease (AD) individuals. JNeuroimmunol 1999;97:163–71.

3 iology o

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

42 G. Ciabattoni et al. / Neurob

24] McAdam BF, Byrne D, Morrow JD, Oates JA. Contribution ofcyclooxygenase-2 to elevated biosynthesis of thromboxane A2 andprostacyclin in cigarette smokers. Circulation 2005;112:1024–9.

25] McKhann G, Drachman D, Folstein M, Katzman R, Price D, StadlanEM. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of the Department of Healthand Human Services Task Force on Alzheimer’s Disease. Neurology1984;34:939–44.

26] Montine KS, Quinn JF, Zhang J, Fessel JP, Roberts II LJ, MorrowJD, et al. Isoprostanes and related products of lipid peroxidation inneurodegenerative diseases. Chem Phys Lipids 2004;128:117–24.

27] Montine TJ, Quinn JF, Milatovic D, Silbert LC, Dang T, Sanchez S, etal. Peripheral F2-isoprostanes and F4-neuroprostanes are not increasedin Alzheimer’s disease. Ann Neurol 2002;52:175–9.

28] Morrow JD, Hill KE, Burk RF, Nammour TM, Badr KF, Roberts II LJ.A series of prostaglandin F2-like compounds are produced in vivo inhumans by a non-cyclooxygenase, free radical-catalyzed mechanism.Proc Natl Acad Sci USA 1990;87:9383–7.

29] Orlandi M, Bartolini G, Belletti B, Spisni E, Tomasi V. Thrombox-ane A2 synthase activity in platelet free human monocytes. BiochimBiophys Acta 1994;1215:285–90.

30] Patrignani P, Filabozzi P, Patrono C. Selective cumulative inhibition ofplatelet thromboxane production by low-dose aspirin in healthy sub-jects. J Clin Invest 1982;69:1366–72.

31] Patrignani P, Santini G, Panara MR, Sciulli MG, Greco A, RotondoMT, et al. Induction of prostaglandin endoperoxide synthase-2 in humanmonocytes associated with cyclo-oxygenase-dependent F2-isoprostaneformation. Br J Pharmacol 1996;118:1285–93.

32] Patrono C, Davı G, Ciabattoni G. Thromboxane biosynthesis andmetabolism in relation to cardiovascular risk factors. Trends Cardio-vasc Med 1992;2:15–20.

33] Patrono C, Garcia Rodriguez LA, Landolfi R, Baigent C. Low-dose aspirin for the prevention of atherothrombosis. N Engl J Med2005;353:2373–83.

34] Patrono C, FitzGerald GA. Isoprostanes: potential markers of oxi-

dant stress in atherothrombotic disease. Arterioscler Thromb Vasc Biol1997;17:2309–15.

35] Pratico D, Lawson JA, FitzGerald GA. Cyclooxygenase-dependentformation of the isoprostane, 8-epi prostaglandin F2�. J Biol Chem1995;270:9800–8.

[

f Aging 28 (2007) 336–342

36] Pratico D, Lee VM, Trojanowski JQ, Rokach J, FitzGerald GA.Increased F2-isoprostanes in Alzheimer’s disease: evidence forenhanced lipid peroxidation. FASEB J 1998;12:1777–84.

37] Pratico D, Clark CM, Lee VM, Trojanowski JQ, Rokach J, FitzGeraldGA. Increased 8, 12-iso-iPF2alpha-VI in Alzheimer’s disease: correla-tion of a noninvasive index of lipid peroxidation with disease severity.Ann Neurol 2000;48:809–12.

38] Rocca B, Secchiero P, Ciabattoni G, Ranelletti FO, Catani L, Guidotti L,et al. Cyclooxygenase-2 expression is induced during human megakary-opoiesis and characterizes newly formed platelets. Proc Natl Acad SciUSA 2002;99:7634–9.

39] Sano M, Ernesto C, Thomas RG, Klauber MR, Schafer K, GrundmanM, et al. A controlled trial of selegiline, alpha-tocopherol, or both astreatment for Alzheimer’s disease. The Alzheimer’s Disease Coopera-tive Study. N Engl J Med 1997;336:1216–22.

40] Skovronsky DM, Lee VM, Pratico D. Amyloid precursor protein andamyloid beta peptide in human platelets. Role of cyclooxygenase andprotein kinase C. J Biol Chem 2001;276:17036–43.

41] Tedgui A, Mallat Z. Platelets in atherosclerosis: a new rolefor beta-amyloid peptide beyond Alzheimer’s disease. Circ Res2002;90:1145–6.

42] Thomas T, Thomas G, McLendon C, Sutton T, Mullan M. �-Amyloid-mediated vasoactivity and vascular endothelial damage.Nature 1996;380:168–71.

43] van Kooten F, Ciabattoni G, Koudstaal PJ, Grobbee DE, Kluft C,Patrono C. Increased thromboxane biosynthesis is associated with post-stroke dementia. Stroke 1999;30:1542–7.

44] Waddington E, Croft K, Clarnette R, Mori T, Martins R. Plasma F2-isoprostanes levels are increased in Alzheimer’s disease: evidence ofincreased oxidative stress in vivo. Alzheimers Rep 1999;2:277–82.

45] Wang Z, Ciabattoni G, Creminon C, Lawson J, Fitzgerald GA, PatronoC, et al. Immunological characterization of urinary 8-epi-PGF2 excre-tion in man. J Pharmacol Exp Ther 1995;275:94–100.

46] Yamaguchi H, Yamazaki T, Lemere CA, Frosch MP, Selkoe DJ. Betaamyloid is focally deposited within the outer basement membrane in

the amyloid angiopathy of Alzheimer’s disease. An immunoelectronmicroscopic study. Am J Pathol 1992;141:249–59.

47] Zubenko GS, Winwood E, Jacobs B, Teply I, Stiffler JS, Hughes III HB,et al. Prospective study of risk factors for Alzheimer’s disease: resultsat 7.5 years. Am J Psychiat 1999;156:50–7.