Journal of Neuroimmunology 240-241 (2011) 114–120

Contents lists available at SciVerse ScienceDirect

Journal of Neuroimmunology

j ourna l homepage: www.e lsev ie r .com/ locate / jneuro im

Galanin and α-MSH autoantibodies in cerebrospinal fluid of patients withAlzheimer's disease☆

Alfredo Costa a,⁎, Paola Bini a, Maria Hamze-Sinno b, Arrigo Moglia a, Diego Franciotta a, Elena Sinforiani a,Sabrina Ravaglia a, Christine Bole-Feysot b, Tomas Hökfelt c, Pierre Déchelotte b, Sergueï O. Fetissov b,⁎⁎a National Institute of Neurology IRCCS C. Mondino, University of Pavia, Italyb ADEN EA 4311 Laboratory, Institute of Medical Research and Innovation, IFR23, Rouen University, Francec Department of Neuroscience, Karolinska Institutet, Stockholm, Sweden

☆ Study funding: This study was supported by the Itala grant of Ricerca Corrente 2009 to the Institute of NeurItaly.⁎ Correspondence to: A. Costa, National Institute of

University of Pavia, Via Mondino, 2, Pavia 27100, Italy.⁎⁎ Correspondence to: S.O. Fetissov, ADEN, Faculté deGambetta, Rouen, 76183 Cedex 1, France. Tel.: +33 2 35

E-mail addresses: alfredo.costa@mondino.it (A. CostSerguei.Fetissov@univ-rouen.fr (S.O. Fetissov).

0165-5728/$ – see front matter © 2011 Elsevier B.V. Aldoi:10.1016/j.jneuroim.2011.10.003

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 27 December 2010Received in revised form 13 September 2011Accepted 12 October 2011

Keywords:Alzheimer's diseaseα-MSHAutoantibodiesCSFGalaninNeurodegeneration

Background: Neuropeptides galanin and α-melanocyte-stimulating hormone (α-MSH) are involved in theregulation of memory and appetite. Increased galanin and decreased α-MSH levels were reported in post-mortem brains of patients with Alzheimer's disease (AD) but the underlying mechanisms are uncertain.Here we studied if autoantibodies (autoAbs) reacting with galanin and α-MSH are altered in AD.Methods: Levels of free and total IgG autoAbs reacting with galanin and α-MSH were measured in sera andcerebrospinal fluid (CSF) of 18 subjects with AD and in 15 age-matched non-demented controls. Valueswere correlated with Mini-Mental State Examination (MMSE) score, body mass index (BMI) and CSF levelsof AD biomarkers.Results: CSF levels of total but not free IgG autoAbs against galanin were increased in AD, resulting inincreased percentage of galanin autoAbs present as immune complexes. CSF levels of galanin total autoAbsand α-MSH free autoAbs correlated negatively with the severity of cognitive impairment as measured by

MMSE. Both total and free autoAbs against galanin and α-MSH in CSF correlated negatively with age in ADpatients but not in controls. CSF levels of galanin autoAbs and free α-MSH AutoAbs negatively correlatedwith CSF levels of t-Tau, p-Tau and ratios of t-Tau/Aβ42 or p-Tau/Aβ42 in AD patients but not in controls.Conclusions: AutoAbs reacting with galanin and α-MSH are present in CSF and are associated with clinicalcharacteristics of AD patients. The functional significance and therapeutic potential of these autoAbs shouldbe further clarified.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Galanin is a biologically active 29 amino acid (30 in humans)neuropeptide (Tatemoto et al., 1983) expressed in mammaliancentral nervous system (CNS), where it modulates several classicalneurotransmitters and has been implicated in the control of feed-ing, pain, seizure threshold, addiction and mood (Hökfelt, 2005).Moreover, galanin is involved in learning andmemorymainly via mod-ulation of the activity of cholinergic basal forebrain (CBF) neurons(Fisone et al., 1987; Dutar et al., 1989; Ögren et al., 2010) that providethe main cholinergic innervation of the cortex and hippocampal

ian Ministry of Health, throughology IRCCS C. Mondino, Pavia,

Neurology IRCCS C. Mondino,

Médecine-Pharmacie, 22, Bld148455; fax: +33 2 35148226.a),

l rights reserved.

formation (HiFo) (Mesulam et al., 1983). Several lines of evidence sug-gest involvement of galanin in the pathogenesis of Alzheimer's disease(AD) (Counts et al., 2008; Crawley, 2008). Interestingly, in later stagesof AD, nerve fibers containing GAL hyperinnervate CBF neurons(Chan-Palay, 1988; Mufson et al., 1993), and elevated levels of galaninin cerebrospinal fluid (CSF) have been observed (Edvinsson et al.,1993). The reasons for the alteration of the galanin system in AD andits functional significance are uncertain.

α-Melanocyte-stimulating hormone (α-MSH) is a 13 amino acidneuropeptide (Harris and Lerner, 1957) produced in the brain mainlyin the arcuate nucleus of the hypothalamus and abundantly presentin afferents to the basal forebrain (Jacobowitz and O'Donohue,1978). α-MSH is known to stimulate satiety and stress-associatedmemory (Wilson and Morgan, 1980; Fan et al., 1997). In contrast togalanin, decreased brain concentrations of α-MSH were reported inAD (Arai et al., 1986; Rainero et al., 1988).

Recent identification of autoantibodies (autoAbs) directed againstappetite-regulating neuropeptides, including galanin and α-MSH, inhealthy subjects has suggested that such autoAbs may represent amechanism underlying the variability in peptidergic signaling (Fetissov

115A. Costa et al. / Journal of Neuroimmunology 240-241 (2011) 114–120

et al., 2008). A functional role of autoAbs reactive with α-MSH in theregulation of appetite and anxiety was shown in rats (Hamze Sinno etal., 2009), while alterations of α-MSH autoAbs levels were associatedwith eating disorders (Fetissov et al., 2005). Altered production ofautoAbs reactive with galanin and α-MSH may also be implicated inpeptidergic dysregulation relevant to AD. We tested this hypothesis inthe present study by measuring levels of galanin and α-MSH autoAbsin serum and CSF of AD patients and of age-matched, non-dementedcontrols. Neuropeptide autoAbs levels were also analyzed for possiblecorrelations with several clinical parameters, as well as with CSF levelsof total Tau (t-Tau), 181-phopsphorylated tau (p-Tau) and amyloidbeta-42 (Aβ42), which are accepted biomarkers of AD (Blennow andZetterberg, 2009).

2. Methods

2.1. Study subjects

After approval of the local Ethical Committee, informed consentwas obtained from each patient and/or caregiver. We includedpatients consecutively seen at our Institute with a diagnosis of prob-able AD, according to the NINCDS–ADRDA criteria (McKhann et al.,1984). All patients underwent a thorough medical history recording,clinical examination, extensive laboratory investigation, lumbarpuncture for CSF analysis, venous blood sample, electrocardiography,electroencephalography and a basal magnetic resonance imaging(MRI) scan. The Mini-Mental State Examination (MMSE) was usedas a measure of global cognitive function (Folstein et al., 1975). Thestudy group consisted of 18 AD patients (7 males and 11 females,Table 1), and 15 sex-matched controls (6 males and 9 females).Control subjects were patients presenting with neurological com-plaints but preserved cognitive function, as also assessed by neuro-psychological examination. The presence of psychiatric disorders,alcoholism, other degenerative and non-degenerative forms of de-mentia, previous head trauma, associated CNS disease, CNS infections,severe systemic diseases and medical conditions possibly influencinglevels of galanin and α-MSH (i.e. eating disorders, obesity, endocrinediseases) were all considered as exclusion criteria. At the time ofinclusion, no patient was on treatment with cholinesterase inhibitors.Bilateral hippocampal atrophy was observed in all AD cases, andin none of controls. Signs of mild chronic vascular leukoencephalopa-thy were observed in 5 AD patients and in 6 controls. Additionally,sera from 43±3 years old, non-demented male subjects (n=40)were studied for the presence of IgG autoAbs against galanin andα-MSH.

Table 1Clinical and neurobiological features of AD patients and controls. Values are expressedas mean±SD.

AD patients Controls

N (% females) 18 (55.5%) 15 (53.3%)Age (years) 70.6±7.5 70.2±7BMI 24.7±3.9 26.2±4.3MMSE 16.6±4.9 29.2±1.4% transfer of CSF albumin/serum albumin 0.6±0.2 0.5±0.2p-tau 181, (pg/ml) 64.6±39.5 35.9±15.5t-tau (pg/ml) 574.6±685.1 241.8±290.8Aß 1–42, (pg/ml) 348±128.7 700.1 ±168.1p-tau/t-tau 0.14±0.06 0.27±0.17p-tau/Aß 1–42 0.22±0.22 0.04±0.02t-tau/Aß 1–42 1.95 ±1.99 0.30±0.09Duration of disease (months) 20.8±11.6 -

2.2. CSF and blood sampling

Lumbar puncture (LP) was performed in all patients and controls, ataround 9.00 a.m., in sitting position, after an overnight fast, between theL3-L4 and L4-L5 intervertebral space, using a 25-gauge needle. CSF wascollected into polypropylene tubes. A routine venous blood sample wasobtained immediately after. Samples were centrifuged at 2000 g for10 min at 4 °C, and then immediately frozen and stored at −80 °Cuntil assay. . The percentage transfer of albumin (CSF albumin/serumalbumin)×100 was calculated as index of integrity of the blood–brainbarrier (BBB) (Thompson, 1988).

2.3. AD biomarkers assay

CSF levels of t-Tau and p-Tau phosphorylated at threonine-181were measured by enzyme-linked immunosorbent assay (ELISA),using a commercially available kit (Innotest, Innogenetics, Ghent,Belgium) with monoclonal antibodies recognizing both the entiremoiety and its fragments (Vanmechelen et al., 2000). CSF levels ofamyloid beta-42 (Aβ42) were measured by ELISA, using a commer-cially available kit (Innotest, Innogenetics). All values are expressedas pg/ml.

2.4. Galanin and α-MSH autoantibody assay

Serum and CSF levels of galanin and α-MSH IgG autoAbs weremeasured using ELISA blinded for the AD diagnosis. The galanin orα-MSH peptides (Bachem AG, Bubendorf, Switzerland) were coatedon Maxisorp plates (Nunc, Rochester, NY) using 100 μl and 2 μg/mlin 100 mM NaHCO3 buffer, pH 9.6, for 72 h at 4 °C. Plates werewashed (3x) in PBS with 0.05% Tween 20, pH 7.4, and incubated over-night at 4 °C with 100 μl of patients and control sera (1:200) or CSF(1:25) diluted in PBS to determine free autoAbs levels, or in dissocia-tive 3 M NaCl, 1.5 M glycine buffer, pH 8.9 to determine total autoAbslevels. Optimal dilutions were determined by prior serial dilutions.Plates were washed (3x) and incubated with 100 μl (1:2000) of rabbitanti human IgG antibodies conjugated with alkaline phosphatase(Sigma, St. Louis, MO) for 3 h at room temperature. Following wash-ing (3x), 100 μl of p-nitrophenyl phosphate solution was added asalkaline phosphatase substrate. After 40 min of incubation at roomtemperature, 3 N NaOH was added. The optical density (OD) wasdetermined at 405 nm using a microplate reader. Blank OD valuesresulting from the reading of plates without addition of human serawere subtracted from sample values. Determinations were done induplicate, with variation between duplicates less than 5%. Presence ofcirculating immune complexes (CIC) was estimated by ratios betweentotal and free autoAbs, and their percentage relatively to total level ofautoAbs was calculated as follows: 100−(OD free autoAbs/OD totalautoAbs)×100.

2.5. Statistical analysis

Statistical analysis was performed in GraphPad Prism 5.02 pro-gram (GraphPad Software Inc., San Diego, CA). Levels of autoAbs inAD patients and controls were compared using the Student's t-testor the Mann–Whitney (MW) test according to a normality test.Kruskal–Wallis and post hoc Dunn's tests were used to analyze differ-ences in autoAbs levels between two age groups. To correlate autoAbsvalues with CSF t-Tau, p-Tau and Aβ42 levels and other continuousvariables (age, disease duration, MMSE score, BMI), Pearson's orSpearman's correlation coefficients were calculated according to thenormality of the variable. Differences were considered significant ifpb0.05.

116 A. Costa et al. / Journal of Neuroimmunology 240-241 (2011) 114–120

3. Results

3.1. Serum and CSF galanin and α-MSH autoAbs

Meanvalues of thepercentage transfer of albuminwere similar in ADpatients and controls (Table 1), and within the reference values (up to0.7%, Thompson, 1988), supporting integrity of BBB in both groups.

CSF and serum levels of free IgG autoAbs against galanin were notdifferent between AD patients and controls (Fig. 1A,B). Althoughlevels of free IgG galanin autoAbs in the CSF were lower comparedto serum, they significantly correlated with serum levels of galaninIgG autoAbs (Pearson's r=0.39, p=0.01).

Total IgG autoAbs against galanin were well detectable in the CSF,and their levels were higher in AD patients (MW-test, p=0.03;Fig. 1C), while serum levels of total IgG galanin autoAbs did not differbetween groups. (Fig. 1D).

An increased percentage of galanin IgG autoAbs present as CIC wasfound in the CSF of AD patients vs. controls (mean±SD, 51±17% vs.37±17%, respectively, t-test, p=0.01; Fig. 1E). In the serum, thepercentage of galanin IgG CIC was similar (87.3±7.9% vs. 87.9±2.9%, respectively; Fig. 1F).

Levels of free and total IgG autoAbs against α-MSH as well as thepercentage of α-MSH IgG CIC in either CSF or serum were not signif-icantly different between AD and control groups (Fig. 2). Levels of freeα-MSH IgG autoAbs in the CSF correlated positively with their serumlevels (Pearson's r=0.6, p=0.0001).

To estimate age effect on formation of CIC, the percentage ofserum IgG CIC with galanin or withα-MSHwas compared in the com-

Fig. 1. CSF (A,C,E) and serum (B,D,F) levels of free (A,B) and total (C,D) galanin IgG autoAbs apb0.05, *t-test, pb0.05. Note that serum samples were 8 times more diluted than CSF.

bined group of AD patients and controls and with 43-year-old controlsubjects, revealing significant differences among groups (KW-test,pb0.0001; Fig. 3). The percentage of CIC with galanin autoAbs washigher in older than in younger subjects (87.6±6% vs. 80.9±7.1%,respectively, Dunn's test p=0.001); that of CIC with α-MSH autoAbswas not different between older and younger subjects after themultiplecomparison tests, but it was significant in a single comparison (72.2±23.7% vs. 62.9±21.8%, MW-test, p=0.004). In both age groups, ahigher percentage of IgG CIC with galanin than with α-MSH wasobserved (Dunn's test p=0.001).

3.2. Correlations of galanin and α-MSH autoAbs with clinical parameters

3.2.1. MMSEIn AD patients, the MMSE score correlated negatively with age

(Spearman's r=−0.42, p=0.04, 1-tail). CSF levels of total and com-plexed IgG galanin autoAbs correlated positively with the MMSE score(Spearman's r=0.53, p=0.03 and r=0.5, p=0.04, respectively,Fig. 4A). CSF levels of free IgGα-MSH autoAbs also correlated positivelywith the MMSE score (Spearman's r=0.5, p=0.04, Fig. 4B). Serumlevels of galanin or α-MSH autoAbs did not show significant correla-tions with MMSE.

3.2.2. Age and disease durationIn AD patients but not in control subjects, CSF levels of both free and

total IgG Abs directed against galanin or against α-MSH showed strongnegative correlations with age (Spearman's r for galanin: IgG free, r=−0.68, p=0.002, IgG total, r=−0.74, p=0.006; Spearman's r for α-

nd percentage of galanin IgG immune complexes in AD and control subjects. #MW test,

Fig. 2. CSF (A,C,E) and serum (B,D,F) levels of free (A,B) and total (C,D) α-MSH IgG autoAbs and percentage of galanin IgG immune complexes in AD and control subjects. Note thatserum samples were 8 times more diluted than CSF.

117A. Costa et al. / Journal of Neuroimmunology 240-241 (2011) 114–120

MSH: IgG free, r=−0.82, pb0.0001, IgG total, r=−0.54, p=0.02).Serum levels of galanin total IgG autoAbs showed a negative associationwith duration of disease (Spearman's r=−0.4, p=0.05, 1 tail).

3.2.3. BMIBoth CSF and serum levels of free IgG galanin Abs correlated pos-

itively with BMI (Pearson's r=0.6, pb0.01 in both cases) in ADpatients, as did CSF levels of α-MSH total IgG autoAbs (Sperman'sr=0.51, p=0.03).

Fig. 3. Percentage of α-MSH and galanin IgG autoAbs present as immune complexes inserum of subjects at two ages, 43 and 71 years old (y.o.). *** Dunn's test, pb0.001.

Fig. 4. Graphs illustrating significant positive correlations between the MMSE score andCSF levels of galanin total IgG autoAbs (A) and free α-MSH IgG autoAbs (B).

118 A. Costa et al. / Journal of Neuroimmunology 240-241 (2011) 114–120

3.2.4. t-Tau, p-Tau and Aβ42CSF t-Tau levels were significantly higher in AD patients than in

controls (574,6±685,1 pg/ml vs. 241.8±290.8 pg/ml, respectively,t-test, pb0.05), as were p-Tau levels (64,6±39,5 pg/ml vs. 35.9±15.5 pg/ml, respectively, t-test pb0.01). On the other hand, CSFconcentrations of Aβ42, were found to be significantly lower in ADpatients than in the control group (348±128,7 vs. 700.1±168.1,respectively, pb0.01).

With regard to ratios, in AD patients both the values of p-Tau/Aβ42 and t-Tau/Aβ42 were significantly higher than in controlsubjects (pb0.01 and pb0.05, respectively).

In AD patients, negative correlations were found between serumlevels of galanin total IgG autoAbs and t-Tau (Pearson's r=−0.54,pb0.05), t-Tau/Aβ42 ratios (Pearson's r=−0.62, pb0.05) and p-Tau/Aβ42 ratios (Pearson's r=−0.6, pb0.05), and between CSFlevels of both free and total galanin IgG autoAbs and t-Tau/Aβ42ratios (Spearman's r=−0.61, pb0.05 and r=−0.7, pb0.01,respectively) as well as p-Tau/Aβ42 ratios (Spearman's r=−0.67,pb0.01 and r=−0.54, pb0.05, respectively). No significant correla-tions between galanin autoAbs and AD biomarkers were found incontrols.

Negative correlations were found between serum levels of α-MSHfree IgG autoAbs and t-Tau (Spearman's r=−0.74, pb0.01) and p-Tau (Spearman's r=−0.76, pb0.001). t-Tau/Aβ42 ratios correlatednegatively with serum levels of both free and total α-MSH IgG autoAbs(Spearman's r=−0.64, pb0.05). CSF levels of freeα-MSH IgG autoAbscorrelated negatively with both p-Tau/Aβ42 and t-Tau/Aβ42 ratios(Spearman's r=−0.64 and r=−0.69, respectively, p=0.01 for both).The only significant correlation for α-MSH autoAbs and AD biomarkersin the control group was found between serum levels of α-MSH totalIgG and CSF levels of Aβ42 (Pearson's r=0.57, pb0.05).

4. Discussion

The main findings of the present study indicate that: i) levels ofgalanin autoAbs present as CIC are increased in CSF of AD patients; ii)severity of cognitive impairment and biomarkers of neurodegenerationin ADpatientswere associatedwith lower CSF levels of galanin total andα-MSH free autoAbs, and iii) aging was associated with decreasedCSF levels of both galanin and α-MSH autoAbs in AD patients butnot in controls. Interpretations of these findings should be cautiousdue to the limited knowledge on the role of neuropeptide autoAbsin general. Indeed, this is the first study on galanin autoAbs aftertheir initial identification in healthy subjects (Fetissov et al., 2008).The present results support our view that neuropeptide autoAbs repre-sent a physiological phenomenon and also extend the concept by show-ing that galanin IgG and α-MSH IgG are normally present in the CSF, atleast in old subjects. The presence of neuropeptide autoAbs in CSF ismost likely a result of their entrance from the general circulation, assupported by their lower CSF vs. serum levels which positively correlate.By separately measuring the levels of free and total galanin and α-MSHautoAbs we also evaluated the presence of CIC. Using this approach, wefound increased CSF levels of total and complexed galanin autoAbs inAD patients, in the presence of normal values of the percentage transferof albumin as index of BBB integrity. This increase may reflect twoprocesses, one physiological as an adaptive response to increasedpresence of antigen, e.g. the peptide galanin, and another one patholog-ical, related to autoAb-mediated cellular toxicity or to inhibition of auto-Abs' normal role. In discussing the putative role of neuropeptide-reactive autoAbs in the brain, however, a major limitation of our studyis related to the absence of assay of corresponding neuropeptides.Such data would have been of importance, to evaluate possible correla-tions between levels of autoAbs and neuropeptides, but presently werely only on published data regarding CSF concentrations of thesepeptides in AD.

4.1. Physiological role of neuropeptide autoantibodies

The physiological role of increased CSF levels of total and com-plexed galanin autoAbs can be explained in analogy to our recentfindings in rats. Increased serum levels of complex-forming vs.free α-MSH autoAbs were associated with increased inhibition ofα-MSH-mediated behavior (Hamze Sinno et al., 2009). Hence, in-creased total and complexed galanin autoAbs would suggest increasedinhibition of central galanin signaling in AD. Importantly, levels ofgalanin total and complexed IgG autoAbs correlated positively withtheMMSE score, implying that increased inhibition of galanin signalingshould be associatedwith a better cognitive performance, in agreementwith the known inhibitory effect of galanin in learning and memory(Ögren et al., 2010). Increased CSF levels of galanin autoAbs in ADmay, therefore, be a compensatory reaction to increased CSF galaninlevels. However, the finding of a decrease of CSF galanin autoAbs levelsassociated with age and increased severity of dementia suggests thatthis compensation may occur only in an early phase of AD.

4.2. Pathological role of neuropeptide autoantibodies

In AD, CBF neurons undergo relatively selective degeneration, andthis process correlates with clinical aspects, including disease dura-tion and degree of cognitive impairment (Whitehouse et al., 1982;Wilcock et al., 1982). Immune complexes with antigen may triggercell lysis following binding to the cell surface via activation of theclassical complement pathway. For the selective loss of a certain neu-ronal population it should contain some specific markers recognizedby the immune system. It is yet unknown whether CIC of galaninreactive autoAbs contain galanin or another galanin-like antigen. Ifgalanin is present,such CIC could potentially bind neurons expressinggalanin receptors (galanin-R), abundantly expressed in this brain areaof rat, although not in CBF neurons (Miller et al., 1997; Mitchell et al.,1999; Mennicken et al., 2002). This may eventually explain increasedoccupancy of galanin-R binding found in AD (McMillan et al., 2004).

In rat dorsal HiFo and dorsal cortical regions, the main sources ofgalanin are the noradrenergic neurons in the locus coeruleus (LC),which in rat express GalR1 and -R2 (O'Donnell et al., 1999) and inhumans GalR3 (LeMaitre et al., 2008). Moreover, LC neurons are affect-ed in AD (Chan-Palay and Asan, 1989). Thus, these neurons can also betargets for galanin CIC, similar to neurons in the ventral forebrain. How-ever, the present results do not support a pathological role of galaninautoAbs, and also in view of the overlap in the levels of galanin autoAbsbetween AD patients and controls, their contribution to neurodegenera-tion may be limited. Moreover, CIC of galanin IgGmay also contain anti-idiotypic autoAbs, representing another type of pathophysiologicalmechanism via blocking the physiological function of neuropeptideautoAbs (Deloumeau et al., 2010).

4.3. Neuropeptide autoantibodies relation to age and BMI

If a pathogenic role of galanin autoAbs on CBF neurons will beestablished, similar effects of CIC with other neuropeptide autoAbsmay occur, provided that targeted neurons express the correspondingneuropeptide receptors. Thus, alterations in neuropeptide autoAbsmay contribute to age-related neurodegeneration (Bouras et al.,2005). Galanin autoAbs may play a particular role in this mechanism,not only due to the galanin innervation of the basal forebrain andHiFo (Mufson et al., 1993), but also because serum levels of galaninCIC autoAbs are elevated in older subjects. In this regard, we showedthat serum levels of galanin CIC autoAbs are higher in 71- than in 43-year old subjects, which is also the case for α-MSH CIC autoAbs. Thepercentage of serum galanin CIC autoAbs was not different betweenAD patients and age-matched controls, suggesting that the age-relatedincrease of serum galanin CIC autoAbs is a natural phenomenon,

119A. Costa et al. / Journal of Neuroimmunology 240-241 (2011) 114–120

which may contribute to the well-established age-related risk factor ofAD. The reason for such an increase requires further investigation.

In contrast to complex-forming autoAbs, the role of the free frac-tion of neuropeptide autoAbs may be to transport the correspondingneuropeptides (Hamze Sinno et al., 2009). We found that CSF levels ofα-MSH free autoAbs correlated positively with the MMSE score. In-creased transportation of α-MSH by autoAbs would therefore resultin a better cognitive performance, in agreement with early data asso-ciating an α-MSH defect with the development of AD (Anderson,1986). The positive correlations found in our study between CSFlevels of α-MSH total autoAbs and BMI may reflect an inhibitoryrole of autoAbs on α-MSH-mediated satiety. In contrast to α-MSH,BMI correlated positively with galanin free autoAbs, suggesting theirparticipation as galanin transporters to stimulate positive energybalance (Crawley, 1999). BMI changes are important in AD patients,weight loss having been proposed as a phenotypic marker of demen-tia (Buchman et al., 2005). The present observations support a role forautoAbs reactive with galanin and α-MSH, which may lead not onlyto cognitive impairment but also to weight loss in AD.

4.4. Neuropeptide autoantibodies and AD biomarkers

CSF levels of t-Tau and its phosphorylated variant p-Tau arebiomarkers of AD in an early phase of disease (Blennow and Zetterberg,2009). In our study, serum galanin total autoAbs correlated negativelywith t-Tau, and serum α-MSH free autoAbs correlated negatively withboth t-Tau and p-Tau. These associations are opposite to those foundbetween CSF galanin and α-MSH autoAbs and the MMSE score, furthersupporting a link between serum neuropeptide autoAbs and centralfunctions which are deteriorating with increased levels of Tau proteins.This also indicates that lower production of both neuropeptide autoAbscoincides with a higher degree of neuronal damage and increasedseverity of AD.

We found no correlation between Aβ42, another biomarker of AD(Hampel et al., 2009) and galanin or α-MSH autoAbs in AD patients.Since combinations of several biomarkers may optimize accuracy ofAD diagnosis (Blennow and Zetterberg, 2009), we observed that in ADpatients, CSF levels of galanin andα-MSHautoAbs correlated negativelywith t-Tau/Aβ42 and p-Tau/Aβ42 ratios, further supporting the viewthat central deficiency in galanin and α-MSH autoAbs parallels theseverity of disease. Interestingly, the only significant correlation be-tween AD biomarkers and neuropeptide autoAbs in controls was thatbetween Aβ42 and serum levels of total α-MSH autoAbs.

In conclusion, the present study shows that CSF levels of autoAbsreactive with galanin and α-MSH correlate positively with the MMSEscore and negatively with AD biomarkers, and decline with age in ADpatients but not in control subjects, suggesting their protective role inthe process of cognitive impairment and neurodegeneration in AD.While data on CSF galanin and α-MSH peptides were not available inthis study, our results indicate that the presence of autoAbs againstgalanin and α-MSH in CSF may influence the corresponding neuro-peptide systems and be to some extent beneficial in AD. In addition,AD patients display elevated CSF levels of autoAbs reactive with galaninbut notwithα-MSH suggesting that this is not a result of disrupted BBB,as also supported by normal mean values of the percentage transfer ofalbumin. These findings, in conjunction with increased serum levels ofgalanin CIC autoAbs with age, warrant further investigations to eluci-date their possible role in age-related neurodegeneration.

Acknowledgments

This studywas supported by the ItalianMinistry of Health, through agrant of Ricerca Corrente 2009 to the Institute of Neurology IRCCS C.Mondino, Pavia, Italy. It received also support from EU INTERREG IVA2 Seas Program (7-003-FR_TC2N).

S.F. was supported by NARSAD Mental Health Research AssociationUSA, 2007 Independent Investigator Award.

References

Anderson, B., 1986. Is alpha-MSH deficiency the cause of Alzheimer's disease? Med.Hypotheses 19, 379–385.

Arai, H., Moroji, T., Kosaka, K., Iizuka, R., 1986. Extrahypophyseal distribution of[alpha]-melanocyte stimulating hormone ([alpha]-MSH)-like immunoreactivityin postmortem brains from normal subjects and Alzheimer-type dementiapatients. Brain Res. 377, 305–310.

Blennow, K., Zetterberg, H., 2009. Cerebrospinal fluid biomarkers for Alzheimer's dis-ease. J. Alzheimer's Dis. 18, 413–417.

Bouras, C., Riederer, B.M., Kövari, E., Hof, P.R., Giannakopoulos, P., 2005. Humoralimmunity in brain aging and Alzheimer's disease. Brain. Res. Brain Res. Rev. 48,477–487.

Buchman, A.S., Wilson, R.S., Bienias, J.L., Shah, R.C., Evans, D.A., Bennett, D.A., 2005.Change in body mass index and risk of incident Alzheimer disease. Neurology 65,892–897.

Chan-Palay, V., 1988. Galanin hyperinnervates surviving neurons of the human basalnucleus ofMeynert in dementias of Alzheimer's and Parkinson's disease: a hypothesisfor the role of galanin in accentuating cholinergic dysfunction in dementia. J. Comp.Neurol. 273, 543–557.

Chan-Palay, V., Asan, E., 1989. Alterations in catecholamine neurons of the locus coeruleusin senile dementia of the Alzheimer type and in Parkinson's diseasewith andwithoutdementia and depression. J. Comp. Neurol. 287, 373–392.

Counts, S., Perez, S., Mufson, E., 2008. Galanin in Alzheimer's disease: neuroinhibitoryor neuroprotective? Cell. Mol. Life Sci. 65, 1842–1853.

Crawley, J.N., 1999. The role of galanin in feeding behavior. Neuropeptides 33,369–375.

Crawley, J., 2008. Galanin impairs cognitive abilities in rodents: relevance to Alzheimer'sdisease. Cell. Mol. Life Sci. 65, 1836–1841.

Deloumeau, A., Bayard, S., Coquerel, Q., Déchelotte, P., Bole-Feysot, C., Carlander, B.,Cochen De Cock, V., Fetissov, S.O., Dauvilliers, Y., 2010. Increased immune com-plexes of hypocretin autoantibodies in narcolepsy. PLoS One 5 (10), e13320.

Dutar, P., Lamour, Y., Nicoll, R.A., 1989. Galanin blocks the slow cholinergic EPSP in CA1pyramidal neurons from ventral hippocampus. Eur. J. Pharmacol. 164, 355–360.

Edvinsson, L., Minthon, L., Ekman, R., Gustafson, L., 1993. Neuropeptides in cerebrospinalfluid of patients with Alzheimer's disease and dementia with frontotemporal lobedegeneration. Dementia 4, 167–171.

Fan, W., Boston, B.A., Kesterson, R.A., Hruby, V.J., Cone, R.D., 1997. Role of melanocorti-nergic neurons in feeding and the agouti obesity syndrome. Nature 385, 165–168.

Fetissov, S.O., Harro, J., Jaanisk,M., Järv, A., Podar, I., Allik, J., Nilsson, I., Sakthivel, P., Lefvert,A.K., Hökfelt, T., 2005. Autoantibodies against neuropeptides are associated with psy-chological traits in eating disorders. Proc. Natl. Acad. Sci. U. S. A. 102, 14865–14870.

Fetissov, S.O., Hamze Sinno, M., Coëffier, M., Bole-Feysot, C., Ducrotté, P., Hökfelt, T.,Déchelotte, P., 2008. Autoantibodies against appetite-regulating peptide hormonesand neuropeptides: putative modulation by gut microflora. Nutrition 24, 348–359.

Fisone, G.,Wu, C.F., Consolo, S., Nordström, O., Brynne, N., Bartfai, T.,Melander, T., Hökfelt, T.,1987. Galanin inhibits acetylcholine release in the ventral hippocampus of the rat: his-tochemical, autoradiographic, in vivo, and in vitro studies. Proc. Natl. Acad. Sci. U. S. A.84, 7339–7343.

Folstein, M.E., Folstein, S.E., McHugh, P.R., 1975. Mini-mental state. A practical methodfor grading the cognitive state of patients for the clinician. J. Psychiatry Res. 12,189–198.

Hampel, H., Blennow, K., Shaw, L.M., Hoessler, Y.C., Zetterberg, H., Trojanowski, J.Q.,2009. Total and phosphorylated tau protein as biological markers of Alzheimer'sdisease. Exp. Gerontol. 45, 30–40.

Hamze Sinno, M., Do Rego, J.C., Coëffier, M., Gilbert, D., Hökfelt, T., Déchelotte, P., 2009.Regulation of feeding and anxiety byα-MSH reactive autoantibodies. Psychoneuroen-docrinology 34 (1), 140–149.

Harris, J.I., Lerner, A.B., 1957. Amino-acid sequence of the alpha-melanocyte-stimulatinghormone. Nature 179, 1346–1347.

Hökfelt, T., 2005. Galanin and its receptors: introduction to the Third InternationalSymposium, San Diego, California, USA, 21–22 October 2004. Neuropeptides 39(3), 125–142.

Jacobowitz, D.M., O'Donohue, T.L., 1978. alpha-Melanocyte stimulating hormone:immunohistochemical identification and mapping in neurons of rat brain. Proc.Natl. Acad. Sci. U. S. A. 75, 6300–6304.

Le Maitre, E., Diaz-Heijtz, R., Palkovits, M., Hökfelt, T., 2008. The galanin system in thehuman locus coeruleus and dorsal raphe nucleus. J. Neural Transm. 115, 1724–1725.

McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., Stadlan, E.M., 1984.Clinical diagnosis of Alzheimer's disease: report of the NINCDS–ADRDA WorkGroup under the auspices of Department of Health and Human Services TaskForce on Alzheimer's Disease. Neurology 34, 939–944.

McMillan, P.J., Peskind, E., Raskind, M.A., Leverenz, J.B., 2004. Increased galanin recep-tor occupancy in Alzheimer's disease. Neurobiol. Aging 25, 1309–1314.

Mennicken, F., Hoffert, C., Pelletier, M., Ahmad, S., O'Donnell, D., 2002. Restricted distri-bution of galanin receptor 3 (GalR3) mRNA in the adult rat central nervous system.J. Chem. Neuroanat. 24, 257–268.

Mesulam, M.M., Mufson, E.J., Levey, A.I., Wainer, B.H., 1983. Cholinergic innervation ofcortex by the basal forebrain: cytochemistry and cortical connections of the septalarea, diagonal band nuclei, nucleus basalis (substantia innominata), and hypothala-mus in the rhesus monkey. J. Comp. Neurol. 214, 170–197.

120 A. Costa et al. / Journal of Neuroimmunology 240-241 (2011) 114–120

Miller, M.A., Kolb, P.E., Raskind, M.A., 1997. GALR1 galanin receptor mRNA is co-expressed by galanin neurons but not cholinergic neurons in the rat basal fore-brain. Mol. Brain Res. 52, 121–129.

Mitchell, V., Bouret, S., Howard, A.D., Beauvillain, J.C., 1999. Expression of the galanin recep-tor subtype Gal-R2 mRNA in the rat hypothalamus. J. Chem. Neuroanat. 16, 265–277.

Mufson, E.J., Cochran, E., Benzing,W., Kordower, J.H., 1993. Galaninergic innervation of thecholinergic vertical limb of the diagonal band (Ch2) and bed nucleus of the stria ter-minalis in aging, Alzheimer's disease and Down's syndrome. Dementia 4, 237–250.

O'Donnell, D., Ahmad, S., Wahlestedt, C., Walker, P., 1999. Expression of the novel gala-nin receptor subtype GALR2 in the adult rat CNS: distinct distribution from GALR1.J. Comp. Neurol. 409, 469–481.

Ögren, S.O., Kuteeva, E., Elvander-Tottie, E., Hökfelt, T., 2010. Neuropeptides in learningand memory processes with focus on galanin. Eur. J. Pharmacol. 626, 9–17.

Rainero, I., May, C., Kaye, J.A., Friedland, R.P., Rapoport, S.I., 1988. CSF alpha-MSH indementia of the Alzheimer type. Neurology 38, 1281–1284.

Tatemoto, K., Rökaeus, Å., Jörnvall, H., McDonald, T.J., Mutt, V., 1983. Galanin — a novelbiologically active peptide from porcine intestine. FEBS Lett. 164, 124–128.

Thompson, E.J., 1988. The CSF proteins: A Biochemical Approach. Elsevier, Amsterdam.Vanmechelen, E., Vanderstichele, H., Davidsson, P., Van Kerschaver, E., Van Der Perre, B.,

Sjögren, M., Andreasen, N., Blennow, K., 2000. Quantification of tau phosphorylatedat threonine 181 in human cerebrospinal fluid: a sandwich ELISA with a syntheticphosphopeptide for standardization. Neurosci. Lett. 285, 49–52.

Whitehouse, P.J., Price, D.L., Struble, R.G., Clark, A.W., Coyle, J.T., Delon, M.R., 1982.Alzheimer's disease and senile dementia: loss of neurons in the basal forebrain.Science 215, 1237–1239.

Wilcock, G.K., Esiri, M.M., Bowen, D.M., Smith, C.C., 1982. Alzheimer's disease. Correla-tion of cortical choline acetyltransferase activity with the severity of dementia andhistological abnormalities. J. Neurol. Sci. 57, 407–417.

Wilson, J.F., Morgan, M.A., 1980. Plasma concentrations of alpha-melanotropin in therat during the acquisition and extinction of conditioned avoidance behaviour andduring the acquisition of maze learning behaviour. Psychopharmacology (Berl)68, 67–72.