From cholesterol to oxidative stress in Alzheimer’s disease

(1)

From cholesterol to oxidative stress in Alzheimer’s disease

A wide perspective on a multifactorial disease Susanne Akterin

Department of Neurobiology, Care Sciences and Society (NVS) KI-Alzheimer’s Disease Research Center

Karolinska Institutet, Stockholm, Sweden

Stockholm 2008

(2)

All previously published papers were reproduced with permission from the publishers.

Paper i: © 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Paper ii: © 2008 The Authors; Journal Compilation © 2008 International Society of Neuropathology. Paper iii: © 2006 Nature Publishing Group.

Figure 3, showing confirmation of apoE3 and apoE4, was reproduced with kind permission of Springer Science and Business Media from Journal of Molecular Neuroscience, Volume 23, 2004, p 194, Apolipoprotein E: Diversity of Cellular Origins, Structural and Biophysical Properties, and Effects in Alzheimer’s Disease, Y. Huang, K. H. Weisgraber, L. Mucke, and R. W. Mahley, Figure 1, © 2004 Humana Press Inc. All rights of any nature whatsoever reserved.

Published by Karolinska Institutet

isbn 978-91-7409-172-4

(3)

Till mormor

nu får du äntligen en egen doktor!

(4)

This is not the end. It is not even the beginning of the end.

But it is, perhaps, the end of the beginning.

Sir Winston Churchill

(5)

Abstract

Epidemiological studies have provided evidence that high cholesterol levels in midlife and lack of antioxidants could render people more susceptible to develop AD. The aim of this thesis project was to get a more profound view of how cholesterol and oxidative stress could modify the development of Alzheimer’s disease (AD) on the molecular level, by studying mechanisms of signal transduction.

In Paper I we studied the effects of ApoE deficiency in combination with high fat/high cholesterol intake on mouse brain, and found that these two factors have synergistic effect on tau phosphorylation, causing hyperphosphorylation.

In Paper II we used microarray to examine how the expression of individual genes in brain is affected by long-term high fat/high cholesterol diet. We found changes in a limited number of genes, of which several have previously been linked to neurodegenerative processes. We focused our investigation on activity- regulated cytoskeletal-associated protein (Arc), that is important for synaptic plasticity and memory. Arc expression was decreased in brains from mice that received a high-fat/high cholesterol diet, as well as in Alzheimer brain. It is likely that this effect was induced by 27-hydroxycholesterol, a cholesterol metabolite that is able to cross the blood-brain barrier.

We went on to study how oxidative stress affects the brain and in Paper III we found that glutaredoxin-1 (Grx1) and thioredoxin-1 (Trx1) are first in line to protect cells against oxidative stress caused by Aβ. Levels of Grx1 and Trx1 were found to be affected in AD brain and this de-regulation could result in activation of apoptosis signal-regulating kinase (ASK)1, with important consequences for Aβ-induced cell death.

The final study concerned thioredoxin-80, a cleavage product of Trx1, earlier found to be produced when Trx1 is oxidized. In Paper IV we found that Trx80 is present in brain, mainly in neurons, that it is produced by cleavage of Trx1, by ADAM10/17 (α-secretase), and that the levels were decreased in AD brain.

Suggested functions of Trx80 in brain are to protect against Aβ-induced cell death, possibly by regulating apoptosis, by degradation of ASK1.

In conclusion, AD is a heterogeneous disease, affecting many different cellular processes and signaling pathways, a notion that is further supported by the results presented in this thesis. Since we do not yet know the underlying mechanisms for the disease development the study of risk factors is an alternative way of addressing this problem. The field of risk factors for AD has lately developed rapidly, but more effort is needed to understand the underlying mechanisms for the risk modifying effects. Further investigation in this area is needed, and will provide knowledge on how one should live to prevent the development of AD as well as suggest new targets for drug therapy.

(6)

Sammanfattning på svenska

Alzheimers sjukdom drabbar hjärnan och leder till senildemens. Demens drabbar främst äldre och i befolkningen över 65 år har 7% någon form av demenssjukdom, för 85-åringarna och äldre är det 20-30%. Alzheimer är den vanligaste formen av demens, ca 50-70% av demenspatienterna har Alzheimer och man uppskattar att det finns ca 90 000 Alzheimerpatienter i Sverige och att detta kommer att öka med en åldrande befolkning och förändrad livstil. Lidandet för patienter och anhöriga är stort och eftersom vården av demenspatienter är mycket dyr, innebär det även stora kostnader för samhället.

De främsta symptomen är problem med minne, språk, orienteringsförmåga och verklighetsuppfattning. Detta kan förklaras av att det är just de områden som styr dessa funktioner i hjärnan som påverkas i ett gradvis förlopp under sjukdomen. Studerar man en hjärna från en person som har dött av Alzheimer ser man främst två typer av sjukliga förändringar, plack och nystan. Placken består av ett ämne som kallas beta-amyloid och som utsöndras från cellerna.

Detta ämne klumpar lätt ihop sig och bildar dessa större ansamlingar som kallas plack. Beta-amyloid är giftigt för hjärnans celler och i placken finns rester från döda celler. Man kan nästan säga att placken kan liknas vid ärr i hjärnan.

Nystanen finns inuti nervcellerna och är långa trådformade fibrer bestående av ett ämne som heter tau. Nystanen hindrar cellen från att fungera normalt, vilket till slut leder till att den dör.

En annan sak som är slående när man studerar en hjärna från en Alzheimer- patient är att den ser ut att ha krympt eller skrumpnat. Detta beror på att en mängd celler i hjärnan har dött av påverkan från beta-amyloid och nystan, vilket påverkat den totala volymen. Hjärnan har en enorm reservkapacitiet, vilket gör att det tar väldigt lång tid för plack och nystan att påverka hjärnan så att man märker av effekterna, troligen flera decennier. Detta leder till att det är svårt att behandla mot Alzheimer, eftersom när symptomen uppträder har sjukdomen readan pågått under en lång tid och effekterna har blivit påtagliga. Nervceller kan inte nybildas naturligt och därför kan inte de delar av hjärnan som skadats återställas. De läkemedel som används idag stoppar inte sjukdomsförloppet utan bromsar upp symtomen under en kortare tid. Främst används så kallade Acetylkolinesterasinhibitorer, som ökar ämnet acetylkolin i hjärnan, vilket bland annat är viktigt för minnesfunktionen. Mängden acetylkolin minskar i hjärnan under Alzheimers sjukdom och med acetylkolinesterasinhibitorer kan man återställa nivåerna till viss del under en tid, men så småningom räcker inte denna effekt till.

Man känner inte till de bakomliggande orsakerna till varför man drabbas av Alzheimer, men man vet om att det finns ett antal faktorer som gör att man har en förhöjd risk att drabbas. Den främsta riskfaktorn är att ha en viss typ av en gen som styr produktionen av apolipoprotein E (ApoE). Alla människor har denna gen, men den finns i tre olika varianter, E2, E3 och E4. Det är just varianten E4 som bidrar till ökad risk för Alzheimer och denna variant finns

(7)

hos ca 15-20% av befolkningen. Det är viktigt att påpeka att man kan leva ett långt liv utan att drabbas av Alzheimer, även om man är apoE4-bärare, men risken att drabbas ökar med ca 3-4 gånger jämfört med de som har de andra genvarianterna. Andra riskfaktorer som är mindre vanliga och inte lika väl kartlagda är om man har någon hjärt-kärlsjukdom, högt blodtryck, diabetes typ 2 eller förhöjda kolesterolvärden. Det finns också studier som har visat att så kallad oxidativ stress i hjärnan och lägre intag av antioxidanter via födan kan öka risken.

Projektet som ligger till grund för denna avhandling har syftat till att ta reda på de mekanismer som ligger bakom hur olika riskfaktorer påverkar utvecklandet av sjukdomen. Jag har främst tittat på hur kolesterol påverkar hjärnan samt vilka mekanismer som är påverkade under oxidativ stress. Nedan föjer en kort sammanfattning av de fyra delarbeten som ingår i avhandlingen.

Studie 1

I den första studien använde vi en musmodell som saknar apoE och eftersom apoE4 hos människan saknar många funktioner jämfört med apoE2 och E3 kan man använda djur som saknar apoE för att härma effekterna av apoE4. Dessa möss fick under nio månaders tid en kost med hög halt av fett, socker och kolesterol, liknande det näringsinnehåll som finns i snabbmat. När vi studerade mössens hjärnor såg vi att tau var kemiskt förändrat på ett typiskt sätt, med ett ökat antal fosfatgrupper (så kallad “hyperfosforylering”). Tau fosforyleras naturligt, men denna överdrivna fosforylering gör att tau inte fungerar som det ska och man har funnit att denna modifiering är ett förstadium till bildandet av de nystan som man ser i Alzheimer-hjärnan. Vi studerade också hur dessa modifieringar uppstått och fann att den typ av ämne som är ansvarigt för att sätta dit fosfatgrupper på tau (så kallade “kinaser”) var överaktivt.

Det är svårt att veta hur detta kan översättas till människa, men en tänkbar slutsats från denna studie är att bildandet av nystan vid Alzheimer skulle kunna orsakas av en samverkan mellan genetiska faktorer, som apoE4, och felaktig kost.

Studie 2

Eftersom vi i studie 1 sett effekter i hjärnan på möss som fått kost med hög kolesterolhalt var vi intresserade av att se vilka gener som påverkas i hjärnan av denna typ av mat. Människor har ca 20 000 – 25 000 gener, men alla är inte påslagna hela tiden, eller i alla celler. Många olika faktorer påverkar vilka gener som är påslagna, bland annat den miljö vi vistas i och den mat vi äter.

I hjärnans blodkärl finns ett spciellt skydd, blod-hjärnbarriären, som gör att ämnen i blodet inte ska kunna komma in i hjärnan. Kolesterol i blodet kan inte gå igenom blod-hjärnbarriären och därför är det oklart hur kolesterol från maten kan påverka hjärnan. Nyligen visade en annan grupp att kolesterol kan ombildas till 27-hydroxykolesterol, som kan passera in i hjärnan.

I denna studie använde vi helt vanliga möss, som fick samma kost som i studie 1.

Vi använde oss av en speciell metod (microarray) med vilken man samtidigt kan

(8)

studera 15 000 gener och se vilka som är av- och påslagna. Av alla dessa gener fann vi att endast 37 gener i hjärnan förändrats i de möss som fått en kost med hög halt av fett och kolesterol, jämfört med möss som fått vanlig kost. Utifrån dessa gener valde vi ut Arc, som är inblandat i minnesinlagring, som en kandidat för fortsatta undersökningar. När vi undersökte nivåerna av Arc i hjärnan såg vi en minskning jämfört med möss som fått vanlig kost. Vidare studerade vi de ämnen som styr produktionen av Arc och såg att även halterna av dessa hade minskat. För att undersöka om det var just kolesterolen i maten som påverat Arc, använde vi nervceller som behandlades med 27-hydroxycholesterol, den form av kolesterol som kan gå in i hjärnan från blodet. Resultaten från detta experiment överensstämde med de som vi tidigare sett för Arc, och de ämnen som styr Arc i mushjärnan. Vår hypotes är att kolesterolet i maten omvandlas till 27-hydroxykolesterol i kroppen, som sedan går in i hjärnan och leder till sänkta nivåer av Arc. Eftersom Arc är inblandat i minnesinlagring kan denna minskning vara av betydelse för utvecklandet av de minnesstörningar som man ser vid Alzheimer.

Studie 3

Vi övergick nu till att studera de mekanismer som styr oxidativ stress och att titta närmare på två kroppsegna antioxidanter som finns i hjärnan och skyddar mot oxidativ stress, glutaredoxin-1 och thioredoxin-1. Vi fann i hjärnor från Alzheimer-patienter att mängden glutaredoxin-1 var förhöjd, medan mängden thioredoxin-1 hade minskat. Vi fortsatte med att studera vilka processer som kan ligga bakom detta fynd och för detta använde vi cellkulturer. Både glutaredoxin och thioredoxin finns i en oxiderad och en reducerad form och det är den reducerade som är aktiv och som kan motverka oxidativ stress i cellen.

Det är känt sedan tidigare att beta-amyloid, den främsta beståndsdelen i placken, är skadlig för celler just för att den bidrar till oxidativ stress. När vi behandlade celler med beta-amyloid fann vi att både glutaredoxin-1 och thioredoxin-1 snabbt övergick till sin oxiderade, inaktiva form, men att de efter 24 timmars behandling återigen blivit reducerade (aktiva). Detta tyder på att glutaredoxin-1 och thioredoxin-1 är ett av kroppens tidigaste och främsta försvar mot beta- amyloid-inducerad oxidativ stress.

Både thioredoxin och glutaredoxin är inblandade i regleringen av en särskild typ av celldöd, så kallad programmerad celldöd (apoptos). Denna celldöd sker när specifika och välreglerade signaler ges i cellen och är inte slumpmässig.

Vi såg att beta-amyloid aktiverade dessa celldödssignaler, men genom att öka mängden thioredoxin-1 eller glutaredoxin-1 i cellen, på artificiell väg, kunde dessa celldödssignaler stoppas och cellerna överlevde.

Våra slutsatser är att thioredoxin-1 och glutaredoxin-1 inte fungerar som de ska i hjärnan hos de som drabbas av Alzheimer och att celldödssignaler därför kan aktiveras och leda till den celldöd som man ser som en följd av sjukdomen.

(9)

Studie 4

I den sista studien valde vi att gå vidare och titta på en nedbrytningsprodukt av thioredoxin-1, ett ämne som kallas thioredoxin-80. Detta ämne bildas genom att den yttersta delen av thioredoxin-1, ca 20%, “klipps” bort. Det ämne som fungerar som sax och utför detta klipp är inte känt. Thioredoxin-80 är ett ämne som upptäcktes för inte så länge sedan och därför vet vi inte så mycket om det, t ex var i kroppen det finns eller vilken funktion det har. Vi började därför med att studera om det fanns i hjärnan och vi såg att thioredoxin-80 främst finns i nervceller. När vi sedan studerade hjärnor från Alzheimer-patienter såg vi en drastisk minskning av thioredoxin-80.

Vi ville ta reda på hur thioredoxin-80 bildas och undersökte ett ämne som heter α-sekretas och vars funktion är nedsatt i Alzheimers sjukdom. Detta ämne fungerar som sax och deltar i olika processer i kroppen, där ämnen ska klippas sönder i mindre delar. Vi fann att när vi på kemisk väg hindrade α-sekretaset att klippa minskade nivåerna av thioredoxin-80. Vi tror därför att α-sekretas deltar i bildningen av thioredoxin-80 från thioredoxin-1.

Vi studerade även möjliga funktioner som thioredoxin-80 skulle kunna vara inblandat i. Thioredoxin har i blodceller visats bidra till att aktivera inflammation, men inga tecken på detta sågs i hjärnan. Eftersom vi i studie 3 såg att thioredoxin-1 är inblandat i celldödssignalering undersökte vi om även thioredoxin-80 kunde vara inblandat i samma process. I en preliminär studie såg vi att ASK1, ett av de ämnen som är särskilt viktiga för just beta-amyloid- aktiverad celldöd, minskade när vi ökade mängden thioredoxin-80 i celler på artificiell väg. En teori är därför att thioredoxin-80 deltar i styrningen av ASK1, men eftersom mängden thioredoxin-80 är minskad i Alzheimer, skulle denna styrning vara satt ur spel och leda till celldöd.

Denna avhandling är främst ämnad att studera grundläggande mekanismer för hur två olika riskfaktorer, höga kolesterolnivåer och oxidativ stress, ökar risken för Alzheimers sjukdom. Generellt har vi funnit att olika riskfaktorer kan samverka, vilket leder till synergieffekter. Mer specifikt har vi sett att möss som fått en kost med mycket fett, kolesterol och socker utvecklar ett förstadium till de nystan som man ser i hjärnan hos Alzheimer-patienter och som kan leda till att cellerna dör. Vi har också sett att två antioxidanter inte fungerar som de ska i hjärnan hos Alzheimer-patienter, vilket även det kan leda till att nervceller dör.

Med denna avhandling vill jag visa att det är viktigt att studera riskfaktorer för Alzheimer, eftersom det är ett bra sätt att öka kunskapen om hur sjukdomen uppstår och för att kunna förhindra detta. I nuläget finns det inga effektiva mediciner mot Alzheimer och sjukdomen i sig är svårbehandlad eftersom döda nervceller inte kan återställas. Dessa faktum gör att det krävs ökade resurser på förebyggande verksamhet. Resultaten i denna avhandling ger vissa indikationer på hur man skulle kunna förebygga Alzheimer, men för att kunna utveckla råd som kan komma allmänheten till del krävs mer forskning på detta område.

(10)

List of abbreviations

a.a . . . .amino acids ABCA1 . . . . ATP-binding cassette A1-transporter ACE . . . . angiotensin I converting enzyme AD . . . . Alzheimer's disease ADAM . . . . a disintegrin and metalloproteinase apoE . . . . apolipoprotein E APP . . . . amyloid precursor protein ASK1 . . . . apoptosis signal-regulating kinase 1 AVP . . . .arginine-vasopressin Aβ . . . . amyloid-β peptide BACE . . . . beta-site APP cleaving enzyme BBB . . . . blood-brain barrier CaMKII . . . .calcium and calmodulin-dependent protein kinase-II cdk5. . . .cyclin dependent protein kinase-5 CK-1 . . . . casein kinase-1 CSF . . . . cerebrospinal fluid DMSO . . . .dimethyl sulfoxide DTT . . . . dithiothreitol ELISA . . . . enzyme-linked immunosorbent assay ERK1/2 . . . . extracellular-regulated kinase 1/2 Grx . . . . glutaredoxin GSH/GSSG . . . . glutathione (reduced/oxidized) GSK3 . . . . glycogen synthase kinase 3 HDL . . . . high density lipoprotein HMG-CoA . . . . 3-hydroxy-3-methylglutaryl coenzyme A ICC . . . . immunocytochemistry IHC . . . . immunohistochemistry IL . . . . interleukin JNK . . . . c-Jun N-terminal Kinase LDLR . . . . low density lipoprotein (LDL) receptor LTP . . . . long term potentiation MAPK . . . .mitogen-activated protein kinase MCI . . . . mild cognitive impairment NFT . . . . neurofibrillary tangles NMDAR . . . . N-methyl-D-aspartate receptors NPC . . . . Niemann-Pick type C disease NQO1 . . . .NAD(P)H:quinone oxidoreductase 1 PHF . . . . paired helical filaments PKA . . . . protein kinase A PP . . . . protein phosphatase PSEN1/2, PS1/2 . . . .presenilin 1/2 RIP . . . . regulated intramembrane proteolysis RNS . . . . reactive nitrogen species ROS . . . .reactive oxygen species SOD . . . . superoxide dismutase TNFα . . . . tumor necrosis factor Trx . . . .thioredoxin TrxR . . . .thioredoxin reductase VLDL . . . . very low density lipoprotein

(11)

List of publications

This thesis is based on the following papers, which will be referred to in the text by their roman numerals.

I. A. Rahman, S. Akterin, A. Flores-Morales, M. Crisby, M. Kivipelto, M. Schultzberg, A. Cedazo-Mínguez. High cholesterol diet induces tau hyperphosphorylation in apolipoprotein E deficient mice.

FEBS Letters (2005) 579, 6411-6416.

II. L. Mateos, S. Akterin, F. Gil-Bea, S. Spulber, A. Rahman, I. Björkhem, M. Schultzberg, A. Flores-Morales, A. Cedazo-Mínguez. Activity- Regulated Cytoskeleton-Associated Protein in Rodent Brain is Down Regulated by High Fat Diet in vivo and by 27-Hydroxycholesterol in vitro.

Brain Pathology (2008). In press. doi:10.1111/j.1750-3639.2008.00174.x III. S. Akterin, R.F. Cowburn, A. Miranda-Vizuete, A. Jiménez,

N. Bogdanovic, B. Winblad, A. Cedazo-Mínguez. Involvement of Glutaredoxin-1 and Thioredoxin-1 in beta-amyloid toxicity and Alzheimer’s disease. Cell Death and Differentiation (2006) 13, 1454-65.

IV. S. Akterin*, F. Gil-Bea*, L. Mateos, D. Lorenzo-Villegas, S. Costa, J. Avila-Cariño, E. Sundström, A. Holmgren, A. Cedazo-Mínguez. The production of thioredoxin-80 from thioredoxin-1 is mediated by alpha secretase and decrease in Alzheimer’s disease. Manuscript.

* These authors contributed equally to the work.

(12)

Introduction

A

lzheimer’s disease (AD) is the most common cause of dementia in the elderly and is responsible for 50-70% of dementia cases [1]. The patients suffer from drastic cognitive deficits, with memory loss, delusions and problems with carrying out functions of normal daily living. There is at present no cure for AD and the available drugs are designed to ameliorate the symptoms. AD is a devastating disease because of the way it affects the mind and the personality. The relatives of a patient can often bear witness to how they have seen the patient slowly drift away, not being able to recognize their loved ones, leaving only the crumbling body to care for.

In addition to the tragic effects for the patients and the closest family, AD is extremely costly for society. The increased incidence of dementia puts pressure on researchers worldwide and poses a financial challenge for the world community as the majority of the patients are now living in developing countries. It was estimated that 24 million people suffered from dementia in 2001, and that this number will double every 20 years, because of an aging population and change in life style. The regional prevalence in people aged 60 years or older is 1.6% in Africa, 3.9% in eastern Europe, 4.0% in China, 4.6%

in Latin America, 5.4% in western Europe, and 6.4% in North America, most likely reflecting the influence of life style factors. [1].

ALzheImer neuropAThoLogy

The definite diagnosis of AD is based on the presence of neuropathological changes in the form of accumulation of extracellular plaques and intracellular neurofibrillary tangles (NFT). Other characteristics are synaptic loss and neuronal cell death, leading to a decrease in brain volume.

Braak-stageing is a commonly used method to stage the pathological

(16)

process, taking into account both the NFT and plaque pathology. The temporal and regional distribution differs between NFTs and plaques.

NFT pathology starts in the medial temporal lobe (entorhinal cortex and hippocampus) and spreads to limbic areas, and finally to neocortical association areas and primary cortex [2]. The plaques are first visible in orbitofrontal and temporal cortices and spread further to parietal cortex and throughout the neocortex. The

cerebellum is usually spared from pathological lesions. The symptoms reflect the neuropathology; the first signs are short-term memory problems, which reflect the early pathology in the hippocampus. The memory problems will later develop into difficulties with executive functions, including planning and initiation of actions, as well as emotional disturbances and apathy. Executive functions are mainly controlled by the prefrontal cortex and interconnected cortical and subcortical brain structures.

One of the striking findings in the AD brain is the selective loss of cholinergic neurons that provide the major cholinergic input to neocortex and cerebral cortex [3]. The major class of drugs prescribed for AD is acetylcholinesterase inhibitors that will increase the availability of acetylcholine, and enhance the cholinergic functions, to counteract the loss of cholinergic innervation.

Amyloid plaques consist of Aβ-peptide

The extracellular plaques found in AD brain are so called amyloid plaques and are made up of amyloid-β peptide (Aβ) [4]. Aβ is a peptide of 39-43 amino acids (a.a) that is able to form β-sheet structures and fibrillar aggregates, hence the term “amyloid” that refers to the property of peptides to form insoluble filaments under physiological conditions, detectable by their binding to Congo Red [5].

There are two forms of amyloid plaques in the AD brain: neuritic plaques (also called senile plaques) and diffuse plaques. The neuritic plaques are extracellular deposits of fibrillar Aβ, with dystrophic neurites that show NFT pathology, as well as activated microglia within the central amyloid core. Reactive astrocytes surround the neuritic plaques. The plaques can also be diffuse, lacking the fibrillar, compacted core and the neuritic dystrophy. It is believed that the diffuse plaques are immature precursors of the neuritic plaques. It is not known how long time is needed to develop neuritic plaques in human brain, but it is likely that it will take months or even years [6]. It has been shown that the activation of microglia by fibrillar Aβ is a very early phenomenon in the AD

Characteristics of the Alzheimer brain

✧ Neuritic plaques (extracellular beta-amyloid)

✧ Neurofibrillary tangles

(intracellular hyperphosphorylated tau)

✧ Inflammation

✧ Oxidative stress

✧ Synaptic loss

✧ Cell loss

(17)

pathogenesis, and that the localization of astrocytes at the neuritic plaques occurs much later, when dementia is developing [7]. One explanation for this could be that microglia has been found to be involved in the clearance of Aβ, by phagocytosis [8]. The activated microglia can also produce toxic products, like reactive oxygen species and pro-inflammatory cytokines that could instead contribute to neurodegeneration [7].

The amyloid cascade hypothesis – pros and cons

In the last 20 years the research into AD has intensified, but the cause or causes of AD are still not known. AD is believed to be multi-factorial, with the involvement of both environmental and genetic factors. Today, the “amyloid cascade hypothesis” [9] is the most prominent of the many theories proposed over the years to explain the pathogenesis of AD. It states that the Aβ peptide acts to drive the neurodegenerative processes that cause AD and that the rest of the disease process results from an imbalance between Aβ production and clearance. The amyloid cascade hypothesis is at present the most dominant theory to explain the pathogenesis of AD.

In support of this hypothesis it should be mentioned that it was proposed in 1992 and has since then been the

focus of many studies. Despite many efforts to elucidate its deficiencies, an alternative hypothesis explaining the cause and early pathogenesis of AD, that has as much experimental support as the Aβ hypothesis, has not emerged. Nevertheless, there are several observations that suggest that it may not hold true in the end, or that it is at least lacking in detail. The biggest concern is that it does not explain the reason for the increased Aβ production in sporadic cases, where no mutations in the genes encoding amyloid precursor protein (APP) or presenilin 1/2 (PSEN1/2) are present. Another drawback is that similar plaque burden as found in patients with mild AD, with or without NFT, can also be present in non-demented subjects without affecting the cognitive performance [10, 11].

Mutations in APP or PSEN1/2

Altered intracellular signaling due to altered neuronal ionic homeostasis

Widespread neuronal dysfunction and cell death Tau hyperphosphorylation due to altered kinase/phosphatase activity Microglia and astrocyte activation

Pro-inflammatory response Subtle effects on synapses

by AB(1-42) toxicity

AB(1-42) oligomerizarion and aggregation Formation of diffuse plaques Increased production of AB(1-42)

Dementia

Figure 1 • Amyloid cascade hypothesis.

(18)

The final test for the amyloid cascade hypothesis in humans remains: to study if the cognitive performance is affected by a reduction in Aβ levels in the brain. Results from a small phase I immunization study, using active Aβ(1-42), were recently published, and although immunization had resulted in clearance of plaques this clearance did not prevent further progressive neurodegenerative changes in the AD patients [12]. Larger immunization studies, also with passive immunizations, are now under way and could potentially come to different conclusions.

Aβ is generated from the Amyloid Precursor Protein (APP), a transmembrane protein with a large extracellular domain and one transmembrane region.

APP can be cleaved by different secretases in an intricate cascade manner, giving rise to an amyloidogenic and to a non-amyloidogenic pathway (Figure 2). The γ-secretase cleavage occurs within the membrane in a process called regulated intramembrane proteolysis (RIP) and is preceded by the rate-limiting ectodomain shedding performed by either α- or β-secretases. If APP is cleaved by β-secretase, followed by γ-secretase, the result is the toxic Aβ peptide. The α-secretase cleaves in the middle of the Aβ-region [13] thus preventing the formation of Aβ peptide.

The molecular identities for all three secretases have now been proposed.

The γ-secretase is a protein complex with presenilin (PS) 1 or 2 as the catalytic subunit, [14]. β-secretase activity is performed by the aspartic proteases beta- site APP cleaving enzyme (BACE)1 and 2 [15]. The identity of the α-secretase is less understood. Enzymes with α-secretase activity have been identified within the A Disintegrin And Metalloproteinase (ADAM)-family and it is suggested that at least ADAM9, 10 and 17 could be partially responsible for α-cleavage of APP. ADAM10 and 17 knockout mice are perinatally lethal and deficiency experiments in cells have resulted in no or small effects on APP cleavage. It has

AB

p3 B AB

AG

sAPPB sAPPA

AICD C99

APP C83

AICD

B

A G

G

Extracellular space Cytoplasm

Figure 2 • Processing of Amyloid Precursor Protein. APP is sequentially cleaved by different secretases (α/γ and β/γ). AICD, APP intracellular domain; C83/99, C-terminal 83 a.a./99a.a; sAPP, soluble APP; p3, peptide fragment, 3 a.a.

(19)

therefore proven difficult to discern the identity of α-secretase. It is believed that α-secretase activity is an act of ADAM9, 10 and 17 in combination, possibly also with help from other ADAM family members, mutually compensating for the lack of one another [16].

neurofibrillary tangles are composed of hyperphosphorylated tau

Both plaque- and neurofibrillary pathologies are required for AD to develop, but the degree of dementia is best correlated to the number of tangles, not the plaques.

NFTs are intraneuronal inclusions of abnormally hyperphosphorylated tau, a microtubuli-associated protein, which self-assembles into paired helical filaments (PHF) making up the tangles. Normal tau binds tubulin and is important for microtubuli assembly, but the hyperphosphorylated form will no longer bind tubulin, or promote the assembly into microtubules. In contrast, hyperphosphorylated tau will instead sequester normal tau, leading to disrupted microtubules. The disruption of the main structure for axonal transport will compromise the transport within cell, with drastic effects in neurons that are large and polarized [17, 18], and can eventually lead to retrograde degeneration with synapse loss and ultimately cell death.

Tau-hyperphosphorylation can be a result of 1) increased kinase activity or 2) decreased phosphatase activity. Tau is phosphorylated at at least 37 different serine or threonine sites [19] and many different kinases are involved: Glycogen synthase kinase 3 (GSK3), cyclin dependent protein kinase-5 (cdk5), casein kinase-1 (CK-1), protein kinase A (PKA), calcium and calmodulin-dependent protein kinase-II (CaMKII), and extracellular-regulated kinase (ERK) 1/2. The sequential phosphorylation of tau by priming kinases (PKA and CaMKII), that will markedly increase the phosphorylation of the other kinases, is believed to be important for hyperphosphorylation to occur [18]. The protein phosphatase (PP)2A is believed to be the major phosphatase for tau, but PP-1 is also involved in the dephosphorylation. The activities of both of these have been reported to be decreased by 20% in AD brain [18]. Tau hyperphosphorylation is most likely caused by a disturbance of the balance between kinases and phosphatases.

geneTICS And rISk FACTorS

There are two subtypes of AD, one caused by mutations (familial AD), and the sporadic form, the most common, which is heterogeneous and multifactorial in its pathogenesis. The cause of sporadic AD is thought to be a combination of genotype with several environmental risk factors, making certain individuals more susceptible to developing AD. Since the discovery of several mutations in APP or PSEN1/2 genes that cause familial forms of AD, much work has been focused on studying how these mutations are involved in AD pathogenesis. A common characteristic of these mutations is that they alter the proteolytical

(20)

processing of APP, which leads to an increased production of total Aβ or of the more aggregate prone Aβ(1-42) [20]. However, though the familial forms of AD closely resemble the clinical manifestations of the sporadic forms, it is not clear how relevant it is to use these mutations as a model to find the cause of AD, since known genetic cases only represent less than 2% of all AD cases [21]. Since the increased Aβ-production in the sporadic cases has a far more complex basis there is a risk that the mutation models of AD will lead to a too simplistic view of the disease and future treatment strategies.

The search for susceptibility genes in AD has been mostly disappointing. Over the years many reports have been published describing new disease-modifying genes in small populations, but few have been reproduced in larger samples or different settings [22]. The lack of success in finding individual genes that affect AD pathogenesis highlights the notion that AD is a complex disease where both genetic and environmental factors combine during life and influence disease development. The combination of different risk factors makes up the risk profile, determining the risk of an individual of developing AD. It is also likely that different individuals will be differently susceptible towards certain risk factors. In recent years much knowledge has been gained about risk factors from large epidemiological studies, but less is understood about the mechanisms by which these factors contribute to the disease development. To be able to use the information from epidemiological studies for future therapies, it is necessary to understand which signaling pathways are affected by certain risk factors and how these pathways can be influenced to change the course of the disease.

The most important genetic risk factor for AD is the presence of one or more APOE ε4 alleles [23-25]. This is the only genetic risk factor that has repeatedly, and in many different populations, been shown to increase the risk of getting AD. Other influencing factors include oxidative stress, high dietary cholesterol intake and medical risk factors such as traumatic brain injury, low cerebral perfusion, cerebral infarct and stroke, cardiovascular disease, depression and possibly also diabetes mellitus, type 2 [26, 27].

ApoLIpoproTeIn e

Apolipoprotein E (apoE) was first discovered in 1973 as a component of triglyceride-rich lipoproteins complexes and as such it was suggested to be

Suggested risk factors for Alzheimer’s disease

✧ Age

✧ Apolipoprotein E4 isoform

✧ High dietary cholesterol intake

✧ Oxidative stress

✧ Low cerebral perfusion

✧ High blood pressure

✧ Traumatic brain injury

✧ Cerebral infarct or stroke

✧ Cardiovascular disease

✧ Depression

✧ Diabetes mellitus, type 2

(21)

important for cholesterol metabolism as it binds and transports cholesterol between different cells and in the blood. The human APOE gene has three alleles (ε2, ε3, ε4) that give rise to six different phenotypes (apoE2/2, E2/3, E2/4, E3/3, E3/4, E4/4). The apoE3 isoform is the most common isoform and E2 the least. Today several new functions have been ascribed for apoE, and since it was discovered that the E4 isoform of apoE is a risk factor for AD, the research has been intense in trying to understand why. Many hypotheses have been formulated that, in general, can be simplified into two: that apoE4 either has a lack of protective or a gain of deleterious functions compared to the other isoforms.

The only difference between the isoforms is a substitution of Cys to Arg at position 112 and 158. This small difference has large consequences for many biological functions in which apoE is involved [28]. Most other species, including the non-human primates, have a sequence similar to apoE4, which is why it is believed that this is the ancestral isoform [29]. Human apoE contains Arg at position 61, in contrast to all other species where apoE has been sequenced, which contain Thr61 instead. This affects the structure of apoE4, introducing a salt bridge between the N- and C-terminus (Arg61 – Glu255), called domain interaction (Figure 3)[30]. The apoE2 and E3 isoforms as well as apoE from other species do not display domain interaction. It is likely that the domain interaction is what allows the apoE4 molecule to acquire the more detrimental properties.

The lipid-binding domain is affected by the domain interaction in such a Figure 3 • Structure of apoE isoforms E3 and E4. The amino acid substitution at position 112 is indicated, as well as the domain interaction between N- and C-terminus in apoE4, see text for details.

(22)

way that apoE4 preferentially binds to very low density lipoprotein (VLDL), and apoE2 and E3 to high density lipoprotein (HDL). In the brain and in cerebrospinal fluid (CSF), the lipoproteins differ considerably compared to serum, and studies have shown that the majority of CSF lipoproteins are in a size-range and density similar to HDL particles [31]. Thus, the relevance of differences between isoforms in lipid binding in brain is not clear.

ApoE is a carrier of lipoproteins and will deliver these to cells by binding to the receptors of the low density lipoprotein (LDL) receptor family (LDLR, VLDLR, apoER2, LDLR-related protein [LRP]) in an isoform-specific manner.

ApoE2 shows a low affinity to LDLR, and lipid-free apoE binds well to VLDLR, but not to the other receptors [32]. Until a few years ago it was believed that the only function of the LDLRs was to supply cells with cholesterol and to remove lipoprotein particles from the blood. It is now known that these receptors also have important functions in signal transduction. The ApoE binding will start a signaling cascade in the complex network of scaffold proteins interacting with the cytoplasmic tail of the LDLRs. Disabled-1 (DAB1), FE-65, c-Jun N-terminal Kinase Interacting Protein 1/2 ( JIP1/2) and PSD-95 can bind the NPxY motif of the LDLR and get recruited to the receptor. The binding can start kinase cascades and activation of c-Jun N-terminal Kinase ( JNK) and GSK-3β etc., leading to neurite extension, cell adhesion, vesicle trafficking, neurotransmission or cell death [28, 33, 34].

ChoLeSTeroL

Several studies have investigated the connection between increased cholesterol levels in serum or plasma and AD [35-40]. It is worth noting that the majority of the prospective studies have found a correlation between cholesterol level, when measured in mid-life, and risk of later developing AD. It is also clear that several cardiovascular risk factors, e.g hypertension, diabetes type 2 or metabolic syndrome, can interact with hypercholesterolemia to further increase the risk of developing AD [40].

Another line of evidence for the cholesterol-AD link is that several long- term retrospective case-control studies [41-43] have reported that statins, inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMG-CoA reductase), lower the risk of AD or dementia, with up to 70%. Since the initial reports several studies have investigated the effects of statins prospectively and also in clinical trials (reviewed by Kuller [44]), but with disappointing results.

However, the growing list of interconnections between cholesterol and AD indicate that statins, that inhibit de-novo cholesterol synthesis in brain, have a disease modifying effect. The discrepancy between the prospective studies and the case-control studies could possibly be explained by the fact that the effects of the statins will be eliminated unless they are given at an early stage of the disease, perhaps as early as mid-life. Too few study subjects and/or too short follow-up time could also be reasons for the lack of positive results.

(23)

It is worth noting that statins also have cholesterol-independent, or pleiotropic effects that could play an even larger role for AD. These effects are linked to the reduced levels of isoprenoid intermediates that will follow by the inhibition of synthesis of mevalonate by the statins [45]. Isoprenoids are involved in the activation of small G-proteins (Ras, Rho, Rac, Rab, etc.), therefore less isoprenoids will result in increased amounts of inactive G-proteins in the cytoplasm. Since these G-proteins are involved in many cell signaling pathways, the theoretical consequences for AD could be beneficiary, with less oxidative stress, inflammation, and apoptosis [46].

Cholesterol metabolism and homeostasis in the brain

Cholesterol metabolism in brain and blood are almost entirely separated, since cholesterol bound to large lipoproteins do not pass the blood-brain barrier (BBB). This means that the brain is self-sufficient in cholesterol and because of the cholesterol-enriched myelin that makes up 70% of the brain cholesterol, the brain is the most cholesterol-rich organ, harboring 25% of the total cholesterol in the body [47]. Astrocytes are the main cholesterol-producing cells in the brain. Neurons can also synthesize cholesterol, especially during development when the cholesterol turn-over is large, due to the generation of neurons and synapses. However, it has been suggested that this ability is turned off, in order to focus on more specialized functions, like generation of electrical activity [47]. Thus, neurons depend on cholesterol delivery from nearby cells, mainly astrocytes, e.g. for axonal regeneration, neurite extension and synaptogenesis.

The half-life of cholesterol in brain is very long, approximately five years, since in the adult, uninjured brain very little cholesterol is used up; instead it is efficiently recycled within the brain.

Astrocytes are also the main apoE-producing cells and they secrete cholesterol bound to lipoprotein particles where apoE is included as well. Extracellular apoE is thus the major cholesterol transporter and therefore tightly linked to the cholesterol homeostasis and the redistribution of cholesterol in the CNS.

Lipoprotein particles are secreted by astrocytes through the ATP-binding cassette (ABC)A1-transporter and the neurons will take up cholesterol through low-density lipoprotein (LDL) receptors on the surface of the neurons, that bind apoE (Figure 4)[48]. The apoE-dependent cholesterol delivery to neurons has isoform-specific differences; apoE3-expressing astrocytes release more cholesterol than apoE4-expressing astrocytes. In addition, the amount of cholesterol in each particle is almost two-fold in apoE3 lipoprotein particles, compared to apoE4 lipoprotein particles. This suggests that apoE3 can supply more cholesterol to the neurons, than apoE4 can [49], something which could have consequences for neurodegeneration after injury and cell damage.

(24)

oxysterols are able to pass the blood-brain barrier effectively

Despite the low turnover of brain cholesterol under normal circumstances, some cholesterol is excreted. To be able to pass the BBB, cholesterol has to be esterified to form oxysterols (Figure 4). The major hydroxylated sterol excreted from the brain is 24(S)-hydroxycholesterol, formed by the conversion of cholesterol 24-hydroxylase (CYP46A1). 24-hydroxylase is expressed by certain types of neurons, such as cortical pyramidal neurons and Purkinje cells in cerebellum. 24(S)-hydroxycholesterol is transported by the blood to the liver and then excreted into the bile [48].

Still the main question remains: how can high serum cholesterol affect the brain and the risk for AD when cholesterol does not pass the BBB? One possibility is that cholesterol could affect the homeostasis of other proteins or metabolites that are able to cross the BBB and thereby affect the brain. Interestingly, it was recently shown that there is a net-flux of the cholesterol derivative 27- hydroxycholesterol into the human brain [50]. 27-hydroxycholesterol is formed from cholesterol, by 27-hydroxylase (CYP27A1), in almost all cells of the body, but the pattern of distribution indicates that the influx from the periphery is more prominent than the in situ-synthesis in the brain [50]. The expression of 27-hydroxylase is decreased in neurons and astrocytes, but increased in oligodendrocytes in AD brain [51]. Despite this decrease in 27-hydroxylase,

Chl

24-OHChl CYP46A1

synaptogenesis neurite

extension

Liver, bile

Neurons

Acetyl-CoA HMG-CoA red.

Astrocytes

27-OHChl

ApoE ABCA1

ApoE LDLR Brain

Chl CYP27A1

BBB

Blood

CYP7B1 7A-OH

Figure 4 • Summary of cholesterol homeostasis in brain. Netfluxes in and out of the brain, of different oxysterols are indicated. For details, see text. 7α-OH, 7α- hydroxy-3-oxo-4-cholestenoic acid; 24(S)-OHChl, 24- h y d r o x y c h o l e s t e r o l ; 27-OHChl, 27-hydroxycholesterol; HMG-CoA red, HMG-Coenzyme A reductase.

(25)

higher levels of 27-hydroxycholesterol has been found in AD brain and in a mouse AD-model overexpressing APP with the Swedish mutation, APP751 [52]. This further indicates that the influx of 27-hydroxycholesterol to the brain may be the link between dietary cholesterol and the risk of developing AD.

27-hydroxycholesterol is also actively metabolized in the brain by neuronal CYP7B1, forming 7α-hydroxy-3-oxo-4-cholestenoic acid (7α-OH) that is excreted from the brain to the blood [53].

It has also been reported that during AD the pathological effects in the brain can disturb the integrity of the BBB, rendering it leaky, with the possibility that proteins from the blood can actually enter the brain. Some evidence for this was presented by Wisniewski et al. and Wada [54, 55].

Cholesterol affects App processing

Several reports have shown that APP processing and Aβ formation are regulated by cholesterol. Refolo et al. [56] showed that hypercholesterolemia increased Aβ accumulation in the brains of double-mutant PS1/APP transgenic mice, caused by an increase in the amyloidogenic processing of APP. There was also an increased deposition of Aβ plaques. Sharma and co-workers just recently demonstrated an up-regulation of levels of both Aβ(1-40) as well as Aβ(1-42) in rabbits in response to cholesterol-enriched diet and that these affects were replicated when hippocampal slice cultures from the same animals were treated with 27-hydroxycholesterol [57]. However, the effect on Aβ production is indirect, since it has been shown that both 27-hydroxycholesterol and 24(S)- hydroxycholesterol inhibit this production, although the latter approximately 1000-fold more potently. In vitro experiments have shown that by depleting cellular cholesterol levels by statin treatment, Aβ levels can be effectively reduced.

Additionally, it was also shown that Aβ generation could be up regulated by reintroducing cholesterol to the same cultures. The mechanisms behind these observations are not fully understood, but there are many reports describing different hypotheses (reviewed by [48, 58, 59]), as summarized below.

Since all cell membranes constitute cholesterol, though to a different extent, a change in membrane cholesterol will affect the ordering, rigidity and fluidity of the membrane. The main hypothesis of how cholesterol is affecting Aβ production suggests that the amyloidogenic processing of APP occurs at lipid rafts. Both β- and γ-secretase are primarily located in cholesterol- and sphingomyelin-enriched areas of intracellular membrane compartments or the plasma membrane, referred to as lipid rafts or detergent resistant membranes.

A small fraction of APP is localized to lipid rafts as well, but the majority is in other regions of the membrane. Increasing amounts of intracellular cholesterol would thus favor processing through the amyloidogenic pathway, via lipid rafts. Another possibility is that since the cleavage site of the γ-secretase is inside the membrane, it could therefore be sensitive to a change in membrane

(26)

constitution. As cholesterol levels increase, the result could be a shift from Aβ(1-40) to Aβ(1-42)-production.

oxIdATIve STreSS

A few prospective studies have been conducted in order to determine whether or not dietary intake of antioxidants can lower the risk of AD [60]. These studies show conflicting results, probably caused by variants in study design. Some studies have not been able to find a link between AD and intake of antioxidants [61, 62], while others have confirmed a correlation between lowered risk of AD and intake of antioxidants in general [63] or vitamin E in particular [64]. These correlations were only true for antioxidants from food, not from supplements.

In the study by Morris et al. [64] vitamin E could only lower the risk in persons lacking ApoE4, which suggest isoform-specific effects of apoE on processes of oxidative stress. Because of the conflicting results in these studies, it is today unclear whether antioxidants can actually affect the risk of AD. Important questions that need to be asked are when and by which mechanisms antioxidants have an effect. To answer these questions the molecular basis of oxidative stress in AD needs to be clarified.

reactive oxygen species and free radicals can cause damage to the cell

Oxidative stress is a state that appears when excessive quantities of reactive oxygen species (ROS) are formed. ROS are intermediate oxygen species that easily react with the macromolecules of the cell, causing a chain reaction of free radical formation. Free radicals are low weight molecules or macromolecules with an unpaired electron that are very harmful to the cell, because of their ability to react with, and modify, all kinds of macromolecules [65]. ROS are primarily formed when ATP is synthesized by reducing oxygen to water in the respiratory chain of the mitochondria. In the brain the major neurotransmitters are the excitatory neurotransmitter glutamate and other excitatory amino acids, which are released from approximately 40% of the synapses. These neurotransmitters will produce ROS when metabolized and can thus be a large source of ROS through-out life, that could eventually lead to excitotoxicity [66].

The most important targets of ROS damage are nucleic acids (DNA/RNA- damage), lipids (lipid peroxidation) and proteins (protein oxidation). In the presence of reduced transition metals (mainly Fe²⁺ and Cu⁺), hydrogen peroxide (H₂O₂) that is one of the major forms of ROS, is reduced to the highly reactive hydroxyl radical (OH^•). Since there is no enzymatic defence system against hydroxyl radicals, the only way cells can protect themselves is to minimize the the formation of these radicals [65, 67].

Free radicals were first associated with aging in 1954 when Harman presented the free radical theory of aging. It postulates that changes associated with aging are a consequence of free radical mediated reactions, which are predominantly

(27)

initiated by the mitochondria, at a rate that is increasing with age. As aging progresses, an unbalance between the production of ROS and the cellular mechanisms that counteract them leads to accumulation of ROS that result in damage to major components of the cell: mitochondrial DNA, nucleus, membranes and cytoplasmic proteins. The life span is determined by the rate of these cellular damages [68, 69].

endogenous antioxidant systems

Oxidative stress is a physiological process and a majority of the free radicals are produced in the mitochondria through the respiratory chain. To be able to handle excess production of ROS the cell has developed several antioxidant systems, both enzymes (catalase, superoxide dismutase, peroxidases and several sulfur-containing enzymes like thioredoxin and glutaredoxin) and low molecular weight compounds (glutathione, NADPH) [65].

Interestingly, activated microglia, that are also important in AD pathology, are a large source of free radicals (e.g., superoxide and nitric oxide). To protect themselves from oxidative stress, that could impair the ability of the brain to defend itself from invading microorganisms and debris from damaged cells, the microglia have high levels of glutathione, superoxide dismutase, catalase and glutathione peroxidase [70].

Superoxide dismutase, catalase and glutathione peroxidase

Superoxide dismutase (SOD) exists in three different isoforms in brain. SOD1 (Cu, Zn SOD) is localized to the cytosol, lysomes and the mitochondrial

H₂O₂ O₂ + e^- H₂O₂

Mn-SOD

O₂^{• -}^Cu,Zn-SOD OH^•

Fe²⁺, Cu⁺

Lipid peroxidation

O₂^{• -}

Damage to DNA and protein

H₂O

CAT

GSH-peroxidase

Mitochondria

Cell m embra

ne

Nucleus free radical

formation

Figure 5 • Endogenous axtioxidant systems. Free radicals and ROS are primarily formed during ATP synthesis in the mitochondria. The cell has developed several antioxidant systems to protect against oxidative damage, see text for details. CAT, catalase; GSH, glutathione; O₂^•-, superoxide ion; OH^•, hydroxyl radical; SOD, superoxide dismutase.

(28)

intramembrane space. SOD2 (MnSOD) is in the mitochondrial matrix and SOD3 is an extracellular form. The function of SOD is to convert superoxide into hydrogen peroxide and oxygen. Hydrogen peroxide can freely cross membranes and has a potential to form hydroxyl radicals that the cell cannot take care of. To lower the risk of hydroxyl radical formation the cell can transform hydrogen peroxide into water.

The enzymes responsible for

this reaction are catalase and glutathione peroxidases. Catalase is localized to peroxisomes, but glutathione peroxidases are more ubiquitous, localized to the cytosol. Glutathione peroxidase is also able to reduce other peroxides (e.g., lipid peroxides in cell membranes) to alcohols. There are four different isoforms, but only two of them are widely expressed in most tissues (glutathione peroxidase 1 and 4) [65, 71].

Glutathione

Glutathione (γ-glutamyl-cysteinylglycine) is the major regulator of intracellular redox status and as such has an important role in the antioxidant cellular defence. Glutathione is a tripeptide thiol consisting of glutamic acid, cysteine and glycine and it is produced in all organs, in 1-10 mM concentration. In cells, total glutathione can be free or bound to proteins. Free glutathione is mainly present in its reduced form (GSH), but this can be converted into oxidized glutathione (GSSG) during oxidative conditions. GSSG is reduced by glutathione reductase in a NADPH-dependent manner. [65, 66, 72]. GSH has several physiological functions. It is the main redox buffer maintaining the cellular reducing environment, which is essential for the activity of most enzymes and other cellular molecules and will also detoxify ROS and scavenge free radicals. GSH itself is resistant to spontaneous oxidation, in contrast to cysteine, one of three amino acids of GSH. This could be a reason for why GSH is the major store of non-protein cysteine and suitable as a cellular redox buffer. The rate limit for GSH synthesis is the availability of cysteine and neurons are dependent on astrocytes to provide a cysteine precursor to increase neuronal GSH synthesis, since neurons have a low capacity to import cysteine [72]. Since GSH-levels are lower in neurons compared to other cells it could be possible that other antioxidant systems are more important to neurons, e.g.

thioredoxins.

Important antioxidants in brain

✧ Glutathione (GSH)

✧ Superoxide dismutase

✧ Catalase

✧ Glutathione peroxidases

✧ Thioredoxin

✧ Glutaredoxin

(29)

Thioredoxin

Thioredoxin (Trx) is a small 12-kDa ubiquitous protein with a redox active dithiol (reduced) / disulfide (oxidized) in its active site sequence, -Cys-Gly- Pro-Cys- (Figure 6, 7)[73].

It operates together with NADPH and thioredoxin reductase (TrxR) as a protein disulfide reducing system [73] and plays a critical role in maintaining the intracellular proteins in their reduced active state, together with the glutaredoxin systems [74]. In addition to the two cysteines in the active site of the protein, human Trx contains three additional structural cysteine residues that can be oxidized to generate dimers or multimers [75, 76].

Thioredoxin is the other major antioxidant in the cell, present at levels of 0.1–2 µM and is essential to the cell, as shown by the embryonal lethality of Trx^-/^- mice [77]. The most important function for Trx as an antioxidant is to reduce thioredoxin peroxidases, enzymes capable of directly reducing peroxides; e.g.

H₂O₂ [65], but Trx can also work as a free radical scavenger by itself. In this way Trx can reduce proteins and restore the function of proteins that have been damaged by oxidative stress.

There is also a mitochondrial form of Trx, called Trx2 [78]. Trx2 constitute a specialized mitochondrial defence mechanism, since the mitochondria is exposed to high concentrations of ROS generated during aerobic respiration [79]. In the rat brain, it has been demonstrated that the highest Trx2 expression is detected in areas exhibiting

the most severe oxidative stress, either due to the toxic action or metabolism of neurotransmitters (e.g. excitatory amino acids, dopamine, NO) or high content of transition metals. These findings imply that thioredoxins could

(dithiol)

Trx

SH SH

(disulfide)

Trx

S S

N

C C69

C62

C73 C32 C35

Figure 7 • Trx1 structure.

Cartoon representation of the structure of fully reduced human thioredoxin-1 with cysteine residues shown as sticks. Cysteines are labeled. C32 and C35 are in the active site. Figure prepared from Protein Data Bank entries 1ERT, by using the program PyMOL (DeLano Scientific LLC).

Figure 6 • Trx1 dithiol/disulfide.

Oxidation of the cysteines in the active site changes the redox state from reduced to oxidized.

(30)

be especially important in neuronal defence, since the other important antioxidant GSH is mostly found in glia [79].

Trx is predominantly a cytosolic protein, but can translocate to the nucleus in response to stress conditions, e.g oxidative stress. Oxidative stress causes an increase in the expression of protective genes, which are regulated by redox-sensitive transcription factors, like NF-κB and AP-1. Thus, translocation of Trx into the nucleus provides a mechanism by which the DNA-binding activity of specific transcription factors can be activated by reducing critical disulfide bonds [80].

Glutaredoxin

Glutaredoxin (Grx) is a 12-kDa dithiol protein with the active site sequence Cys-Pro-Tyr-Cys (Figure 8). Its primary function is to maintain the reduced state of cysteines in cellular proteins. Grx has many similarities to Trx and most of the functions are the same, which can be explained by the fact that Grx structurally belongs to the same superfamily as Trx. GSH reduces oxidized Grx and Grx is capable of reducing protein-glutathionyl-mixed disulfides, proteins modified by binding to GSSG (protein-SSG), which are formed and accumulated during oxidative stress [65, 81, 82]. So far, two isoforms of human Grx have been found, Grx1 (cytosolic) and Grx2, that exist as two different splice variants (mitochondrial, cytosolic and nuclear) [83-85].

oxidative stress in Alzheimer’s disease

A number of factors act in unison to make the neurons particularly sensitive to ROS and oxidative stress: (1) The brain and neurons are highly dependent on oxygen and use 20% of the total body consumption of oxygen, although only constituting 2% of the body weight. (2) Excitatory neurotransmitters, e.g. glutamate, forms ROS when metabolized. (3) The neuronal cell membrane contains a high proportion of polyunsaturated fatty acids that are targets of lipid peroxidation, by hydroxyl radicals. (4) Some areas of the brain have a high

N

C C83

C79

C26

C23 C8

Figure 8 • Grx1 structure.

Cartoon representation of the structure of fully reduced human glutaredoxin-1 with cysteine residues shown as sticks. Cysteines are labeled.

Figure prepared from Protein Data Bank entries 1ERT, by using the program PyMOL (DeLano Scientific LLC).

From cholesterol to oxidative stress in Alzheimer’s disease