Pro-inflammatory effect of freshly solubilized β-amyloid peptides in the brain

Prostaglandins & other Lipid Mediators 70 (2002) 1–12

Pro-inflammatory effect of freshly solubilized�-amyloid peptides in the brain

Daniel Paris∗, Kirk P. Townsend, Demian F. Obregon,James Humphrey, Michael Mullan

The Roskamp Institute, University of South Florida, 3515 E. Fletcher Avenue, Tampa, FL 33613, USA

Received 10 August 2001; received in revised form 21 August 2001; accepted 22 October 2001

Abstract

It has recently been shown that the level of soluble�-amyloid (A�) peptides correlates well withthe severity of synaptic loss and the density of neurofibrillary tangles observed in Alzheimer’sdisease (AD) brain. However, the biological activity of soluble forms of A� peptides in the brainremains to be determined. We have investigated ex vivo the effect of freshly solubilized A�1–40

peptides (fsA�) on prostaglandin E2 (PGE2) production in rat brain slices. PGE2 levels increasedrapidly following treatment with fsA�, an effect that was prevented by SB202190, a selective in-hibitor of p38 mitogen-activated protein kinase (p38 MAPK), and by NS-398, which preferentiallyinhibits cyclooxygenase-2 (COX-2) compared to COX-1. In an attempt to determine the cellularsystems of the brain responsible for prostaglandin production in response to fsA�, the effect of fsA�was tested on isolated brain microvessels, primary cultures of brain smooth muscle cells/pericytesand endothelial cells, and a human neuron-like cell line (IMR32). Our data show that fsA� ex vivocan stimulate prostaglandin accumulation in incubates of isolated rat brain microvessels. In addi-tion, fsA� appears to cause a concentration-dependent enhancement of prostaglandin accumulationin primary cultures of brain microvessel-derived smooth muscle cells/pericytes but not of brainendothelial cells. Finally, fsA� also stimulated PGF2� accumulation in cultures of differentiatedIMR32 cells, but to a lesser extent than in brain smooth muscle cell/pericyte cultures. Depositionof aggregated forms of A� in the brain has been thought to trigger an inflammatory response whichaccompanies the neuropathologic events of AD. Our data provide evidence that fsA� triggers apro-inflammatory reaction in rat brain, and suggest that the cerebrovasculature may constitute animportant source of pro-inflammatory eicosanoids.© 2002 Elsevier Science Inc. All rights reserved.

Keywords: Alzheimer’s disease; Inflammation; p38 MAPK; Cyclooxygenase; Prostaglandin; Brain slices;Microvessels

∗ Corresponding author. Tel.:+1-813-974-3722; fax:+1-813-974-3915.E-mail address: [email protected] (D. Paris).

0090-6980/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved.PII: S0090-6980(02)00111-9

2 D. Paris et al. / Prostaglandins & other Lipid Mediators 70 (2002) 1–12

1. Introduction

In Western countries Alzheimer’s disease (AD) is the major cause of dementia in theelderly. It is characterized by the progressive accumulation of intracellular neurofibril-lary tangles, extracellular parenchymal senile plaques, and congophilic cerebrovasculardeposits[1]. The principal components of senile plaques and cerebrovascular deposits arethe�-amyloid peptides (A�). It has become increasingly evident that A� peptides may playan important role in mediating the initial pathogenic events in AD, since all the mutationsassociated with familial forms of AD affect A� production[2] and since A� levels areincreased in AD brain before the occurrence of significant tau pathology[3]. Furthermore,the quantity of soluble A� peptides is increased in AD brain and correlates well with theamount of synaptic loss and the density of neurofibrillary tangles observed in AD, whereasthe amount of aggregated A� does not[4,5]. These data suggest that both soluble andaggregated forms of A� peptides might exert a deleterious action in AD.

Release of arachidonic acid (the precursor of the 2-series prostaglandins) from mem-brane phospholipids has been associated with different acute cerebral pathologies that con-stitute risk factors for AD including transitory ischemia, convulsions, and head trauma[6]. Increased levels of PGE2 have been reported in AD cerebrospinal fluid[7,8], and im-munoreactivity for cyclooxygenase-2 (COX-2), a key enzyme in prostanoid biosynthesis, iselevated in AD brain[9]. Prostaglandins in the brain are a key element in the induction andmaintenance of a chronic inflammatory state[10] which may exacerbate AD pathology. Wehave previously shown that freshly solubilized A� peptides (fsA�) induce vasoconstric-tion in rat isolated aorta by stimulating a pro-inflammatory pathway involving activationof p38 MAPK, COX-2, and the formation of pro-inflammatory eicosanoids[11]. We thusinvestigated whether fsA� might stimulate a similar pro-inflammatory pathway in the brain.

2. Methods

2.1. Rat brain slice cultures

Male 9-month-old Sprague–Dawley rats (Zivic Miller, Zelienople, PA) were anesthetizedwith isofluorane, and perfused intracardially with ice-cold PBS (pH 7.4). The brains were re-moved aseptically and sectioned coronally into eight 2-mm slices using a customized rodentbrain matrix (rat brain slicer: Muromachi Kikai, Tokyo, Japan). Slices were washed severaltimes in cold PBS (pH 7.4) and maintained in 7 ml of artificial cerebrospinal fluid (NaCl,129 mM; KCl, 3 mM; NaHPO4, 1.25 mM; MgSO4, 1.8 mM; CaCl2, 1.6 mM; NaHCO3,21 mM; glucose, 10 mM; pH 7.4) at 37◦C and continuously gassed with 5% CO2 in O2.Some sections were treated with 1�M of fsA�1–40 (Biosource, Inc., CA); control sectionswere untreated, and other sections were treated with NS-398 (20�M) or SB202190 (5�M)alone or in combination with fsA�1–40. Prostaglandin E2 and F2� levels were determinedin 100�l aliquots of Krebs’ buffer surrounding the brain sections using the respectivecompetitive enzyme immunoassay kits (Cayman Chemical Company, Ann Arbor, MI).Aliquots were collected 30, 60 and 120 min after treatment with fsA�. Data are expressedas means± S.E.M. picogram of prostaglandin/milliliter per gram of fresh tissue.

D. Paris et al. / Prostaglandins & other Lipid Mediators 70 (2002) 1–12 3

2.2. Rat and bovine brain microvessels isolation

Brain microvessels were isolated as described previously[12]. Rat brains (n = 3) or1 cm3 pieces of bovine cerebral cortex were homogenized with a dounce homogenizerin 20–25 ml of Hank’s balanced salt solution (HBSS) at 4◦C. Briefly, the homogenateswere centrifuged at 2000× g and the pellets were resuspended in 4 volumes of HBSScontaining 15% dextran and 5% fetal calf serum (FCS). Following 20-min centrifugationat 4000× g, the pellets were resuspended in 2 volumes of HBSS. Microvessels werethen collected on 53-�m filters and washed with 30 ml of HBSS. Microvessels were thencentrifuged at 2000×g for 15 min and resuspended in Dulbecco’s Modified Eagle Medium(DMEM, Life Technologies, Inc., Rockville, MD) containing 5% FCS. Microvessels werethen transferred to 24-well tissue culture plates and incubated at 37◦C following treatmentwith 2�M of fsA�1–40 (control microvessels were untreated). A total of 100�l of cellculture medium were collected at 2, 4 and 6 h of incubation and assayed for PGE2. Atthe end of the experiments, microvessels were collected by centrifugation and sonicated inmammalian lysis buffer (Pierce, Rockford, IL). Protein concentrations were determined inthe lysis buffer using Biorad reagent, and results were standardized against the amount ofprotein.

2.3. Primary cultures of smooth muscle cells/pericytes from rat and bovinebrain microvessels

Primary cultures of smooth muscle cells/pericytes and endothelial cells, from rat andbovine brain microvessels, were established essentially as previously described[12]. Briefly,for the preparation of smooth muscle cells/pericytes, brain microvessels were treated with1 mg/ml collagenase/dispase (Life Technologies, Inc.) for 2 h at 37◦C and transferred inDMEM containing 10% FCS to 6-well tissue culture plates. For the isolation of endothelialcells, microvessels were included in Matrigel (Becton Dickinson, Bedford, MA) and cov-ered with EBM-2 medium (Clonetics Corporation, Walkersville, MD). Pieces of Matrigelcontaining proliferating endothelial cells were collected and maintained in EBM medium(Clonetics Corporation) supplemented with 4% FCS. Smooth muscle cell/pericyte culturesand endothelial cultures were further characterized by double immunofluorescence usingtwo antibodies, one against factor VIII, and another against�-smooth muscle cell actin(data not shown). Smooth muscle cell/pericyte cultures were positive for�-smooth musclecell actin and negative for factor VIII. Endothelial cells were positive for factor VIII andnegative for�-smooth muscle cell actin. Confluent smooth muscle cell/pericyte culturesfrom rat brain were then treated with 0.5, 1 or 2�M fsA�, 20�M NS-398, a combinationof fsA� (2�M) and NS-398, 5�M SB202190, a combination of SB202190 and fsA�,100�M of valeryl salicylate (a high concentration), or a combination of fsA� and valerylsalicylate. A total of 50�l of medium were collected for PGE2 and PGF2� determinationfollowing 2, 4 and 6 h incubation. Confluent bovine brain endothelial cells were treated(controls untreated) with 2�M fsA� for 2, 4 and 6 h, and cell culture medium was analyzedfor PGE2 and PGF2� using competitive ELISA following the recommendation of the manu-facturer (Cayman Chemical Company). Results are expressed as picogram of prostaglandinper milligram of cellular protein.


2.4. IMR32 cells culture

The human neuronal cell line IMR32 (purchased from American Type Culture Collec-tion, VA) was cultivated in DMEM/Ham’s F12 medium supplemented with 10%fetal bovine serum (FBS), penicillin (50 U/ml), streptomycin (50�g/ml), and 2 mM glu-tamine (Life Technologies, Inc.). When the IMR32 cells were 70–80% confluent, theywere differentiated for a week in the medium described above supplemented with 10�Mof 5-bromo-2′-deoxyuridine (Sigma, St. Louis, MO). The medium was changed every 3days and immediately prior to treatment with fsA�. Differentiated IMR32 cells were thentreated with 0.5 or 1�M fsA� or were untreated (controls). Then, 100�l of cell culturemedium were collected following 2, 4 and 6 h treatment with fsA� and assayed for PGF2�

as above.

Fig. 1. (A) Effect of a COX-2 inhibitor on PGE2 levels in incubates of rat brain sections following treatmentwith 1�M of freshly solubilized A�1–40. The number of sections studied for each treatment is indicated inparentheses in the figure key. ANOVA revealed main effects of time (P < 0.01), A� (P < 0.04), and NS-398(P < 0.001). Post hoc comparisons revealed differences between control and fsA� (P < 0.005), and betweencontrol and NS-398 (P < 0.05), but no significant difference between NS-398 and NS-398+ fsA� (P = 0.999).Thus, NS-398 completely inhibited PGE2 accumulation stimulated by fsA�. (B) Effect of a p38 MAPK inhibitoron PGE2 accumulation in incubates of rat brain slices. ANOVA revealed effects of time (P < 0.005), fsA�

(P < 0.001), and SB202190 (P < 0.001). Post hoc comparisons revealed differences between control and fsA�

(P < 0.005), but no significant difference between control and SB202190 (P = 0.295) or between SB202190and SB202190+ fsA� (P = 0.456). Thus SB202190 completely inhibited the increase of PGE2 accumulationstimulated by fsA�.


2.5. Statistical analysis

Comparisons among groups were evaluated using ANOVA, and post hoc comparisonswere made using Scheffe’s or Bonferroni’s method where appropriate. Levene’s test forequality of variance, followed by at-test for independent samples, were used for singlemean comparisons. The alpha level was 0.05, and analyses were performed using SPSS forWindows release 10.0.

3. Results

3.1. Freshly solubilized Aβ stimulated eicosanoid production by rat brain slices

Our data show that rat brain slices produce PGE2 in a time-dependent manner, andthat fsA� enhanced the accumulation of PGE2 in the incubation fluid (Fig. 1A). NS-398,which preferentially inhibits COX-2 compared to COX-1[13–15], reduced the spontaneousrelease of PGE2 from rat brain slices, suggesting that an important fraction of PGE2 pro-duced by the brain is COX-2 derived (Fig. 1A). Because the increased PGE2 accumulationinduced by fsA� was completely inhibited by NS-398, COX-2 probably contributes to thepro-inflammatory effect of fsA� in rat brain.

Since p38 MAPK regulates COX-2 induction[16], we tested the possible involvement ofthis kinase in the stimulation of prostaglandin accumulation triggered by fsA�. SB202190,a highly selective inhibitor of p38 MAPK[17] compared with other kinases, did not signifi-cantly affect spontaneous PGE2 production by rat brain slices (Fig. 1B). However, 5�M ofSB202190 (at this concentration SB202190 selectively inhibits p38 MAPK activity with-out affecting ERK2, SAPK3 or SAPK4 activities[18,19]) prevented the increased PGE2accumulation induced by fsA� (Fig. 1B). Similar results were observed with PGF2� (datanot shown).

Fig. 2. The fsA� stimulated PGE2 accumulation in incubates of isolated rat brain microvessels ex vivo.P < 0.01for the comparison of control rat brain microvessels and microvessels challenged with fsA� for 2 h (t-test forindependent samples).


Fig. 3. (A) Concentration-dependent effect of fsA� on PGE2 production by rat brain smooth muscle cells(SMC)/pericytes. ANOVA revealed effects of fsA� (P < 0.001) and fsA� concentration (P < 0.001). One-wayANOVA across the different time-points revealed between-groups differences (P < 0.001), and post hoc com-parisons revealed differences between (a) control and 500 nM fsA� (P < 0.001), (b) control and 1�M fsA�

(P < 0.001), (c) control and 2�M fsA� (P < 0.001), (d) 500 nM fsA� and 1�M fsA� (P < 0.001), and (e)1�M fsA� and 2�M fsA� (P < 0.001). Similar results were obtained for PGF2� (data not shown). These resultsshow that fsA� stimulated the accumulation of PGE2 and PGF2� in primary cultures of rat SMC/pericytes isolatedfrom brain microvessels in a concentration-dependent manner. (B) COX-1 does not appear to contribute to theproduction of PGE2 stimulated by fsA� in primary cultures of rat brain SMC/pericytes, since ANOVA revealedan effect of fsA� (P < 0.001) and time (P < 0.001) but no significant effect of valeryl salicylate (labeled asvaleryl) (P = 0.227). One-way ANOVA across the different time-points revealed between-groups differences(P < 0.001), and post hoc comparisons revealed differences between (a) control and fsA� (P < 0.001), and(b) control and valeryl salicylate+ fsA� (P < 0.001) but no significant differences between fsA� and valerylsalicylate+ fsA� (P = 0.997) or between control and valeryl salicylate (P = 0.269). These data indicatethat COX-1 probably does not mediate the release of PGE2 induced by fsA� in primary cultures of rat brainSMC/pericytes.


Fig. 4. (A) COX-2 appears to mediate the release of PGE2 induced by fsA� in primary cultures of rat brainSMC/pericytes. ANOVA revealed effects of fsA� (P < 0.001), NS-398 (P < 0.001) and time (P < 0.02),and interaction between NS-398, fsA� and time (P < 0.007). One-way ANOVA across the different time-pointsrevealed between-groups differences (P < 0.001), and post hoc comparisons revealed differences between controland fsA� (P < 0.001), and between control and NS-398 (P < 0.001), but no significant differences betweenNS-398 and NS-398+ fsA� (P = 0.941). These data show that the preferential COX-2 inhibitor NS-398 totallyinhibited the accumulation of PGE2 stimulated by fsA� in primary cultures of rat brain SMC/pericytes. Similarresults were obtained for PGF2� (data not shown). (B) The p38 MAPK inhibitor SB202190 prevented the increase ofPGE2 accumulation induced by fsA� in primary cultures of bovine brain SMC/pericytes. ANOVA revealed effectsof fsA� (P < 0.001), SB202190 (P < 0.001) and time (P < 0.001), as well as a interaction between SB202190and fsA� (P < 0.001). One-way ANOVA across the different time-points revealed between-groups differences(P < 0.001), and post hoc comparisons revealed differences between control and fsA� (P < 0.001), and betweencontrol and SB202190 (P < 0.007), but no significant difference between SB202190 and SB202190+ fsA�

(P = 0.926). These data show that the highly selective p38 MAPK inhibitor SB202190 completely inhibitedPGE2 accumulation stimulated by fsA� in primary cultures of cow brain SMC/pericytes. Similar results wereobtained for PGF2� (data not shown).


3.2. Contribution of the cerebrovasculature to the eicosanoid production induced byfreshly solubilized Aβ

We previously showed that fsA� peptides stimulate eicosanoid production in the pe-ripheral vasculature, resulting in an increased vasoconstriction by endothelin-1[11]. In thepresent study, we wondered whether fsA� could also stimulate prostaglandin formation inthe cerebrovasculature, and thus we incubated rat and bovine brain microvessels with fsA�.Our data show that fsA� increased the accumulation of PGE2 in both tissues (rat,Fig. 2;bovine, data not shown). The fsA� also enhanced the accumulation of PGF2� and PGE2in primary cultures of both rat and bovine cerebrovascular smooth muscle cells/pericytes(Figs. 3 and 4) whereas the cerebrovascular-derived endothelial cells were unresponsive(data not shown). Furthermore, prostaglandin accumulation stimulated by fsA� was inhib-ited by 5�M of the p38 MAPK inhibitor SB202190 (Fig. 4B). The preferential COX-2inhibitor NS-398 blocked the accumulation of prostaglandin triggered by fsA� in primarycultures of brain microvessel smooth muscle cells/pericytes whereas the preferential COX-1inhibitor valeryl salycilate[20] was ineffective (Figs. 3B and 4A).

3.3. Effect of freshly solubilized Aβ on PGF2� production by differentiatedIMR32 cells

We observed that fsA� stimulated the accumulation of PGF2� in cultures of differentiatedIMR32, suggesting that neurons can be a source of pro-inflammatory eicosanoids in thebrain (Fig. 5). We then investigated the effect of fsA� on a human astrocytoma cell lineand observed that fsA� did not significantly affect their production of PGF2� (data notshown).

Fig. 5. fsA� stimulated the formation of PGF2� in differentiated IMR32 cells. ANOVA revealed effects of fsA�

concentration (P < 0.001) and time (P < 0.001). One-way ANOVA across the different time-points revealedbetween-groups differences (P < 0.001), and post hoc comparisons revealed differences between control and1�M fsA� (P < 0.001), and between 500 nM and 1�M fsA� (P < 0.002).


4. Discussion

Some non-steroidal anti-inflammatory drugs (NSAIDs) for the treatment of arthriticdiseases can slow the progression of AD or reduce the risk of its development[21,22],suggesting that inflammatory events may be important in the initiation of AD pathology.In particular, NSAIDs reduce prostaglandin production by inhibiting cyclooxygenase[23],suggesting that prostaglandins may contribute to AD development. Several converging linesof evidence suggest that A� peptides and inflammation may be linked in AD pathogenesis.For example, senile plaques observed in AD brain are sites of classical inflammatory pro-cesses, as shown by the presence of numerous degenerating neurons, reactive microglia andastrocytes, cytokines, and complement proteins[24,25]. However, measures of aggregatedA� are not good predictors of AD dementia severity, whereas soluble A� levels correlatewell with the synaptic decline observed in AD brain. Thus soluble forms of A� may play animportant role in AD pathogenesis[4,5]. Although it is generally regarded that aggregatedand fibrillar A� deposits in amyloid plaques can trigger inflammation, soluble forms of thepeptide have not been specifically investigated in relation to inflammation. For this reason,we tested the effect of fsA� on prostaglandin production in rat brain slices. At physiologicpH, A�1–40 forms some soluble oligomeric structures including dimers and tetramers, whichappear to be stable for an extended period (4 weeks)[26]. Under our experimental paradigm,A�1–40 is dissolved at concentrations of 0.5, 1 or 2�M in Krebs’ buffer or cell culture mediaand incubated for 6 h or less, corresponding to a time frame and pH which do not favor A�aggregation. In our preparations, fsA� probably exists as dimeric and tetrameric structures.Our data reveal that fsA� can stimulate the accumulation of PGE2 and PGF2� in culturesof rat brain slices, suggesting that soluble forms of A� may exhibit pro-inflammatory ef-fects in the brain. In addition, this increased production of PGE2 and PGF2� in rat brainslices following fsA� treatment can be opposed by SB202190 or NS-398, which prefer-entially inhibit p38 MAPK and COX-2 respectively. To determine the cell types likely tobe responsible for the increased prostaglandin production, we studied the effect of fsA�on isolated brain microvessels and human neuron-like cells (IMR32). The fsA� enhancedPGE2 accumulation in cultures of brain microvessels, and this was inhibited by SB202190or NS-398, suggesting the involvement of p38 MAPK/COX-2-dependent pathway(s). Inter-estingly, the pro-inflammatory effect of fsA� in these brain capillaries is essentially drivenby smooth muscle cells/pericytes but not by endothelial cells. The preferential COX-1 in-hibitor valeryl salicylate[20] did not affect the accumulation of prostaglandins triggered byfsA� in cultures of brain smooth muscle/pericytes, whereas NS-398, which preferentiallyblocks COX-2 compared to COX-1, prevented the increased prostaglandin accumulationinduced by fsA�.

Since neurons constitutively express COX-2[13], we wondered whether they also re-spond to fsA� by producing more prostaglandins. We tested this using differentiated hu-man neuroblastoma IMR32 cells, which can form neurofibrillary tangles in vitro[27],express functional nicotinic receptors[28], and are human neuron-like cells. They pro-duced PGF2� in response to fsA�, suggesting that neurons may also be an important sourceof brain pro-inflammatory eicosanoids. In addition to the cerebrovasculature and neuronsperhaps other cell types also contribute to the pro-inflammatory effect of soluble A�. Forinstance, we showed previously that fsA� enhances eicosanoid production by microglia


[29], and others have shown that A� stimulates prostaglandin production by astrocytes[10].

The increase of prostaglandin production with fsA� in the brain seems to involve stim-ulation of a p38 MAPK/COX-2-dependent pathway. However, at the concentration used(20�M), NS-398 may have some effect on COX-1 activity[15], and it can partially reduceCOX-1 activity in primary cultures of COX-2-deficient endothelial cells (data not shown).Thus COX-1 might also contribute to the accumulation of PGE2 triggered by fsA� in ratbrain slice cultures. The baseline spontaneous production of prostaglandins by rat brainslices was unaffected by the p38 MAPK inhibitor SB202190, suggesting that basal COXactivity is regulated via another mechanism. NF-�B for example might play an importantrole in controlling the basal COX-2 level[30]. We previously found that before the appear-ance of aggregated deposits of A� peptides, p38 MAPK activity was increased in the brainsof transgenic mice which overexpressed a human APP transgene bearing the “Swedish”mutation. Soluble forms of A� may therefore be able to activate p38 MAPK in vivo[11].In addition, we found that p38 MAPK also mediates the vasoactive properties of fsA�[11]. The stress-activated enzyme p38 MAPK (responsible for transducing inflammatorysignals and initiating apoptosis) is activated in neurofibrillary tangle-bearing neurons ofAD brains[31]. Our data suggest that p38 MAPK activity is required for the stimulation ofprostaglandin production induced by soluble forms of A� in the brain and could contributeto AD neuroinflammation.

An elevated PGE2 concentration in the cerebrospinal fluid has been found in patientswith probable AD[7,8], and COX-2 immunoreactivity was elevated in hippocampal neu-rons from AD brains[9]. The role of elevated prostaglandin production in AD pathogenesisis far from understood. Prostaglandins are important modulators of arterial vasotonus; forexample PGF2� can reduce cerebral blood flow by inducing vasoconstriction in the cere-brovasculature[32,33]. Cerebral blood flow is depressed in AD brain, and the Mini-MentalStatus Examination score (a sensitive marker of neuropsychological deficit) correlates withthe hippocampal perfusion level[34]. Perhaps induced vasoconstriction contributes to ADpathophysiology. Prostaglandins might also contribute to AD pathophysiology via othersmechanisms; for instance, PGE2 can activate the expression of the amyloid precursor pro-tein in astrocytes[35], and thus may play an important role in modulating A� production.Activated microglia probably contribute to the neurodegeneration observed in AD, and weshowed previously that an inhibitor of COX-2 or p38 MAPK can suppress release of tumornecrosis factor-� from activated microglia[36]. Chronic neuroinflammation in rats leadsto the destruction of cholinergic neurons and the activation of microglia and astrocytesvia a COX-2-dependent mechanism, suggesting that prostaglandins could contribute to thedegeneration of forebrain cholinergic neurons in AD[37,38]. Interestingly, PGE2 can stim-ulate the release of glutamate by astrocytes[39], which can activate glutamate receptorsand the entry of toxic amounts of calcium into neurons.

NSAIDs can reduce the risk of developing AD, apparently through inhibition of COX ac-tivity and reduction of prostaglandin formation[22]. Chronic oral administration of ibupro-fen, a commonly used NSAID, reduced microglial activation, astrogliosis and amyloidpathology in a transgenic model of AD[40], suggesting that prostaglandins may be im-portant in the development of AD pathology. Our data suggest that fsA� can induce apro-inflammatory reaction in the brain resulting in increased prostaglandin production,


which is in part mediated by the neurons and the cerebrovasculature. The levels of solubleA� might be elevated during the AD process before the appearance of any aggregated formsof A�, and this soluble A� might contribute to the AD process by stimulating inflammation.Although additional studies are required to determine the relative contributions of multipleneuroinflammatory phenomena occurring in AD brains, prostaglandins could be importantin the exacerbation of neurodegeneration in AD.

Acknowledgements

We wish to thank Bob and Diane Roskamp for their generous support, which helped tomake this work possible. This work was supported in part by National Institutes of HealthGrant AG19250-01 and by a Veterans’ Administration Merit Award.

References

[1] Sisodia SS, Price DL. Role of the beta-amyloid protein in Alzheimer’s disease. FASEB J 1995;9:66–70.[2] Selkoe DJ. Amyloid beta-protein and the genetics of Alzheimer’s disease. J Biol Chem 1996;271:8295–8.[3] Naslund J, Haroutunian V, Mohs R, et al. Correlation between elevated levels of amyloid beta-peptide in the

brain and cognitive decline. J Am Med Assoc 2000;283:571–7.[4] Lue LF, Kuo YM, Roher AE, et al. Soluble amyloid beta peptide concentration as a predictor of synaptic

change in Alzheimer’s disease. Am J Pathol 1999;155:853–62.[5] McLean CA, Cherny RA, Fraser FW, et al. Soluble pool of A� amyloid as a determinant of severity of

neurodegeneration in Alzheimer’s disease. Ann Neurol 1999;46:860–6.[6] Bazan NG, Rodriguez de Turco EB, Allan G. Mediators of injury in neurotrauma: intracellular signal

transduction and gene expression. J Neurotrauma 1995;12:791–814.[7] Ho L, Luterman JD, Aisen PS, Pasinetti GM, Montine TJ, Morrow JD. Elevated CSF prostaglandin E2 levels

in patients with probable AD. Neurology 2000;55:323.[8] Montine TJ, Sidell KR, Crews BC, et al. Elevated CSF prostaglandin E2 levels in patients with probable AD.

Neurology 1999;53:1495–8.[9] Kitamura Y, Shimohama S, Koike H, et al. Increased expression of cyclooxygenases and peroxisome

proliferator-activated receptor-gamma in Alzheimer’s disease brains. Biochem Biophys Res Commun1999;254:582–6.

[10] Landolfi C, Soldo L, Polenzani L, et al. Inflammatory molecule release by beta-amyloid-treated T98Gastrocytoma cells: role of prostaglandins and modulation by paracetamol. Eur J Pharmacol 1998;360:55–64.

[11] Paris D, Town T, Mori T, Parker TA, Humphrey J, Mullan M. Soluble�-amyloid peptides mediate vasoactivityvia activation of a pro-inflammatory pathway. Neurobiol Aging 2000;21:183–97.

[12] Diglio CA, Liu W, Grammas P, Giacomelli F, Wiener J. Isolation and characterization of cerebral resistancevessel endothelium in culture. Tissue Cell 1999;25:833–6.

[13] Kaufmann WE, Worley PF, Pegg J, Bremer M, Isakson P. COX-2, a synaptically induced enzyme, is expressedby excitatory neurons at postsynaptic sites in rat cerebral cortex. Proc Natl Acad Sci USA 1996;93:2317–21.

[14] Futaki N, Takahashi S, Yokoyama M, Arai I, Higuchi S, Otomo S. NS-398, a new anti-inflammatory agent,selectively inhibits prostaglandin G/H synthase/cyclooxygenase (COX-2) activity in vitro. Prostaglandins1994;47:55–9.

[15] Warner TD, Giuliano F, Vojnovic I, Bukasa A, Mitchell JA, Vane JR. Nonsteroid drug selectivities forcyclo-oxygenase-1 rather than cyclo-oxygenase-2 are associated with human gastrointestinal toxicity: a fullin vitro analysis. Proc Natl Acad Sci USA 1999;96:7563–8.

[16] Ridley SH, Dean JL, Sarsfield SJ, Brook M, Clark AR, Saklatvala J. A p38 MAP kinase inhibitor regulatesstability of interleukin-1-induced cyclooxygenase-2 mRNA. FEBS Lett 1998;439:75–80.


[17] Manthey CL, Wang SW, Kinney SD, Yao Z. SB202190, a selective inhibitor of p38 mitogen-activated proteinkinase, is a powerful regulator of LPS-induced mRNAs in monocytes. J Leukoc Biol 1998;64:409–17.

[18] Li Z, Jiang Y, Ulevitch RJ, Han J. The primary structure of p38 gamma: a new member of p38 group of MAPkinases. Biochem Biophys Res Commun 1996;228:334–40.

[19] Kumar S, McDonnell PC, Gum RJ, Hand AT, Lee JC, Young PR. Novel homologues of CSBP/p38 MAPkinase: activation substrate specificity and sensitivity to inhibition by pyridinyl imidazoles. Biochem BiophysRes Commun 1997;235:533–8.

[20] Bhattacharyya DK, Lecomte M, Dunn J, Morgans DJ, Smith WL. Selective inhibition of prostaglandinendoperoxide synthase-1 (cyclooxygenase-1) by valeryl salicylic acid. Arch Biochem Biophys 1995;317:19–24.

[21] Andersen K, Launer LJ, Ott A, Hose AW, Breteler MMB, Hofman A. Do non-steroidal anti-inflammatorydrugs decrease the risk for Alzheimer’s disease? The Rotteerdam study. Neurology 1995;45:1441–5.

[22] Breitner JCS, Gau BA, Welsh KA, Plassman BL, McDonald WM, Helms MJ, et al. Inverse association ofanti-inflammatory treatments and Alzheimer’s disease: initial results of a co-twin control study. Neurology1994;44:227–32.

[23] Meade EA, Smith WL, Dewitt DL. Differential inhibition of prostaglandin endoperoxide synthase(cyclooxygenase) isozyme by aspirin and other non-steroidal anti-inflammatory drugs. J Biol Chem1993;268:6610–4.

[24] Itagaki S, McGeer PL, Akiyama H, Zhu S, Selkoe D. Relationship of microglia and astrocytes to amyloiddeposits of Alzheimer’s disease. J Neuroimmunol 1989;24:173–82.

[25] McGeer PL, McGeer EG. Inflammation of the brain in Alzheimer’s disease: implications for therapy. JLeukoc Biol 1999;65:409–15.

[26] Huang TH, Yang DS, Plaskos NP, Go S, Yip CM, Fraser PE, et al. Structural studies of soluble oligomers ofthe Alzheimer beta-amyloid peptide. J Mol Biol 2000;297:73–87.

[27] Ko LW, Sheu KV, Young O, Thaler H, Blass JP. Expression in cultured human neuroblastoma cell of epitopesassociated with affected neurons in Alzheimer’s disease. Am J Pathol 1990;136:867–79.

[28] Gotti C, Briscini L, Verderio C, Oortgiesen M, Balestra B, Clementi F. Native nicotinic acetylcholine receptorsin human Imr32 neuroblastoma cells: functional, immunological and pharmacological properties. Eur JNeurosci 1995;7:2083–92.

[29] Paris D, Town T, Parker TA, et al. Inhibition of Alzheimer’s�-amyloid induced vasoactivity and microglialinflammatory response in microglia by a cGMP-dependent mechanism. Exp Neurol 1999;157:211–21.

[30] Kim HJ, Kim KW, Yu BP, Chung HY. The effect of age on cyclooxygenase gene expression: NF-kappaBactivation and IkappaBalpha degradation. Free Radic Biol Med 2000;28:683–92.

[31] Hensley K, Floyd RA, Zheng NY, et al. p38 kinase is activated in the Alzheimer’s disease brain. J Neurochem1999;72:2053–8.

[32] Hoffman SW, Moore S, Ellis EF. Isoprostanes: free radical-generated prostaglandins with constrictor effectson cerebral arterioles. Stroke 1997;28:844–9.

[33] Li DY, Hardy P, Abran D, et al. Key role for cyclooxygenase-2 in PGE2 and PGF2alpha receptor regulationand cerebral blood flow of the newborn. Am J Physiol 1997;273:R1283–90.

[34] Rodriguez G, Nobili F, Copello F, et al. 99mTc-HMPAO regional cerebral blood flow and quantitativeelectroencephalography in Alzheimer’s disease: a correlative study. J Nucl Med 1999;40:522–9.

[35] Lee RK, Knapp S, Wurtman RJ. Prostaglandin E2 stimulates amyloid precursor protein gene expression:inhibition by immunosuppressants. J Neurosci 1999;19:940–7.

[36] Paris D, Town T, Mullan M. Novel strategies for opposing murine microglial activation. Neurosci Lett2000;278:5–8.

[37] Wenk GL, McGann K, Mencarelli A, Hauss-Wegrzyniak B, Del Soldato P, Fiorucci S. Mechanisms to preventthe toxicity of chronic neuroinflammation on forebrain cholinergic neurons. Eur J Pharmacol 2000;402:77–85.

[38] Willard LB, Hauss-Wegrzyniak B, Danysz W, Wenk GL. The cytotoxicity of chronic neuroinflammationupon basal forebrain cholinergic neurons of rats can be attenuated by glutamatergic antagonism orcyclooxygenase-2 inhibition. Exp Brain Res 2000;134:58–65.

[39] Bezzi P, Carmignoto G, Pasti L, et al. Prostaglandins stimulate calcium-dependent glutamate release inastrocytes. Nature 1998;391:281–5.

[40] Lim GP, Yang F, Chu T, et al. Ibuprofen suppresses plaque pathology and inflammation in a mouse modelfor Alzheimer’s disease. J Neurosci 2000;20:5709–14.

Documents

Pro-inflammatory effect of freshly solubilized β-amyloid peptides in the brain