Clinical Chemistry Link to Randox Laboratories Web Site
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Clinical Chemistry 51: 1650-1660, 2005. First published July 14, 2005; 10.1373/clinchem.2005.051201
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Data Supplements
Right arrow All Versions of this Article:
clinchem.2005.051201v1
51/9/1650    most recent
Right arrow Submit an electronic Letter to
the Editor about this paper
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (14)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Vanderstichele, H.
Right arrow Articles by Vanmechelen, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Vanderstichele, H.
Right arrow Articles by Vanmechelen, E.
Related Collections
Right arrow General Clinical Chemistry
Right arrow Proteomics and Protein Markers
(Clinical Chemistry. 2005;51:1650-1660.)
© 2005 American Association for Clinical Chemistry, Inc.

Proteomics and Protein Markers

Amino-Truncated ß-Amyloid42 Peptides in Cerebrospinal Fluid and Prediction of Progression of Mild Cognitive Impairment

Hugo Vanderstichele1,a, Geert De Meyer1, Niels Andreasen2, Vesna Kostanjevecki1, Anders Wallin3, Annika Olsson4, Kaj Blennow4 and Eugeen Vanmechelen1

1 Innogenetics NV, Gent, Belgium.
2 Neurotec, Department of Geriatric Medicine, Huddinge University Hospital, Stockholm, Sweden.
3 Department of Clinical Neuroscience, Department of Psychiatry, Göteborg University, Mölndal, Sweden.
4 Department of Clinical Neuroscience, Experimental Neuroscience, Sahlgrenska University Hospital, Mölndal, Sweden.

aAddress correspondence to this author at: Industriepark Zwijnaarde 7, Box 4, B-9052 Gent, Belgium. Fax 32-9-241-0907; e-mail hugovdr{at}innogenetics.be.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background: Early identification of patients with mild cognitive impairment (MCI) progressing to Alzheimer disease (MCI-AD) by use of biomarkers in cerebrospinal fluid (CSF) is an essential step toward improving clinical diagnosis and drug development. We evaluated whether different ß-amyloid42 (Aß42) peptides can add further information to the combined use of tau and Aß1–42 for predicting risk of progression of MCI to AD.

Methods: We used xMAP® technology to simultaneously quantify different Aß42 peptides modified at the amino terminus, tau, and phosphorylated tau (P-tau181P) in CSF. Aß42 peptide concentrations were measured by use of immunoreactivity toward Aß monoclonal antibodies [3D6 (Aß42-3D6), WO2 (Aß42-WO2), 6E10 (Aß42-6E10), and 4G8 (Aß42-4G8)]. The discriminant ability of the markers was evaluated by ROC curve analysis.

Results: The areas under the curves for the separation of MCI-AD from nonprogressing MCI (MCI-N) were significantly higher when we used Aß42-3D6/Aß42-WO2, Aß42-3D6/Aß42-6E10, or Aß42-3D6/Aß42-4G8 compared with Aß42-3D6. In addition, differentiation of MCI-N from MCI-AD was improved by quantification of full-length Aß1–42 (Aß42-3D6) compared with Aß42-WO2, Aß42-6E10, or Aß42-4G8. Several Aß42 peptides truncated at the amino terminus (Aß11–42 and Aß8–42) were identified in CSF by surface-enhanced laser desorption/ionization time-of-flight technology.

Conclusion: The CSF markers tau, Aß42 forms, and P-tau181P, when used as adjuncts to clinical diagnosis, have the potential to help identify AD pathology and could be a valuable asset for early AD diagnosis.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Therapeutic strategies for Alzheimer disease (AD)1 should be based in part on the prevention of disease progression by initiating treatment as early as possible. It is therefore crucial to accurately identify incipient AD in patients with mild cognitive impairment (MCI). Autopsy data have revealed that AD pathology (e.g., hippocampal atrophy, neurofibrillary tangles in the hippocampus and entorhinal cortex, neuritic plaques in the neocortex, synapse loss, and oxidative stress) can be present as long as 20 years before the clinical onset of dementia (1). However, MCI-diagnosed persons do not ultimately share the same fate; whereas some may develop AD, others progress to another dementia type [e.g., dementia with Lewy bodies (DLB), frontotemporal lobe dementia, or Parkinson disease (PD)], and some will never progress to any significant extent. Individuals who present at dementia clinics with MCI symptoms have a 50% probability of progressing to symptomatic AD within a 4-year period (2).

According to a published consensus report, an ideal biological marker should have a clinical sensitivity ≥80% for detecting AD and a clinical specificity ≥80% for distinguishing AD from control or other dementia types (3). These criteria can be achieved by use of protein determinations of cerebrospinal fluid (CSF), which is a continuum of the brain and the spinal cord interstitial fluid. CSF is an obvious source of markers reflecting central neuropathologic features of the disease [e.g., neuronal degeneration, the disturbance of ß-amyloid (Aß) metabolism with formation of plaques, and hyperphosphorylation of tau leading to formation of neurofibrillary tangles]. The most promising markers at present are total tau, phosphorylated tau, and ß-amyloid42 (Aß42). Different immunoassays for these proteins have been described (4).

Since the discovery of Aß as the major constituent of neuritic plaques in AD (5), several amino (N)-terminally truncated as well as carboxy (C)-terminally elongated Aß peptides have been identified in the brain, cell culture supernatant, and CSF by different methodologies (6)(7)(8). N-Truncated Aß has been described in 35% of AD patients, based on immunoreactivity toward specific monoclonal antibodies (mAbs) (9); these truncated forms represent more than 60% of all Aß species and function as a nucleus for Aß oligomerization and deposition into neuritic plaques (8). The specific chemical nature and relative distributions of Aß peptides at the very early stages of Aß deposition, their presence in biological fluids in different stages of the disease process, and their discriminating properties are still unclear.

We used the xMAP technology (Luminex) (10) to design and evaluate the ability of a new multiplexed bead-based assay format to discriminate Aß peptides. The previously described bead immunoassay format (INNO-BIA Alzbio3; for research use only) provided possibilities for simultaneous quantification of Aß1–42, total tau, and tau phosphorylated at Thr181 (P-tau181P) in a small volume of sample with high precision (11). This report presents results obtained with a bead immunoassay for quantification of Aß42 forms, tau, and P-tau181P. A research study was performed with CSF samples obtained from healthy controls, patients with AD, patients with MCI who progressed to AD (MCI-AD) within a period of 3 years, patients with MCI but who did not progress to AD (MCI-N), and patients with other neurologic disorders (PD or DLB).


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
chemicals
All reagents were of analytical grade. Synthetic peptides were obtained from Bachem, Neosystems, or AnaSpec. The characteristics of the calibrators (recombinant tau, Aß1–42, and a phosphorylated tau peptide) have been described in detail elsewhere (12)(13). mAbs 6E10 and 4G8 were obtained from Signet Laboratories, and WO2 was obtained from The Genetics Company. The INNOTEST® ß-AMYLOID(1–42), INNOTEST hTAU Antigen, and INNOTEST PHOSPHO-TAU(181P) were from Innogenetics NV.

antibody characteristics and coupling on microspheres
The characteristics of the mAbs generated against tau (HT7 and BT2) and P-tau181P (AT270) have been described previously in detail (12). The analytical specificities of the N-terminal–specific mAbs (3D6, 6E10, 4G8, and WO2) used in the present study were further evaluated by use of sandwich immunoassays (ELISA format). Plates were coated with the N-terminal–specific mAbs in phosphate-buffered saline (PBS; 8 mmol/L Na2HPO4 · 2 H2O, 1.6 mmol/L KH2PO4, 2.8 mmol/L KCl, 0.137 mmol/L NaCl, pH 7.4) solution at a concentration of 5 mg/L. After a block-and-wash step, the Aß peptides were incubated simultaneously with the biotinylated detector antibody (21F12) for 1 h at room temperature. After a wash step, peroxidase-labeled streptavidin (RDI) was used for quantification. The Aß42-peptide concentration between assay formats was normalized with respect to immunoreactivity toward binding to 4G8. Aß42 peptides binding to 3D6, WO2, 4G8, or 6E10 are further described in the text as Aß42-3D6, Aß42-WO2, Aß42-4G8, or Aß42-6E10.

xMAP technology
The xMAP technology is a flow cytometric method involving covalent coupling of a capture mAb to spectrally specific fluorescent microspheres. Because different microsphere sets can be combined within one method and each bead number is linked with only one antibody, unequivocal identification of the analyte in the mixture is possible. Differences from classic immunoassays include multianalyte testing, lower sample volumes, fewer handling steps, less total hands-on time, and only one test protocol for all analytes. Carboxylated beads were chemically coupled with mAbs, as described by Olsson et al. (11). More technical details on selected assay components are described in Table 1 .


View this table:
[in this window]
[in a new window]
  		Table 1. Assay format.

methods for bead-based xMAP technology
All CSF incubations were performed at room temperature (defined as 18–30 °C) and in the dark (plates covered with aluminum foil). A 96-well filter plate (Millipore Corporation) was prewetted with 225 µL of wash buffer. The wash buffer was removed from the plates by use of a vacuum manifold (Millipore). Beads were sonicated and vortex-mixed before addition to the filter plates (3000 beads/well of each bead region in 100 µL of solution), followed by removal of buffer. A mixture of 2 biotinylated detector mAbs (21F12 and HT7) was added to the wells of the filter plate (25 µL/well). Subsequently, calibrators, blanks, or CSF samples were added (75 µL/well) in duplicate and incubated overnight on a plate shaker. The plate was then washed 3 times with 225 µL of wash buffer. Phycoerythrin-labeled streptavidin (Caltag Laboratories) was added (100 µL/well) and incubated for 1 h on a plate shaker, after which the plate was washed 3 times with 225 µL of wash buffer. Finally, 100 µL of PBS was added to each well. Samples were incubated for 2 min on a plate shaker and then analyzed on the Luminex 100 IS; 100 beads were analyzed for each bead region. Median fluorescence values were reported. Results were excluded if the counted number for a specific bead region was too low.

assay calibration
The calibrators used in the present assay formats were identical to those used in the ELISA methods for the respective proteins [INNOTEST ß-AMYLOID(1–42), INNOTEST hTAU Ag, and INNOTEST PHOSPHO-TAU(181P)]. The preparation and characteristics of these calibrators have been described in detail previously (12)(13)(14)(15); however, the 3 proteins are combined in one vial for the research multiplexed immunoassays. All biomarker data are presented as ng/L and do not take into account possible differences in concentrations observed between ELISA formats and multiplexed products as was described in Olsson et al. (11).

surface-enhanced laser desorption/ionization technology
Surface-enhanced laser desorption/ionization time-of-flight (SELDI-TOF) technology was used to identify Aß42 species (16). mAb 4D7A3 (Innogenetics) or a control mouse IgG was covalently linked to the PS10 ProteinChip array. Briefly, 1 µg of 4D7A3 was applied to the array spot and incubated in a humidity chamber (3 h at room temperature) to allow covalent binding to the PS10 ProteinChip array. The arrays were washed twice with Dulbecco PBS (DPBS) containing 1 mL/L Triton X-100 and twice with PBS (pH 8.0). Unreacted sites were then blocked by incubating 3 µL of 10 g/L bovine serum albumin in DPBS for 2 h at room temperature. Excess bovine serum albumin was removed by 2 washes with DPBS containing 5 mL/L Triton X-100 followed by 3 washes with DPBS. Blocking was performed with 0.5 mol/L Tris (pH 8.0) for 2 h at room temperature, after which the wash procedure described above was repeated. CSF samples (100 µL) in 0.1 mol/L urea containing 1 g/L CHAPS were loaded on a spot by use of the ProteinChip bioprocessor and incubated overnight at 4 °C with constant shaking. Two wash steps, as described above, followed by 2 additional washes with 50 mmol/L HEPES (pH 7.0) completed the reaction. After the arrays had dried, a 20% saturated solution of {alpha}-cyano-4-hydroxycinnamic acid (Ciphergen Biosystems) in 5 mL/L trifluoroacetic acid–500 mL/L acetonitrile–495 mL/L water was applied to each spot. Mass analysis was performed on a ProteinChip reader (Model PBS II; Ciphergen). For calibration purposes, 7 fmol of Aß9–42 peptide (AnaSpec) and 6 fmol of bovine insulin (Ciphergen) were applied and used for data calibration.

research study
Samples.
The study was performed on CSF samples archived at the Sahlgrenska University Hospital, Göteborg, Sweden. Lumbar puncture was performed in the morning under standard conditions. CSF samples were taken at baseline (first visit to the memory clinic). A 12-mL sample of CSF was collected and gently mixed to avoid gradient effects. CSF samples with >500 erythrocytes/µL were excluded from the analysis. All CSF samples were centrifuged at 2000g for 10 min to remove cells and debris. CSF was stored in polypropylene tubes (to avoid adsorption of proteins to the test tube walls) at –80 °C until analysis. All CSF samples (75 µL/sample) were analyzed on the same day. No specific training was required.

Specimen collection.
The diagnostic performance of the new xMAP multianalyte bead-based immunoassay was evaluated with CSF samples from 66 AD patients, 29 healthy controls, 15 patients with DLB, 15 with PD, and 41 with MCI. Recruitment of the patients was based on admission to the hospital for evaluation of cognitive symptoms. The control material was collected from healthy volunteers. Of the MCI patients who were followed up for a period of 3 years, 15 progressed to AD (MCI-AD), whereas 26 patients did not (MCI-N). Clinical data for the patients are given in Table 2 . The control group consisted of individuals without histories, symptoms, or signs of psychiatric or neurologic disease, malignant disease, or systemic disorders (e.g., rheumatoid arthritis or infectious diseases). All patients underwent a thorough clinical investigation, which included a medical history; physical, neurologic, and psychiatric examinations; screening laboratory tests; an electroencephalogram; and a computerized tomography scan of the brain. The diagnosis of probable AD was made according to the National Institute of Neurological and Communitive Disorders and Stroke and Alzheimer’s Disease and Related Disorders Association criteria (17). No patient with AD had a family history of dementia, suggestive of autosomal-dominant AD. MCI diagnoses were made according to Petersen et al. (18) if the patient met the following criteria: (a) a subjective memory complaint, preferably corroborated by an informant; (b) an objective memory impairment adjusted for age and education; (c) normal general cognitive function; (d) preservation of activities of daily living; and (e) no dementia. In the MCI patient group, no symptom other than memory impairment was reported at baseline investigation, i.e., only cases with the amnestic form of MCI were included in the study. None of these patients fulfilled DSM-IV criteria for dementia. MCI patients for whom follow-up data were available were selected for this study. The mean follow-up period of the patients was 3 years. The diagnosis of DLB was made according to the criteria of McKeith et al. (19) and that of definite PD according to the British Brain Bank diagnostic criteria (20). Cognitive status was examined by use of the Mini Mental State Examination (21). Clinical evaluations and diagnoses were made without knowledge of the results of the biochemical analyses. Patients were recruited from February 1997 to September 2001.


View this table:
[in this window]
[in a new window]
  		Table 2. Demographic data for the patients enrolled in the present study.

ethics
The Ethics Committees of the Universities of Göteborg and Umeå approved the study. All patients or their relatives gave informed consent to participate in the study.

statistics
For the different analytes, the concentrations were compared among the treatment groups by all possible pairwise t-tests. The normality assumptions were verified by use of QQ plots. For each analyte, a Bonferroni adjustment was made to adjust for multiple testing. Analyses were performed in SAS (Ver. 9.1). Correlations are reported in terms of the Spearman rank correlation coefficient together with a test for difference from 0. The relationship between sensitivity and specificity for the biomarkers was described by ROC curve analysis. The area under the curve (AUC) was calculated by use of the Medcalc Program (Medcalc Software).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
antibody characterization
Epitope mapping of different N-terminal–specific mAbs (3D6, 6E10, 4G8, and WO2) against the Aß42 peptide was done with sandwich immunoassays with biotinylated 21F12 as detection antibody; the results are shown in Fig. 1 . In contrast to 3D6, all other assay formats for Aß42 detection were not specific for Aß1–42. mAbs 4G8, 6E10, and WO2 recognized not only the full-length Aß1–42, but also Aß42 peptides truncated at the NH2 terminus. The antibody pair 3D6/21F12 preferably quantified full-length Aß1–42, whereas all other combinations (21F12/6E10, 21F12/4G8, and 21F12/WO2) quantified AßN–42. The 6E10/21F12 antibody pair was able to detect Aß42 peptides from Aß1–42 to Aß5–42, but Aß8–42 and Aß9–42 were not detected. The WO2/21F12 pair detected only Aß1–42 and Aß2–42. The 4G8/21F12 combination detected all peptides used in the present study, independent of the NH2 terminus.



View larger version (21K):
[in this window]
[in a new window]
  		Figure 1. Epitope mapping of N-terminal–specific amyloid mAbs by ELISA.

Aß42 peptides: {blacksquare}, Aß1–42; {circ}, Aß2–42; +, Aß3–42; x, Aß4–42; •, Aß5–42; {triangleup}, Aß8–42; {triangledown}, Aß9–42. Experimental details are described in the Materials and Methods. Results are the mean of 2 measurements.

comparison of formats
The concentrations of Aß1–42, total tau, and P-tau181P measured by a research version of the INNO-BIA AlzBio3 (11) and those obtained by the present assay format were correlated: r = 0.824 (P <0.01; n = 159) for Aß1–42, 0.922 (P <0.01; n = 159) for tau, and 0.921 (P <0.01; n = 143) for P-tau181P. In addition, the correlations (r) between results of the present assay format and the classic ELISA immunoassays (INNOTEST) for the same analytes were 0.863 (P <0.01; n = 166) for Aß1–42, 0.865 (P <0.01; n = 165) for tau, and 0.900 (P <0.01 for values <200 ng/L in the INNOTEST; n = 163) for P-tau181P.

research study
We used the xMAP-based assay format to compare the concentration of each biomarker in CSF for different patient groups. Aß42-3D6 was significantly decreased in the AD (P <0.01), DLB (P <0.01), PD (P = 0.0072), MCI-N (P <0.01), and MCI-AD (P <0.01) groups compared with controls, and was significantly different in the MCI-N (P <0.01), MCI-AD (P <0.01), and PD (P <0.01) groups compared with the AD group.

Aß42-WO2 was significantly decreased in AD (P <0.01), DLB (P <0.01), MCI-N (P <0.01), and PD (P <0.01) patients compared with controls, and was significantly higher in MCI-N (P = 0.018), MCI-AD (P <0.01), and PD (P <0.01) patients compared those with AD. Aß42-4G8 was significantly decreased in the AD (P <0.01), DLB (P <0.01), PD (P = 0.033), and MCI-N (P <0.01) groups, but not the MCI-AD group (P = 0.431), compared with controls and was significantly higher in patients with MCI-N (P = 0.0458), MCI-AD (P <0.01), or PD (P <0.01) compared with those with AD. Aß42-6E10 was significantly decreased in the patients with AD (P <0.01), DLB (P <0.01), PD (P <0.01), MCI-N (P <0.01), or MCI-AD (P = 0.039) compared with controls, and was significantly higher in patients with MCI-N (P = 0.028), MCI-AD (P <0.01), or PD (P = 0.021) compared with those with AD.

For tau, concentrations were significantly increased in the AD (P <0.01) and MCI-AD (P <0.01) groups compared with controls, and were significantly lower in patients with DLB (P <0.01), MCI-N (P <0.01), or PD (P <0.01 compared with those with AD). No difference was observed for AD vs MCI-AD (P = 0.574).

P-tau181P concentrations were significantly increased in patients with AD (P <0.01) or MCI-AD (P <0.01) compared with controls and were lower in patients with DLB (P <0.01), MCI-N (P <0.01), or PD (P <0.01) compared with those with AD (Fig. 1Up in the Data Supplement that accompanies the online version of this article athttp://www.clinchem.org/content/vol51/issue9/ ), whereas concentrations in MCI-AD patients were not statistically different (P = 0.704 vs AD).

In subsequent analyses, we compared CSF from MCI-AD patients with CSF from MCI-N patients. In contrast to tau (P <0.01), P-tau181P (P <0.01), or a minor difference for Aß42-4G8 (P = 0.029), there was no evidence of a difference between MCI-AD and MCI-N patients for the concentrations of Aß42-3D6 (P = 0.364), Aß42-6E10 (P = 0.239), or Aß42-WO2 (P = 0.106; Fig. 2 ). However, when we used the ratios Aß42-3D6/Aß42-6E10 (P <0.01), Aß42-3D6/Aß42-4G8 (P <0.01), and Aß42-3D6/Aß42-WO2 (P <0.01), we observed a difference among the groups (Fig. 3 ). This noteworthy result needs to be confirmed in further studies with MCI patients, including an age-matched contrast group. In the present study, the MCI-N group was younger than the MCI-AD group.



View larger version (21K):
[in this window]
[in a new window]
  		Figure 2. Box plots representing xMAP results for individual proteins.

C, healthy controls. Significant differences: vs AD, *, P <0.05; **, P <0.01; vs healthy controls, +, P <0.05; ++, P <0.01.



View larger version (19K):
[in this window]
[in a new window]
  		Figure 3. ROC curve analysis for the discrimination ability of ß-amyloid forms for differentiation of MCI-AD and MCI-N patients (left), and box plots representing results obtained for Aß1–42/AßN–42 ratios as measured with different mAbs (right).

Significant difference: vs AD, **, P <0.01; vs MCI-AD, +, P <0.05; ++, P <0.01; NS, not significant.

There was no evidence of a difference in P-tau181P (P = 0.704), tau (P = 0.574), Aß42-3D6/Aß42-6E10 (P = 0.103), Aß42-3D6/Aß42-WO2 (P = 0.212), or Aß42-3D6/Aß42-4G8 (P = 0.142) between AD and MCI-AD patients (Figs. 2Up and 3Up ), confirming the progression of the MCI-AD group to AD pathology.

ROC analysis was performed on samples for which all Aß values were available (AD, n = 59; MCI-AD, n = 14; MCI-N, n = 24). This analysis revealed that the AUC for the separation of MCI-AD and MCI-N patients was significantly higher for the ratios Aß42-3D6/Aß42-6E10 [0.876; 95% confidence interval (CI), 0.729–0.960; P <0.01], Aß42-3D6/Aß42-4G8 [0.943 (0.817–0.991); P <0.01], and Aß42-3D6/Aß42-WO2 [0.943 (0.817–0.991); P <0.01] compared with Aß42-3D6 alone [0.576 (0.405–0.734; Fig. 3Up ]. At optimal cutoffs (maximum sensitivity and specificity), the sensitivity/specificity values were 71.4%/91.7%, 92.9%/79.2%, and 78.6%/100% for the ratios Aß42-3D6/Aß42-6E10, Aß42-3D6/Aß42-4G8, and Aß42-3D6/Aß42-WO2, respectively. In addition, differentiation of MCI-N from AD was improved by quantification of the full-length Aß42-3D6 [AUC = 0.850 (95% CI, 0.755–0.919)] compared with Aß42-6E10 [0.655 (0.543–0.756); P <0.01], Aß42-4G8 [0.661 (0.549–0.762); P <0.01], or Aß42-WO2 [0.656 (0.544–0.757); P <0.01]. Similar results were obtained for the comparison between the control and AD groups.

identification of Aß42 peptides by seldi-tof technology
We further investigated the presence of truncated Aß42 peptides, using SELDI-TOF technology. For this purpose, CSF samples from controls and AD patients were used. Application of the mAb 4D7A3 on the ProteinChip array revealed 3 distinct Aß42 peptide peaks (Fig. 4 ), which could be assigned on the basis of their molecular masses as Aß1-42, Aß11-42, and Aß8-42 (Table 3 ). The mass accuracy was calculated as the mean value for each peptide in 30 CSF samples.



View larger version (24K):
[in this window]
[in a new window]
  		Figure 4. SELDI-TOF spectra of CSF.

The N-truncated Aß42 peptides Aß8–42, Aß11–42, and Aß1–42 are as assigned in CSF. The first spectrum (A) shows results from the reaction of CSF with a nonrelevant (anti-amyloid) antibody. The second (B) and third (C) spectra show results from reaction of CSF with anti-Aß42 mAb 4D7A3. Bovine insulin with a single- and double-positive charge (5733.6 and 2866.8 Da, respectively; represented by *) and synthetic Aß9–42 (5733.6 Da) were added to CSF for internal spectra calibration. External calibration was performed with dynaphorin (2147.5 Da), human adrenocorticotropic hormone(1–24) (2933.5 Da), and bovine insulin ß-chain (3495.9 Da) for mass accuracy calculation.


View this table:
[in this window]
[in a new window]
  		Table 3. Molecular masses and mass accuracy for Aß peptides detected in human CSF by SELDI-TOF.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
There is a growing need to shorten the lead times for development of new AD drugs. The increasing costs of research and development have encouraged pharmaceutical companies to consider the use of biomarkers in phase 1 clinical trials with the ultimate goal of finding surrogate markers that are influenced by the drug treatment protocol. A surrogate endpoint or marker is defined as a laboratory measurement or physical sign that is used in therapeutic trials as a substitute for a clinically meaningful endpoint that is a direct measure of how a patient feels, functions, or survives and is expected to predict the effect of the therapy (22). Diagnostic approaches are aimed at (a) the identification of incipient AD in patients with MCI or age-associated memory impairment (early diagnosis), (b) the differential diagnosis of AD from other types of dementia (DLB, frontotemporal lobe dementia, and PD), or (c) the establishment of surrogate markers for patient management or therapy follow-up. At present, no single biomarker provides all of the necessary information, but the diagnostic accuracy is significantly improved if the results of measurement of several different proteins (tau, P-tau, and Aß42) can be combined (4).

In the present study, the multiplexed bead-based xMAP technology was used to develop assays for simultaneous quantification of Aß42 peptides, tau, and P-tau181P in human CSF. The release of tau is considered a reflection of the intensity of neuronal damage and degeneration; P-tau concentrations reflect the phosphorylation state of tau; Aß42 is a marker of plaque formation and reflects the stage of the disease (23); and identification of truncated Aß42 points to the initiation of plaque formation (8).

We used different assay formats to determine the concentrations of biomarkers in CSF samples obtained from several patient groups. Notwithstanding the overlap in concentrations when different diagnostic groups were compared, the outcomes of this research study suggest that the combined use of tau, P-tau, and Aß1–42 concentrations can lead to efficient diagnosis of AD. The concentrations of these 3 CSF biomarkers may change very early in the disease process and may, after further validation, be of value to differentiate MCI-AD cases from benign MCI cases (4). In the present study, we extended the panel of probable useful biomarkers by incorporating the ratios between different Aß42 peptides. In a multiprotein assay format, it was shown that MCI patients with progression to AD could be identified some years before the onset of clinical dementia by incorporation of the ratios among different Aß42 peptides (modified at the NH2 terminus) as outcome variables.

The observation that quantification of full-length Aß provides a higher discriminating ability (for differentiation of AD from controls) than do shorter forms of the Aß42 peptide explains some of the differences among published results obtained with different Aß assay formats. Efficient standardization of CSF assays should improve interpretation when data derived from different clinical centers are compared (4).

The combination of tau, P-tau, and the Aß42 peptides provides a panel of biomarkers reflecting different pathologic hallmarks, relevant in different phases of the disease process, that can be used for diagnostic purposes as well as for patient management. It is important to perform an evaluation study with a large number of patient samples in which all biomarkers are quantified, to design algorithms for efficient diagnosis or patient follow-up. The discriminating ability of each protein should be compared with that of a panel of markers analyzed in clinical routine. In addition, the added value for each biomarker in the diagnostic work-up should be validated in detail.

The different Aß42 peptides were measured by use of mAbs to different epitopes located at the NH2 terminus of Aß. To our knowledge, this is the first study in which 4 different mAbs (3D6, 6E10, 4G8, and WO2) were compared for performance. mAb 3D6 (in contrast to 6E10, 4G8, or WO2) selectively identified Aß peptides with a free amino group at Asp1, which confirmed previously published data obtained by peptide mapping experiments (13) or 2-dimensional gel electrophoresis (8). mAbs 6E10 and 4G8 are able to capture shorter peptides as well as Aß1–42. According to Clarke and Shearman (24), residues Gly8 and Tyr9 of the Aß sequence are considered as an essential part of the epitope of mAb 6E10 (24); the absence of reactivity against Aß8–42 and Aß9–42 points to a broadening of the epitope of mAb 6E10 toward positions 6 and 7. Assays developed with mAb WO2 also have an epitope directed to the NH2 terminus of Aß (positions 3–8) (25).

Because the differentiation of MCI-AD from MCI-N was comparable when mAb WO2, 6E10, or 4G8 was used as the capture mAb, we expect no assay bias caused by potential interference by soluble amyloid precursor protein. For mAb 3D6, an absence of reactivity toward soluble amyloid precursor protein has been described previously (13). In addition, ROC analysis revealed no statistical difference for diagnosis of AD vs MCI-N for Aß42-3D6 [AUC = 0.826 (95% CI, 0.732–0.898); n = 64] compared with Aß42-3D6 quantified with a research version of the INNO-BIA AlzBio3 (11) [0.838 (0.745–0.907); n = 26; P = 0.804]; our multianalyte approach thus does not appear to introduce any assay bias. Furthermore, MCI-AD can also be differentiated by single-analyte (ELISA) analysis (our unpublished observations).

The mAb combinations WO2/21F12, 4G8/21F12, and 6E10/21F12 identified a greater number of different N-modified forms than did the combination 3D6/21F12, which detected only Aß1–42. Discrimination of MCI-AD patients from patients with MCI-N was improved by use of the ratio Aß1–42/AßN–42 instead of Aß1–42 concentrations alone. The relevance of the mAb approach was validated with SELDI-TOF technology, which confirmed the presence of truncated Aß42 peptides. Apart from Aß1–42, several other peptides, modified at the NH2 terminus, were identified: Aß11–42 and Aß8–42. In their study using mAb 1E8, Wiltfang et al. (9) found that there is some diagnostic value for quantification of CSF Aß2–42, although no epitope mapping results were reported. However, when we used the SELDI technology with mAb 4D7A3, we could not confirm the presence of Aß2–42 in CSF. In addition, there was no evidence of secretion of p3 (Aß17–42) in CSF. Gouras et al. (26) also failed to detect Aß17–42 in rodent primary neuronal cell culture media derived from day-17 embryonic rats, but the presence of Aß11–42 is in line with our data.

The improved discrimination of MCI-AD from MCI-N obtained with the ratio Aß1–42/AßN–42 can be explained by the different amounts of N-truncated Aß42 peptides present in CSF from the different diagnostic groups or by the onset of oligomerization/protofibril formation, a process identified by the use of mAbs. At present, we lack sufficient experimental evidence to make claims about differences in the detection of oligomers, protofibrils, or small aggregates by these mAbs. Aß peptides have many structural and biological complexities. The loss of the NH2 terminus of Aß can have important implications for the progression of plaque morphology or interaction with binding proteins (apolipoprotein E, complement component C1q, heparin, proteoglycans) and Zn2+ ions present and is implicated in plaque formation (27). N-Terminal degradation has been observed in Aß peptide deposits from AD patients (28) and is possibly one of the first steps in the disease process toward amyloid oligomerization and deposition in neuritic plaques (8). Soluble Aß peptides beginning at residues Glu3 and Glu11, with cyclization of the N-terminal Glu, have been detected in AD and Down syndrome, but not in control brains (29)(30). Wiltfang et al.(9) detected Aß peptides truncated at position 2 in 35% of AD patients. N-Truncated forms of Aß42 are also present in biological fluids. Studies have shown that N-terminally modified Aß42 is selectively increased in CSF, whereas Aß2–42 is increased in the brains and CSF of some AD patients (9)(31). Furthermore, low CSF Aß42 concentrations, high CSF total-tau concentrations, or altered Aß40/Aß42 ratios are already present several years before the onset of clinical dementia (4)(32). The diagnostic relevance of C-terminally elongated peptides, identified in CSF and brain tissue from AD patients, needs to be further validated (6)(7).

Other possible types of Aß modifications that could explain the observed results are the formation of oligomers or protofibrils, which are involved in the pathology of the disease. Oligomers are increased in the brains of persons with AD (33). The evidence for secretion of oligomers in CSF or blood or both is still unclear. Using fluorescence correlation spectroscopy with labeled Aß protein as a probe, Pitschke et al. (34) showed that Aß aggregates might be present in CSF samples from AD patients, but not in CSF from non-AD controls, although this result has not been confirmed by other research groups. Small oligomers accumulate in conditioned media of cell cultures and increase after transfection with mutant familial AD presenilins (35). Protofibrils, first discovered as an intermediate in Aß40 amyloidogenesis, can cause oxidative stress and neuronal cell death in cell cultures (36). Larner(37) suggested that subtle disturbances in the formation of Aß peptides (N-truncation, oligomer formation, amyloid precursor protein mismetabolism, C-terminal forms) may lead to a shift from neuroprotection to neurodegeneration, possibly through disruption of negative feedback pathways. It is not clear at present whether these observations can be extrapolated to plasma or serum and, if so, whether they are a reflection of pathophysiologic changes in AD brains. More studies are needed to select the best mAb to design an assay for quantification of AßN–42.

As in other areas of medicine, CSF markers for AD should not be used as isolated tests, but should be added to the information from the clinical examination. Results on the diagnostic capacity of CSF markers come from studies with clinically diagnosed patients. As a consequence, the diagnostic performance of the selected markers cannot be higher than the accuracy of the clinical diagnostic criteria used. In most of the cases, there was an overlap between results for each variable. The selected biomarkers described in the present study are related to different hallmarks of AD. Tau proteins are considered "state" markers, whereas Aß42 proteins can be used as "stage" markers (4). Because of the diagnostic uncertainties and the heterogeneity of the disease process, it will be necessary to validate and implement an algorithm based on a combination of biomarkers to reduce the overlap among diagnostic groups (38). Validation of such a model will require a larger number of samples.

The presence of truncated Aß42 peptides in CSF samples from patients with early AD points to a role in the acceleration of Aß deposition. The multianalyte assay format that we describe, which allows quantification of tau, P-tau, and Aß42 peptides, could be a useful adjunct in the clinical diagnosis of AD pathology and could become a valuable asset for early AD diagnosis.


   Acknowledgments
 
We thank Fred Shapiro, Anne Farmer, and Karen De Vreese for editing and useful discussions and are grateful to Leentje Demeyer, Lies Vandecasteele, Pierre Grognet, Sandy Van Remoortere, and Vicky Mortier for performing the assays. The study was supported in part by Instituut voor aanmoediging van Innovatie door Wetenschap en Technologie (IWT) Grant IWT990220.


   Footnotes
 
1 Nonstandard abbreviations: AD, Alzheimer disease; MCI, mild cognitive impairment; DLB, dementia with Lewy bodies; PD, Parkinson disease; P-tau181P, tau phosphorylated at Thr181; MCI-AD, mild cognitive impairment progressing to Alzheimer disease; MCI-N, mild cognitive impairment not progressing to Alzheimer disease; PBS, phosphate-buffered saline; SELDI-TOF, surface-enhanced laser desorption/ionization time-of-flight; DPBS, Dulbecco phosphate-buffered saline; AUC, area under the curve; and 95% CI, 95% confidence interval.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Davies L, Wolska B, Hilbich C, Multhaup G, Martins R, Simms G, et al. A4 amyloid protein deposition and the diagnosis of Alzheimer’s disease: prevalence in aged brains determined by immunocytochemistry compared with conventional neuropathologic techniques. Neurology 1988;38:1688-1693.[Abstract/Free Full Text]
  2. DeCarli C. Mild cognitive impairment: prevalence, prognosis, aetiology, and treatment. Lancet Neurol 2003;2:15-21.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  3. . Consensus report of the Working Group on "Molecular and Biochemical Markers of Alzheimer’s Disease". The Ronald and Nancy Reagan Research Institute of the Alzheimer’s Association and the National Institute on Aging Working Group. Neurobiol Aging 1998;19:109-116.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  4. Blennow K, Hampel H. CSF markers for incipient Alzheimer’s disease. Lancet Neurol 2003;2:605-613.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  5. Glenner GG, Wong CW. Alzheimer’s disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun 1984;120:885-890.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  6. Lewczuk P, Esselmann H, Meyer M, Wollscheid V, Neumann M, Otto M, et al. The amyloid-ß (Aß) peptide pattern in cerebrospinal fluid in Alzheimer’s disease: evidence of a novel carboxyterminally elongated Aß peptide. Rapid Commun Mass Spectrom 2003;17:1291-1296.[CrossRef][Medline] [Order article via Infotrieve]
  7. Lewczuk P, Esselmann H, Groemer TW, Bibl M, Maler JM, Steinacker P, et al. Amyloid ß peptides in cerebrospinal fluid as profiled with surface enhanced laser desorption/ionization time-of-flight mass spectrometry: evidence of novel biomarkers in Alzheimer’s disease. Biol Psychiatry 2004;55:524-530.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  8. Sergeant N, Bombois S, Ghestem A, Drobecq H, Kostanjevecki V, Missiaen C, et al. Truncated beta-amyloid peptide species in pre-clinical Alzheimer’s disease as new targets for the vaccination approach. J Neurochem 2003;85:1581-1591.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  9. Wiltfang J, Esselmann H, Cupers P, Neumann M, Kretzschmar H, Beyermann M, et al. Elevation of ß-amyloid peptide 2–42 in sporadic and familial Alzheimer’s disease and its generation in PS1 knockout cells. J Biol Chem 2001;276:42645-42657.[Abstract/Free Full Text]
  10. Oliver KG, Kettman JR, Fulton RJ. Multiplexed analysis of human cytokines by use of the FlowMetrix system. Clin Chem 1998;44:2057-2060.[Free Full Text]
  11. Olsson A, Vanderstichele H, Andreasen N, De Meyer G, Wallin A, Holmberg B, et al. Simultaneous measurement of ß-amyloid(1–42), total tau, and phosphorylated tau (Thr181) in cerebrospinal fluid by the xMAP technology. Clin Chem 2005;51:336-345.[Abstract/Free Full Text]
  12. Vanmechelen E, Vanderstichele H, Davidsson P, Van Kerschaver E, Van Der Perre B, Sjögren M, et al. Quantification of tau phosphorylated at threonine 181 in human cerebrospinal fluid: a sandwich ELISA with a synthetic phosphopeptide for standardization. Neurosci Lett 2000;285:49-52.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  13. Vanderstichele H, Van Kerschaver E, Hesse C, Davidsson P, Buyse M-A, Andreasen N, et al. Standardization of measurement of ß-amyloid(1–42) in cerebrospinal fluid and plasma. Amyloid 2000;7:245-258.[ISI][Medline] [Order article via Infotrieve]
  14. Blennow K, Wallin A, Agren H, Spenger C, Siegfried J, Vanmechelen E. Tau protein in cerebrospinal fluid: a biochemical marker for axonal degeneration in Alzheimer disease?. Mol Chem Neuropathol 1995;26:231-245.[ISI][Medline] [Order article via Infotrieve]
  15. Vandermeeren M, Mercken M, Vanmechelen E, Six J, Van de Voorde A, Martin J-J, et al. Detection of {tau} proteins in normal and Alzheimer’s disease cerebrospinal fluid with a sensitive sandwich enzyme-linked immunosorbent assay. J Neurochem 1993;61:1828-1834.[ISI][Medline] [Order article via Infotrieve]
  16. Frears ER, Stephens DJ, Walters CE, Davies H, Austen BM. The role of cholesterol in the biosynthesis of ß-amyloid. Neuroreport 1999;10:1699-1705.[ISI][Medline] [Order article via Infotrieve]
  17. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 1984;34:939-944.[Abstract/Free Full Text]
  18. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999;56:303-308.[Abstract/Free Full Text]
  19. McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, et al. Consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international workshop. Neurology 1996;47:1113-1124.[Abstract/Free Full Text]
  20. Hughes AJ, Daniel SE, Kilford L, Lees AJ. Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992;55:181-184.[Abstract]
  21. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189-198.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  22. Temple R. Are surrogate markers adequate to address cardiovascular disease drugs?. JAMA 1998;282:790-795.
  23. Wahlund L-O, Blennow K. Cerebrospinal fluid biomarkers for disease stage and intensity in cognitively impaired patients. Neurosci Lett 2003;339:99-102.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  24. Clarke EE, Shearman MS. Quantitation of amyloid-ß peptides in biological milieu using a novel homogeneous time-resolved fluorescence (HTRF) assay. J Neurosci Methods 2000;102:61-68.[Medline] [Order article via Infotrieve]
  25. Ida N, Hartmann T, Pantel J, Schröder J, Zerfass R, Förstl H, et al. Analysis of heterogeneous ßA4 peptides in human cerebrospinal fluid and blood by a newly developed sensitive western blot assay. J Biol Chem 1996;271:22908-22914.[Abstract/Free Full Text]
  26. Gouras GK, Xu H, Jovanovic JN, Buxbaum JD, Wang R, Greengard P, et al. Generation and regulation of ß-amyloid peptide variants by neurons. J Neurochem 1998;71:1920-1925.[ISI][Medline] [Order article via Infotrieve]
  27. Larner AJ. Hypothesis: amyloid ß-peptides truncated at the N-terminus contribute to the pathogenesis of Alzheimer’s disease. Neurobiol Aging 1999;20:65-69.[CrossRef][Medline] [Order article via Infotrieve]
  28. Kalback W, Watson MD, Kokjohn TA, Kuo YM, Weiss N, Luehrs DC, et al. APP transgenic mice Tg2576 accumulate Aß peptides that are distinct from the chemically modified and insoluble peptides deposited in Alzheimer’s disease senile plaques. Biochemistry 2002;41:922-928.[CrossRef][Medline] [Order article via Infotrieve]
  29. Harigaya Y, Saido TC, Eckman CB, Prada CM, Shoji M, Younkin SG. Amyloid ß protein starting pyroglutamate at position 3 is a major component of the amyloid deposits in the Alzheimer’s disease brain. Biochem Biophys Res Commun 2000;276:422-427.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  30. Russo C, Saido TC, DeBusk LM, Tabaton M, Gambetti P, Teller JK. Heterogeneity of water-soluble amyloid ß-peptide in Alzheimer’s disease and Down’s syndrome brains. FEBS Lett 1997;409:411-416.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  31. Tamaoka A, Sawamura N, Fukushima T, Shoji S, Matsubara E, Shoji M, et al. Amyloid ß protein 42(43) in cerebrospinal fluid of patients with Alzheimer’s disease. J Neurol Sci 1997;48:41-45.
  32. Skoog I, Davidsson P, Aevarsson O, Vanderstichele H, Vanmechelen E, Blennow K. Cerebrospinal fluid beta-amyloid 42 is reduced before the onset of sporadic dementia: a population-based study in 85-year-olds. Dement Geriatr Cogn Disord 2003;15:169-176.[CrossRef][Medline] [Order article via Infotrieve]
  33. Enya M, Morishima-Kawashima M, Yoshimura M, Shinkai Y, Kusui K, Khan K, et al. Appearance of sodium dodecyl sulfate-stable amyloid ß-protein (Aß) dimer in the cortex during aging. Am J Pathol 1999;154:271-279.[Abstract/Free Full Text]
  34. Pitschke M, Prior R, Haupt M, Riesner D. Detection of single amyloid ß-protein aggregates in the cerebrospinal fluid of Alzheimer’s patients by fluorescence correlation spectroscopy. Nat Med 1998;4:832-834.[CrossRef][ISI][Medline] [Order article via Infotrieve]
  35. Xia W, Zhang J, Kholodenko D, Citron M, Podlisny MB, Teplow DB, et al. Enhanced production and oligomerization of the 42-residue amyloid ß-protein by Chinese hamster ovary cells stably expressing mutant presenilins. J Biol Chem 1997;272:7977-7982.[Abstract/Free Full Text]
  36. Walsh DM, Hartley DM, Kusumoto Y, Fezoui Y, Condron MM, Lomakin A, et al. Amyloid ß-protein fibrillogenesis. Structure and biological activity of protofibrillar intermediates. J Biol Chem 1999;274:25945-25952.[Abstract/Free Full Text]
  37. Larner AJ. Physiological and pathological interrelationships of amyloid ß peptide and the amyloid precursor protein. Bioassays 1995;17:819-824.[CrossRef][Medline] [Order article via Infotrieve]
  38. De Meyer G, Shapiro F. Biomarker development: the road to clinical utility. Current Drug Discovery 2003;:23-27MAY.



The following articles in journals at HighWire Press have cited this article:


Home page
 J Biomol ScreenHome page
P. G. Szekeres, K. Leong, T. A. Day, A. E. Kingston, and E. H. Karran
Development of Homogeneous 384-Well High-Throughput Screening Assays for A{beta}1-40 and A{beta}1-42 Using AlphaScreenTM Technology
J Biomol Screen, February 1, 2008; 13(2): 101 - 111.
[Abstract] [PDF]

Home page
 NeurologyHome page
R. J. Bateman, G. Wen, J. C. Morris, and D. M. Holtzman
Fluctuations of CSF amyloid-ss levels: Implications for a diagnostic and therapeutic biomarker
Neurology, February 27, 2007; 68(9): 666 - 669.
[Abstract] [Full Text] [PDF]

Home page
 J Geriatr Psychiatry NeurolHome page
T. Sunderland, H. Hampel, M. Takeda, K. T. Putnam, and R. M. Cohen
Biomarkers in the Diagnosis of Alzheimer's Disease: Are We Ready?
J Geriatr Psychiatry Neurol, September 1, 2006; 19(3): 172 - 179.
[Abstract] [PDF]

Home page
 AM J ALZHEIMERS DIS OTHER DEMENHome page
A. Marcos, P. Gil, A. Barabash, R. Rodriguez, M. Encinas, C. Fernandez, and J. A. Cabranes
Neuropsychological Markers of Progression From Mild Cognitive Impairment to Alzheimer's Disease
American Journal of Alzheimer's Disease and Other Dementias, May 1, 2006; 21(3): 189 - 196.
[Abstract] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Data Supplements
Right arrow All Versions of this Article:
clinchem.2005.051201v1
51/9/1650    most recent
Right arrow Submit an electronic Letter to
the Editor about this paper
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (14)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Vanderstichele, H.
Right arrow Articles by Vanmechelen, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Vanderstichele, H.
Right arrow Articles by Vanmechelen, E.
Related Colle