Increased CSF-BACE1 Activity Associated with Decreased ... - IOS Press

Journal of Alzheimer’s Disease 25 (2011) 373–381 DOI 10.3233/JAD-2011-091153 IOS Press

373

Increased CSF-BACE1 Activity Associated with Decreased Hippocampus Volume in Alzheimer’s Disease Michael Ewersa,b,c,∗ , Xin Chengd,e,# , Zhenyu Zhongd,# , Hikmet F. Nurald , Cathal Walshf , Thomas Meindlg , Stefan J. Teipelh,i , Katharina Buergerb , Ping Hed,j , Yong Shend,j and Harald Hampelk a Discipline

of Psychiatry, School of Medicine & Trinity College Institute of Neuroscience (TCIN), Laboratory of Neuroimaging & Biomarker Research, Trinity College, University of Dublin, The Adelaide and Meath Hospital Incorporating The National Children’s Hospital (AMiNCH), Tallaght, Dublin, Ireland & Dementia Research Section and Memory Clinic, Dublin, Ireland b Alzheimer Memorial Center, Department of Psychiatry, Ludwig Maximilian University, Munich, Germany c Department of Radiology, University of California, San Francisco, CA, USA d Haldeman Laboratory of Molecular and Cellular Neurobiology, Sun Health Research Institute, Sun City, AZ, USA e Department of Neurology, Institute of Neurology, Huashan Hospital, Fudan University Shanghai Medical College, Shanghai, China f Department of Statistics, Trinity College, University of Dublin, Dublin, Ireland g Department of Radiology, Ludwig Maximilian University, Munich, Germany h Department of Psychiatry, University of Rostock, Rostock, Germany i Deutsches Zentrum für Neurodegenerative Erkrankungen (DZNE), Germany j Center for Advanced Therapeutic Strategies for Brain Disorders, Roskamp Institute, Sarasota, FL, USA k Department of Psychiatry, Psychosomatic Medicine & Psychotherapy, Goethe University, Frankfurt, Germany

Accepted 8 March 2010

Abstract. The enzyme ␤-secretase (BACE1) is essentially involved in the production of cerebral amyloidogenic pathology in Alzheimer’s disease (AD). The measurement of BACE1 activity in cerebrospinal fluid (CSF) has been reported, which may render CSF measurement of BACE1 a potential biomarker candidate of AD. In order to investigate whether BACE1 protein activity is correlated with regional brain atrophy in AD, we investigated the association between CSF levels of BACE1 and MRI-assessed hippocampus volume in patients with AD (n = 30). An increase in CSF-BACE1 activity was associated with decreased left and right hippocampus volume corrected for global head volume in the AD patients. Boot-strapped regression analysis showed that increased CSF levels of BACE1 activity were associated with increased CSF concentration of total tau but not amyloid-␤1-42 in AD. White matter hyperintensities did not influence the results. BACE1 activity and protein levels were

# Note:

Both authors contributed equally to the study. to: Michael Ewers, Ph.D., University of California, San Francisco, Department of Radiology, VA Medical Center, Center for Neuroimaging of Neurodegenerative Diseases, San Francisco, CA 94121; Tel.:415-221-4810 x3831; fax: 415 668 2864. E-mail: [email protected]; [email protected]. ∗ Correspondence

ISSN 1387-2877/11/$27.50 © 2011 – IOS Press and the authors. All rights reserved

374

M. Ewers et al. / CSF-BACE1 and Hippocampus

significantly increased in AD compared to 19 elderly healthy controls. Thus, the CSF biomarker candidate of BACE1 activity was associated with hippocampus atrophy in AD in a robust manner and may reflect neurotoxic amyloid-␤-related processes. Keywords: BACE1, ␤-secretase, cerebrospinal fluid, hippocampus

INTRODUCTION Alzheimer’s disease (AD) is a neurodegenerative disorder leading to global cognitive decline. AD is histochemically characterized by two major brain lesions including the deposition of the amyloid-␤ peptide (A␤1-42 ) as well as neurofibrillary changes including neurofibrillary tangles and alterations of the neuropil. The enzyme ␤-secretase (BACE1) is an enzyme that is essentially involved in amyloid-␤ protein precursor (A␤PP) cleavage that leads to the synthesis of A␤1-42 [1]. BACE1 activity and protein concentration were found upregulated in the postmortem brain of AD patients and related to neuronal death [2], suggesting that BACE1 may reflect or contribute to brain atrophy in AD [2]. BACE1 activity and protein levels were found to be increased due to oxidative stress [3] and after ischemic lesions, and increased apoptosis as shown by us and others [4, 5]. Soluble A␤ may form downstream of the A␤ dimers and oligomers, which have been shown to lead to hippocampus synaptic loss [6] and disruption of hippocampus neuronal function [7]. Increased BACE1 activity in the brain may initiate or trigger such A␤-related pathological cascades and neuronal degeneration [8], eventually leading to hippocampus volume reduction in AD [9]. Recently, cerebrospinal fluid (CSF)-based assays allowing for the in vivo measurement of BACE1 were developed [10–12]. CSF levels of BACE1 were previously found to be changed in mild cognitive impairment (MCI) and clinically manifest AD [10, 11, 13]. Thus, the CSF measure of BACE1 is a promising new candidate as a biomarker of early AD pathology. Well established biomarker candidates including CSF levels of total tau, phospho tau, or A␤1-42 have previously been found to be associated with hippocampus volume in AD [14]. Here we hypothesize that increased levels of BACE1 activity are associated with MRI-assessed hippocampus volume as a surrogate measure of neurodegeneration in AD. The hippocampus volume is one of the best established MRI-based biomarkers of brain atrophy in AD [15–17]. We also examined the association between CSF levels of BACE1 activity and CSF levels of A␤1-42 and total tau. Total tau in CSF has been found to be associated with the amount of cerebral neu-

rofibrillary tangles within the brain and is thought to reflect neurodegeneration in the brain. Thus, a correlation between CSF-tau and BACE1 may further support the hypothesis of an association between BACE1 and neurodegeneration in the brain. Previous studies have shown that the increased development of A␤1-42 is an early event within the cascade of AD pathology [18]. Consistent with this hypothesis, the elevation of BACE1 activity was already observed in early stages of AD [10, 13]. Similarly, neuronal loss is especially observed within the medial temporal lobe in the early stages of AD before it becomes more widespread [15, 19]. Thus, in the current study, we have focused on the assessment of the association between CSF-BACE1 activity and hippocampal volume in the mild stage of AD. MATERIALS AND METHODS Patients CSF samples were obtained from a total of 30 patients diagnosed with probable AD according to the NINCDS-ADRDA criteria (mean Mini Mental Status Examination (MMSE) = 24.2, SD = 3.8) and 19 cognitively normal controls (NC; mean MMSE = 28.7, SD = 1.2), recruited at the memory clinic of the Department of Psychiatry, University of Munich, Germany. The NC performed within 1 SD of the age-adjusted norm of the CERAD test battery and no subjective memory impairment was reported. The NC subjects included a clinical control group who underwent either spinal anesthesia for the primary purpose of surgery of the urinary tract or lower extremities or were diagnosed with neuropathies. A comparison of the CSF concentration of tau and A␤1-42 between the current sample of healthy controls and normative values of elderly healthy controls [20] showed no study differences (see Results section below), suggesting that the current control CSF values are representative for elderly normal subjects of about 65 years of age. It is still important to keep in mind that these are not physically healthy control subjects and the results may not extrapolate necessarily to BACE1 activity for which we do not have normative values.


Note that handedness was not assessed in the current study, even though it may potentially introduce a laterality bias in hippocampus volume. However, since handedness is unlikely to be correlated with BACE1 activity, this was not deemed a confounding factor in the current study. All subjects were included in the study after written consent. The study was approved by the institutional review board (IRB) of the University of Munich. MRI volume measures All patients were scanned on a 1.5 T scanner (Magnetom Vision, Siemens, Erlangen), using a gradient echo sequence. Hippocampus volume was obtained via manual volumetry according to a standardized protocol [21], with an intraclass correlation coefficient of >0.9. Prior to volumetric measurements, all MR volumes were corrected for image intensity non-uniformities, mapped by affine stereotaxic transformation into coordinates based on the Talairach Atlas, resampled onto a 1 mm voxel-grid, and each second scan was flipped in the left-right orientation. These preprocessing steps reduce inter-scanner variability due to scan artifacts, correct for effects of whole brain atrophy, and preclude rater-bias for the left or the right hemisphere. We calculated the determinant of the 12-parameter affine transformation matrix which represents the overall volume effect associated with the transformation of the brain into standard space. The measurements of hippocampus volume obtained after global transformation that accounted for differences in head shape (and global brain volume) were used for subsequent analysis. BACE1 protein and activity in CSF The BACE1 assays are based on our recently reported methods [10] with minor modification. Briefly, two BACE1 protein sandwich-ELISAs were established: one used a combination of anti-BACE1 polyclonal antibody SECB2 as a capture antibody and biotinylated anti-BACE1 polyclonal antibody SECB1 as a detection antibody. The other ELISA was established by using anti-BACE1 polyclonal antibody B280 as a capture antibody and anti-BACE1 monoclonal antibody (R&D) as a detection antibody. For BACE activity, assays were performed by using synthetic peptide substrates containing the BACE1 cleavage site (MCA-Glu-Val-Lys-Val-AspAla-Glu-Phe-(Lys-DNP)-OH) at a 50 mM reaction buffer (50 mM acetic acid pH4.1, 100 mM NaCl). Flu-

375

orescence was observed by a fluorescent microplate reader as described previously [10]. Total tau and Aβ1-42 in CSF Total tau and A␤1-42 were measured with the commercially available immunoassays INNOTEST® *hTau Ag and INNOTEST® ␤-Amyloid (1-42) (Innogenetics, Belgium) [22]. White matter rating Regional white matter rating within the temporal lobe and basal ganglia was conducted by an experienced radiologist (TM) according to a standardized regional white matter rating scale, which ranges from 0 (no white matter lesions) to 3 (diffuse involvement of entire region or, in case of the basal ganglia, confluent lesions) [23]. Statistics In a first step, all variables of interest including BACE1 activity, BACE1 protein, and the unilateral hippocampus volumes were tested for outliers in AD patients. Two extreme outliers that lay more than three times the interquartile range (IQR) to the left and right from the first and third quartiles were identified for the left hippocampus (volume = 2179 mm2 and 2209 mm2 ) and excluded from further analysis. To attain normal distribution of the data, the CSF levels of total tau and A␤1-42 were transformed using the natural logarithm. Linear regression models including the CSF concentrations of BACE1 activity and BACE1 protein, and the nuisance variables age, MMSE, and gender, were built to predict right and left normalized hippocampus volume. Binarized apolipoprotein E (ApoE) genotype (n = 17 ApoE ␧4 non-carriers without any ␧4 allele and n = 11 ApoE ␧4 allele carriers) were entered as an additional predictor. In order to examine the robustness of our findings, we also resampled the data 1000 times, using bootstrapping. This ensures that the findings are not an artifact of small sample size or bias that could be caused by unusual individual observations. The 95% confidence interval estimates (95% CI) of the predictors of left and right hippocampus volume are provided. ANCOVA-based group comparison in CSF measures between NC and AD were controlled for age and gender. Bonferroni correction was applied with ␣ = 0.05. Because the results on the correlation between BACE1 activity and total tau in CSF of AD patients vary between studies, a meta-analysis of the

376


Pearson moment correlation of these two measures of the current and previous studies [10, 13] was computed. To this end, the Spearman correlation coefficients r were first transformed into a standard normal metric Zri = 1/2loge ((1 + ri )/(1 − ri )). The Z values were weighted by the sample size with the factor wi = n – 3 ¯r = tocompute theweighted mean Z value with Z (( ki=1 wi zi )/( ki=1 wi )). The standard error of the ¯ r ) = 1/ ki=1 wi Z values was calculated with SE(Z and the standard z score of the mean Z value was computed by z = z/SE to determine the probability value within the standard normal distribution [24]. RESULTS Since the controls were cognitively normal but physically not healthy, it is possible that these subjects had conditions that may have influenced the concentration of CSF biomarkers of AD pathology. In order to test the representativeness of the current control group, we compared the CSF total tau and A␤1-42 concentration to previously published normative CSF concentrations in healthy elderly control subjects without any medical, neurological, or psychiatric disorder [20]. Both the CSF total tau and A␤1-42 concentration in the current control group (Table 1) were within 1 SD of the concentration of the reference group of healthy control subjects. The total tau concentration was similar in the current group (mean = 249.8 ng/ml, SD = 136) compared to the reference healthy control subjects who were between 51 and 70 years old (mean 243.6 ng/ml, SD = 127) [20]. The mean CSF concentration of A␤1-42 was 753.9 ng/ml (SD = 201) in the current control group and 790.6 ng/ml (SD = 228) in the reference group, thus again showing similar and statistical non-significantly different mean values. Association between CSF measures and hippocampus volume in AD Increased CSF concentration of BACE1 activity was associated with decreased hippocampus volume within

the left brain hemisphere (p = 0.02, B = –1137.9, 95% CI = (–2023.4, –252.3)) and the right brain hemisphere (p = 0.03, B = –1178.7, 95% CI = (–2171.6, –185.8), Fig. 1). Male AD patients had a larger hippocampus volume corrected for global head volume for the left (p = 0.01) and right hemisphere (p < 0.01) compared to female patients. The interaction term of BACE1 activity and gender, however, was not significant. Bootstrapping of the regression model confirmed BACE activity as a predictor of left hippocampus volume (95% CI = (–1955.1, –304.3)) and right hippocampus volume (95% CI = (–2210.5, –127)). Standard diagnostics were run after fitting the regression in order to ensure the validity of modeling assumptions for this (moderately sized) data set. These included a plot of residuals against fitted values, a Scale-Location plot of square root of the absolute value of residuals against fitted values, a Normal Q-Q plot, a plot of residuals against leverages, and Cook’s distances. No abnormal deviations were evident having run these diagnostics. Pearson moment correlations showed that an increase in CSF levels of BACE1 activity was correlated with an increase in log-transformed CSF-tau (r = 0.4, p = 0.04, Fig. 2A) but not CSF levels of A␤1-42 (p = 0.8). In addition, an increase in log-transformed CSF-tau levels was correlated with decreased right hippocampus volume (r = –0.42, p = 0.03, Fig. 2B) but not left hippocampus volume (r = –0.31, p = 0.11). CSF levels of A␤1-42 were not associated with left or right hippocampus volume (p > 0.77). Whereas the absence of a correlation between CSF-BACE1 activity and A␤1-42 in AD is consistent with previous studies [13], the results for the association between CSF levels of BACE1 activity and total tau vary between studies, with some studies reporting a significant correlation (r = 0.7, p < 0.001) [13] and another a non-significant finding (r = –0.09, p = 0.47) [10]. In order to test whether the correlation between these two measures is significant when combined across different studies, a meta-analysis was done. The meta-analytical mean correlation coefficient transformed into a z score was 2.55 (p < 0.001).

Table 1 Mean and standard deviation (in brackets) of demographic and CSF biomarker concentration for AD and HC subjects Group AD HC

Sample size

Age

Gender (f/m)

MMSE

CSF BACE1 activity (pmol/min*␮l)

CSF BACE1 protein (ng/ml)

CSF total tau (ng/ml)

CSF A␤1-42

28 19

74.1 (8.7) 65.4 (7.5)

17/11 6/13

24.2 (3.8) 28.7 (1.2)

0.18 (0.08) 0.01 (0.003)

0.8 (0.58) 0.12 (0.12)

652.3 (462.7) 249.8 (136)

648.6 (331.1) 753.9 (201)

f = female, m = male (mean age in AD: 74.1 yrs, SD = 8.7, gender: 17 female, 11 males; for NC: mean age = 65.4 yrs, SD = 7.5, gender: 6 females, 13 males).


A

377

B

Fig. 1. Hippocampus volume for the left (A) and right (B) hemisphere as a function of level of BACE1 activity in CSF in AD patients.

A

B

Fig. 2. Regression plots for CSF BACE1 activity versus CSF total tau concentration (A) and right hippocampus volume versus CSF total tau concentration (B) are displayed. Higher CSF total tau concentration is associated with higher CSF BACE1 levels (A) but lower hippocampus volume (B).

Inﬂuence of comorbidities, white matter hyperintensities, and drug-treatment in AD The most frequent comorbidity was hypertension (n = 7) diagnosed according to the WHO guidelines in AD [25]. We tested whether this condition may have been a confounding factor in the current analysis. Results of a two-sampled t-test showed that the mean values of CSF-BACE1 levels as well as hippocampus volume were virtually the same between AD patients with hypertension and those without (p values for each comparison were >0.77). Four AD patients had depressive symptoms or manifest depression, but were within 1 standard deviation of

the group mean of left or right hippocampus volume or CSF-BACE1 measures of the non-depressed AD patients. One patient had type II diabetes and showed no significant divergence from the remainder of the group in terms of the variables of interest. A total of 10 AD patients were treated with acetylcholinesterase inhibitors including rivastigmine (n = 1), donepezil (n = 3), or galantamine (n = 1), or were included in a clinical trial including galantamine or memantine treatment (n = 5). We could not detect any differences in hippocampus volume (p > 0.13) or CSF-BACE measures (p > 0.52) between the treated and untreated groups by two-sample t-tests. Two patients were treated for depression but lay within one

378


standard deviation of the group mean of left or right hippocampus volume or CSF BACE1 measures of the patients not treated for depression. Rating of white matter hyperintensity according to a standardized regional white matter rating scale [23] showed for three AD patients a rating score of 1 within the left temporal lobe and for one subject within the right temporal lobe. For the basal ganglia, 2 out of 28 patients had a score of 1 and 2 other patients a score of 2. All other patients had a score of 0. The low frequency of lesions did not allow for a further assessment of the association between white matter lesions and CSF biomarker levels or hippocampus volume. Group differences in CSF biomarkers In comparison to NC, AD subjects showed significantly increased CSF levels of BACE1 activity (F(1,42) = 56.29, p = 0.001) and BACE1 protein (F(1,40) = 12.22, p = 0.001). CSF-tau levels were significantly increased in AD patients when compared to HC subjects (F(1,42) = 15, p < 0.001). CSF-levels of A␤1-42 were numerically but not significantly lower in AD compared to HC (F(1,42) = 3.44, p = 0.07). DISCUSSION The current study shows for the first time that BACE1 activity is associated with a decrease in hippocampus volume in mild to moderate AD. The findings were extensively examined to ensure statistical validity of the current findings. Bootstrapping analysis ensured that the results are not an artifact of a moderate sample size or bias that could be caused by unusual individual observations. Neither hippocampus volume nor CSF levels of BACE1 activity were correlated with white matter lesions, the most frequent comorbidities or treatment of AD, which supports the robustness of the current finding of the association between CSFBACE1 activity and hippocampus volume. Increased CSF levels of BACE1 may indicate increased BACE1 activity and A␤-related pathology in the brain. We found that CSF-BACE1 activity was positively correlated with CSF-total tau levels, which is thought to be a marker of neurodegeneration. Our meta-analysis on the correlation between these two measures across the current and previous studies [10, 13] shows that this association holds when taking between-study variation into account. We have previously shown that higher brain levels of BACE1 activity are associated with higher A␤ burden in the

brain [26]. Increased BACE1 activity in the brain may initiate or trigger A␤-related pathological cascades and neuronal degeneration [8] eventually leading to hippocampus volume reduction in AD [9]. Preclinical studies in bigenic 5XFAD mouse model of AD showed that BACE gene knockout is associated with a reduction in levels of A␤ in the brain, neuronal loss, and hippocampal-dependent memory performance [2]. One possible mechanism by which BACE1 may trigger neurotoxic processes may relate to the formation of A␤ oligomers. Human CSF-derived A␤ oligomers have been shown to be disrupt hippocampus long-term potentiation in vivo [27], and the formation of A␤ oligomers and ensuing long-term potentiation disruption could be prevented by blocking the amyloidogenic pathway of A␤PP processing by ␥-secretase inhibition [28]. Interestingly, A␤ oligomers, but not insoluble amyloid plaques, derived from the postmortem brain of AD patients showed a disruption of hippocampus long-term potentiation [29]. Increased BACE1 activity may be related to increased formation of A␤ oligomers and thus may exert synaptoxic effects [8]. Thus, increased levels of BACE1 in CSF may reflect such increased A␤ related neurotoxic processes. It should be noted that BACE1 is regulated by many factors. For example, BACE1 is upregulated by physiological stress that may be associated with neurodegeneration [30, 31]. BACE1 activity and protein levels were found to be increased due to oxidative stress [3] and after ischemic lesions and increased apoptosis [4, 5]. It has been further suggested that BACE1 may be involved in remyelination of the axons [32], although its role in AD is unclear. From that perspective, BACE1 activity may result from other neurodegenerative processes rather than playing the major role in triggering neurotoxic effects. It is currently unknown which factors lead to increased BACE1 activity in the brain and what the specific consequences are. The current results of increased CSF BACE1 activity in AD are consistent with the findings in the majority of previous studies in AD [11, 13, 33, 34]. In addition, we have previously found increased CSF-BACE1 activity levels in MCI [10], and BACE1 activity levels are higher in MCI patients who convert to AD than those who remain stable over a period of 4–6 years [13]. Divergent from this predominant result pattern are reports of no differences between controls and AD patients in BACE1 activity [10] or, as reported in a single study, even a decrease of CSF-BACE1 activity in AD patients [35]. It is currently unclear why BACE1 activity is sometimes not elevated in AD. Possible factors include differences in assays, sample


size, or patient characteristics such as age or disease severity [35]. In order to investigate the association of BACE1 activity with other core CSF biomarkers of AD pathology, a correlation between CSF-BACE1 activity and CSF-tau and A␤1-42 concentration was tested. While we found an increase in BACE1 activity within the CSF of AD patients, a trend toward reduced CSF-levels of A␤1-42 in AD compared to controls was observed. It should be noted that the absence of the statistical significance of the group difference, albeit the presence of a trend, should be interpreted in the context of the statistical power of the relatively small sample size in the current study. Multiple previous studies have shown that CSF levels of A␤1-42 are decreased in AD [36] and lower CSF concentration of A␤1-42 is predictive of increased risk of AD in subjects with MCI [37, 38]. BACE1 activity was correlated with CSF-tau but not CSF concentration of A␤1-42 . This result pattern is similar to previous findings. In the study by Zetterberg et al. [13], the BACE 1 activity measured in CSF was increased in AD patients when compared to healthy controls, but CSF-levels of A␤1-42 were decreased. BACE1 levels were not correlated with A␤1-42 in the CSF; in fact, the correlation coefficient ranged between negative and positive values among the different diagnostic groups tested. BACE1 activity was, however, positively correlated with ␣A␤PP and ␤A␤PP in the CSF [13]. Thus, in the current and previous study, no correlation between CSF-BACE1 activity and CSFA␤1-42 levels could be detected. It is possible that BACE1 activity levels and A␤1-42 are regulated by different factors. Lower antemortem CSF-A␤1-42 concentration is associated with higher A␤ burden in the postmortem brain [39] and higher PIB-PET binding [40]. This finding has led to the hypothesis that sequestration of fibrillar A␤1-42 into insoluble A␤ deposition in cerebral plaques accounts for the reduced levels of soluble A␤1-42 within the CSF. Once A␤ aggregates, it tends to become “sticky”, thus showing increased binding to cell membranes or other proteins, i.e., APOE, TNF receptors, etc., in the brain. Thus, the amount of A␤ that flows within the CSF is decreased. However, BACE1 is mainly secreted or cleaved from neurons (recently many studies reported that BACE1 is present in glial cells as well) [1, 41]. Elevated BACE1 in CSF can result from several sources: 1) neurons are damaged in AD brains, so that BACE1 can be “leaked” from broken or dead neurons; 2) activated glial cells may “spin off” BACE1 [41, 42]; 3) neurons in AD brain are hyperactive and activated neurons are able to make more BACE1; or 4)

379

increased inflammatory responses the AD brains may trigger increased BACE1 production [43, 44]. Therefore, the elevated BACE1 levels within the CSF in AD may be derived from multiple sources and may lead to increased levels in CSF, where the mechanisms of secretion are different for BACE1 and A␤. In conclusion, the current results demonstrate an association between CSF-based BACE1 activity and a decrease in hippocampus volume in AD. In addition, increased levels of BACE1 activity were associated with increased CSF-tau levels, suggesting that BACE1 activity is associated with neurodegenerative processes. These results encourage larger validation studies to provide further support for the notion that CSF-levels of BACE1 activity may show utility to assess in vivo core AD pathology in association with neurodegenerative processes, an important feature for the application of a biomarker to track AD disease processes such as needed in clinical trials on treatment of A␤ pathology. ACKNOWLEDGMENTS The authors thank W. Kretschmann, Y. Hoessler, B. Asam, F. Jancu, L. Jertila-Aqil, and C. Sänger for technical assistance. We thank also K. Duggan for editing the manuscript. The study was supported by a grant from the Federal Agency of Education and Research (Bundesministerium fuer Bildung und Forschung, BMBF 01 GI 0102) to the Competence Network of Dementia (to HH, SJT, and ME), grants from Adelaide and Meath Hospital incorporating the National Children’s Hospital (AMNCH) (to HH), the Health Service Executive (HSE) (to HH), Trinity College Dublin, Ireland (to HH), Science Foundation Ireland (08/IN.1/B1846, to HH), National Institute on Aging (NIHRO1AG025888) (YS), Alzheimer’s Association Zenith Award (YS, HH), and the Arizona Alzheimer Research Consortium (YS). Authors’ disclosures available online (http://www.jalz.com/disclosures/view.php?id=366). REFERENCES [1]

[2]

Vassar R, Bennett BD, Babu-Khan S, Kahn S, Mendiaz EA, Denis P, Teplow DB, Ross S, Amarante P, Loeloff R, Luo Y, Fisher S, Fuller J, Edenson S, Lile J, Jarosinski MA, Biere AL, Curran E, Burgess T, Louis JC, Collins F, Treanor J, Rogers G, Citron M (1999) Beta-secretase cleavage of Alzheimer’s amyloid precursor protein by the transmembrane aspartic protease BACE. Science 286, 735-741. Ohno M, Cole SL, Yasvoina M, Zhao J, Citron M, Berry R, Disterhoft JF, Vassar R (2007) BACE1 gene deletion pre-

380

[3]

[4]

[5]

[6]

[7]

[8] [9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

M. Ewers et al. / CSF-BACE1 and Hippocampus vents neuron loss and memory deficits in 5XFAD APP/PS1 transgenic mice. Neurobiol Dis 26, 134-145. Tamagno E, Bardini P, Obbili A, Vitali A, Borghi R, Zaccheo D, Pronzato MA, Danni O, Smith MA, Perry G, Tabaton M (2002) Oxidative stress increases expression and activity of BACE in NT2 neurons. Neurobiol Dis 10, 279-288. Tesco G, Koh YH, Kang EL, Cameron AN, Das S, SenaEsteves M, Hiltunen M, Yang SH, Zhong Z, Shen Y, Simpkins JW, Tanzi RE (2007) Depletion of GGA3 stabilizes BACE and enhances beta-secretase activity. Neuron 54, 721-737. Wen Y, Onyewuchi O, Yang S, Liu R, Simpkins JW (2004) Increased [beta]-secretase activity and expression in rats following transient cerebral ischemia. Brain Res 1009, 1-8. Shankar GM, Bloodgood BL, Townsend M, Walsh DM, Selkoe DJ, Sabatini BL (2007) Natural oligomers of the alzheimer amyloid-beta protein induce reversible synapse loss by modulating an NMDA-type glutamate receptor-dependent signaling pathway. J Neurosci 27, 2866-2875. Shankar GM, Li S, Mehta TH, Garcia-Munoz A, Shepardson NE, Smith I, Brett FM, Farrell MA, Rowan MJ, Lemere CA, Regan CM, Walsh DM, Sabatini BL, Selkoe DJ (2008) Amyloid-[beta] protein dimers isolated directly from Alzheimer’s brains impair synaptic plasticity and memory. Nat Med 14, 837-842. Walsh DM, Selkoe DJ (2007) A beta oligomers – a decade of discovery. J Neurochem 101, 1172-1184. Jack CR Jr., Petersen RC, Xu YC, O’Brien PC, Waring SC, Tangalos EG, Smith GE, Ivnik RJ, Thibodeau SN, Kokmen E (1998) Hippocampal atrophy and apolipoprotein E genotype are independently associated with Alzheimer’s disease. Ann Neurol 43, 303-310. Zhong Z, Ewers M, Teipel S, Burger K, Wallin A, Blennow K, He P, McAllister C, Hampel H, Shen Y (2007) Levels of beta-secretase (BACE1) in cerebrospinal fluid as a predictor of risk in mild cognitive impairment. Arch Gen Psychiatry 64, 718-726. Holsinger RM, McLean CA, Collins SJ, Masters CL, Evin G (2004) Increased beta-Secretase activity in cerebrospinal fluid of Alzheimer’s disease subjects. Ann Neurol 55, 898-899. Verheijen JH, Huisman LGM, van Lent N, Neumann U, Paganetti P, Hack CE, Bouwman F, Lindeman J, Bollen ELEM, Hanemaaijer R (2006) Detection of a soluble form of BACE-1 in human cerebrospinal fluid by a sensitive activity assay. Clin Chem 52, 1168-1174. Zetterberg H, Andreasson U, Hansson O, Wu G, Sankaranarayanan S, Andersson ME, Buchhave P, Londos E, Umek RM, Minthon L, Simon AJ, Blennow K (2008) Elevated cerebrospinal fluid BACE1 activity in incipient Alzheimer disease. Arch Neurol 65, 1102-1107. Herukka SK, Pennanen C, Soininen H, Pirttila T (2008) CSF Abeta42, tau and phosphorylated tau correlate with medial temporal lobe atrophy. J Alzheimers Dis 14, 51-57. Jack CR Jr., Dickson DW, Parisi JE, Xu YC, Cha RH, O’Brien PC, Edland SD, Smith GE, Boeve BF, Tangalos EG, Kokmen E, Petersen RC (2002) Antemortem MRI findings correlate with hippocampal neuropathology in typical aging and dementia. Neurology 58, 750-757. Jack CR Jr., Petersen RC, Xu Y, O’Brien PC, Smith GE, Ivnik RJ, Boeve BF, Tangalos EG, Kokmen E (2000) Rates of hippocampal atrophy correlate with change in clinical status in aging and AD. Neurology 55, 484-489. Jack CR Jr., Petersen RC, Xu YC, O’Brien PC, Smith GE, Ivnik RJ, Boeve BF, Waring SC, Tangalos EG, Kokmen E (1999) Prediction of AD with MRI-based hippocampal volume in mild cognitive impairment. Neurology 52, 1397-1403.

[18]

[19]

[20]

[21]

[22]

[23]

[24] [25]

[26]

[27]

[28]

[29]

[30]

[31]

Braak H, Braak E (1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathologica 82, 239259. Jack CR Jr., Lowe VJ, Weigand SD, Wiste HJ, Senjem ML, Knopman DS, Shiung MM, Gunter JL, Boeve BF, Kemp BJ, Weiner M, Petersen RC, the Alzheimer’s Disease Neuroimaging I (2009) Serial PIB and MRI in normal, mild cognitive impairment and Alzheimer’s disease: implications for sequence of pathological events in Alzheimer’s disease. Brain 132, 1355-1365. Sjogren M, Vanderstichele H, Agren H, Zachrisson O, Edsbagge M, Wikkelso C, Skoog I, Wallin A, Wahlund L-O, Marcusson J, Nagga K, Andreasen N, Davidsson P, Vanmechelen E, Blennow K (2001) Tau and A{beta}42 in cerebrospinal fluid from healthy adults 21–93 years of age: establishment of reference values. Clin Chem 47, 1776-1781. Pruessner JC, Li LM, Serles W, Pruessner M, Collins DL, Kabani N, Lupien S, Evans AC (2000) Volumetry of hippocampus and amygdala with high-resolution MRI and three-dimensional analysis software: minimizing the discrepancies between laboratories. Cereb Cortex 10, 433-442. Vanderstichele H, van Kerschaver E, Hesse C, Davidsson P, Buyse MA, Andreasen N, Minthon L, Wallin A, Blennow K, Vanmechelen E (2000) Standardization of measurement of beta-amyloid(1-42) in cerebrospinal fluid and plasma. Amyloid 7, 245-258. Wahlund LO, Barkhof F, Fazekas F, Bronge L, Augustin M, Sjogren M, Wallin A, Ader H, Leys D, Pantoni L, Pasquier F, Erkinjuntti T, Scheltens P (2001) A new rating scale for age-related white matter changes applicable to MRI and CT. Stroke 32, 1318-1322. Hedges LV, Olkin I (1985) Statistical Methods for MetaAnalysis, Academic Press, Orlando. Whitworth JA (2003) 2003 World Health Organization (WHO)/International Society of Hypertension (ISH) statement on management of hypertension. J Hypertens 21, 1983-1992. Yang LB, Lindholm K, Yan R, Citron M, Xia W, Yang XL, Beach T, Sue L, Wong P, Price D, Li R, Shen Y (2003) Elevated beta-secretase expression and enzymatic activity detected in sporadic Alzheimer disease. Nat Med 9, 3-4. Klyubin I, Betts V, Welzel AT, Blennow K, Zetterberg H, Wallin A, Lemere CA, Cullen WK, Peng Y, Wisniewski T, Selkoe DJ, Anwyl R, Walsh DM, Rowan MJ (2008) Amyloid beta protein dimer-containing human CSF disrupts synaptic plasticity: prevention by systemic passive immunization. J Neurosci 28, 4231-4237. Walsh DM, Klyubin I, Fadeeva JV, Cullen WK, Anwyl R, Wolfe MS, Rowan MJ, Selkoe DJ (2002) Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term potentiation in vivo. Nature 416, 535-539. Shankar GM, Li S, Mehta TH, Garcia-Munoz A, Shepardson NE, Smith I, Brett FM, Farrell MA, Rowan MJ, Lemere CA, Regan CM, Walsh DM, Sabatini BL, Selkoe DJ (2008) Amyloid-beta protein dimers isolated directly from Alzheimer’s brains impair synaptic plasticity and memory. Nat Med 14, 837-842. Guglielmotto M, Aragno M, Autelli R, Giliberto L, Novo E, Colombatto S, Danni O, Parola M, Smith MA, Perry G, Tamagno E, Tabaton M (2009) The up-regulation of BACE1 mediated by hypoxia and ischemic injury: role of oxidative stress and HIF1 alpha. J Neurochem 108, 1045-1056. Tamagno E, Parola M, Bardini P, Piccini A, Borghi R, Guglielmotto M, Santoro G, Davit A, Danni O, Smith MA, Perry G, Tabaton M (2005) Beta-site APP cleaving enzyme


[32]

[33]

[34]

[35]

[36] [37]

[38]

up-regulation induced by 4-hydroxynonenal is mediated by stress-activated protein kinases pathways. J Neurochem 92, 628-636. Willem M, Garratt AN, Novak B, Citron M, Kaufmann S, Rittger A, DeStrooper B, Saftig P, Birchmeier C, Haass C (2006) Control of peripheral nerve myelination by the betasecretase BACE1. Science 314, 664-666. Holsinger RMD, Lee JS, Boyd A, Masters CL, Collins SJ (2006) CSF BACE1 activity is increased in CJD and Alzheimer disease versus other dementias. Neurology 67, 710-712. Verheijen JH, Huisman LG, van Lent N, Neumann U, Paganetti P, Hack CE, Bouwman F, Lindeman J, Bollen EL, Hanemaaijer R (2006) Detection of a soluble form of BACE1 in human cerebrospinal fluid by a sensitive activity assay. Clin Chem 52, 1168-1174. Wu G, Sankaranarayanan S, Tugusheva K, Kahana J, Seabrook G, Shi X-p, King E, Devanarayan V, Cook JJ, Simon AJ (2008) Decrease in age-adjusted cerebrospinal fluid [beta]secretase activity in Alzheimer’s subjects. Clin Biochem 41, 986-996. Blennow K, Hampel H (2003) CSF markers for incipient Alzheimer’s disease. Lancet Neurol 2, 605-613. Shaw LM, Vanderstichele H, Knapik-Czajka M, Clark CM, Aisen PS, Petersen RC, Blennow K, Soares H, Simon A, Lewczuk P, Dean R, Siemers E, Potter W, Lee VM, Trojanowski JQ (2009) Cerebrospinal fluid biomarker signature in Alzheimer’s disease neuroimaging initiative subjects. Ann Neurol 65, 403-413. Hampel H, Teipel SJ, Fuchsberger T, Andreasen N, Wiltfang J, Otto M, Shen Y, Dodel R, Du Y, Farlow M, Moller HJ, Blennow K, Buerger K (2004) Value of CSF beta-amyloid142 and tau as predictors of Alzheimer’s disease in patients

[39]

[40]

[41]

[42]

[43]

[44]

381

with mild cognitive impairment. Mol Psychiatry 9, 705710. Strozyk D, Blennow K, White LR, Launer LJ (2003) CSF Abeta 42 levels correlate with amyloid-neuropathology in a population-based autopsy study. Neurology 60, 652-656. Fagan AM, Mintun MA, Mach RH, Lee SY, Dence CS, Shah AR, LaRossa GN, Spinner ML, Klunk WE, Mathis CA, DeKosky ST, Morris JC, Holtzman DM (2006) Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans. Ann Neurol 59, 512-519. Yan R, Bienkowski MJ, Shuck ME, Miao H, Tory MC, Pauley AM, Brashler JR, Stratman NC, Mathews WR, Buhl AE, Carter DB, Tomasselli AG, Parodi LA, Heinrikson RL, Gurney ME (1999) Membrane-anchored aspartyl protease with Alzheimer’s disease [beta]-secretase activity. Nature 402, 533-537. Bourne KZ, Ferrari DC, Lange-Dohna C, Rossner S, Wood TG, Perez-Polo JR (2007) Differential regulation of BACE1 promoter activity by nuclear factor-kappaB in neurons and glia upon exposure to beta-amyloid peptides. J Neurosci Res 85, 1194-1204. Heneka M, Sastre M, Dumitrescu-Ozimek L, Dewachter I, Walter J, Klockgether T, Van Leuven F (2005) Focal glial activation coincides with increased BACE1 activation and precedes amyloid plaque deposition in APP[V717I] transgenic mice. J Neuroinﬂam 2, 22. Sastre M, Dewachter I, Landreth GE, Willson TM, Klockgether T, van Leuven F, Heneka MT (2003) Nonsteroidal antiinflammatory drugs and peroxisome proliferator-activated receptor-gamma agonists modulate immunostimulated processing of amyloid precursor protein through regulation of beta-secretase. J Neurosci 23, 9796-9804.