Analysis of Human Tau in Cerebrospinal Fluid - American Chemical

Jan 13, 2010 - autopsy. Clinically used biomarkers present in cerebrospinal fluid (CSF), that is, unphosphorylated or phosphorylated tau and Aβ-pepti...
1 downloads 0 Views 3MB Size
Download PDF
Analysis of Human Tau in Cerebrospinal Fluid Katja Hanisch,† Hilkka Soininen,‡ Irina Alafuzoff,‡,§ and Ralf Hoffmann*,† Institute of Bioanalytical Chemistry, Center for Biotechnology and Biomedicine, Faculty of Chemistry and Mineralogy, Leipzig Universität, 04103 Leipzig, Germany, Department of Neurology, Kuopio University, Kuopio, Finland, and Section of Neuropathology, Kuopio University, Kuopio, Finland, Department of Genetics and Pathology, Uppsala University, S-75185, Uppsala, Sweden Received November 4, 2009

Alzheimer’s disease (AD) is the most common form of dementia. Neuropathologically, it is characterized by two major hallmarks: neurofibrillary tangles (NFT) formed from hyperphosphorylated versions of the tau-protein, and neuritic plaques (NP) containing mostly β-amyloid peptides (Aβ) that are formed from the amyloid precursor protein (APP) by enzymatic cleavage. Despite much progress in recent years, the causes of sporadic (i.e., nonfamiliar) AD are still unclear and its valid diagnosis still relies on autopsy. Clinically used biomarkers present in cerebrospinal fluid (CSF), that is, unphosphorylated or phosphorylated tau and Aβ-peptides of different lengths, lack the necessary specificity and sensitivity. Here, we describe a novel strategy to characterize tau versions present in CSF with respect to their molecular mass and isoelectric point. Aliquots of 1 mL CSF (i.e., 700 to 1300 pg tau) from nondemented persons and histopathologically confirmed AD cases were depleted for six dominant proteins, separated by two-dimensional gel electrophoresis, and then electro-transferred onto PVDF-membranes. Tau was detected with monoclonal antibody (mAb) HT7 conjugated with horseradish peroxidase (HRP). In this way, a complex tau pattern was identified in CSF that was very similar to the tau preparations from autopsy brain samples. The presented strategy enables the analysis of the phosphorylation and processing status of tau in CSF samples from healthy people and patients diagnosed with different neurological disorders. This more-detailed information on circulating tau versions and their clearance rates may facilitate the development of new diagnostic tools. Keywords: Alzheimer’s disease (AD) • Immunoblot • Protein depletion • Two-dimensional gel electrophoresis (2-DE)

Introduction Cerebrospinal fluid (CSF) is a translucent fluid in the ventricle of the brain and spinal cord, with a total volume from 150 to 270 mL in adults.1 CSF is diffused through the brain and thus contains proteins and peptides, as well as other metabolites that reflect the status of the central nervous system (CNS). Thus, CSF contain proteins that are potential biomarkers for different CNS diseases. Identification of valid proteins as biomarkers in CSF is challenging because of their low concentrations in a highly complex sample that includes proteolytic degradation. In addition, a large amount of blood-derived proteins permeate the blood-brain barrier and appear in CSF at high concentrations. Albumin for example, which is present at a concentration of 150-350 µg/mL, accounts for 35-80% of the total protein content in CSF (150-450 µg/mL), while immunoglobulins account for another 10-20%.2,3 * To whom correspondence should be addressed. Prof. Dr. Ralf Hoffmann. Institute of Bioanalytical Chemistry, Center for Biotechnology and Biomedicine, Faculty of Chemistry and Mineralogy, Leipzig University, Deutscher Platz 5, 04103 Leipzig, Germany. Tel. #49 (0) 341 9731330, Fax. #49 (0) 341 9731339. E-mail: [email protected]. † Leipzig University. ‡ Kuopio University. § Uppsala University.

1476 Journal of Proteome Research 2010, 9, 1476–1482 Published on Web 01/13/2010

Alzheimer’s disease (AD) is the most common form of dementia. It is generally suspected based on the clinical history, including information on memory testing, presence of characteristic neurological and neuropsychological features, and absence of other diseases with similar symptoms.4 Advanced medical imaging, such as computed tomography (CT), magnetic resonance imaging (MRI), single photon emission computed tomography (SPECT), and positron emission tomography (PET) can be used to exclude other cerebral pathologies or subtypes of dementia. A final diagnosis of AD can, however, only be obtained postmortem by neuropathological examinations. Thus, identification of valid biomarkers, which ideally would reflect the pathology of AD, is extremely important for the early diagnosis and clinical evaluation of patients with dementia. Besides the diagnostic value, CSF markers may also reveal underlying molecular mechanisms of the disease. AD is characterized by an enormous loss of neurons and synapses in the cerebral cortex and certain subcortical regions, and the massive formation of both neuritic plaques (NP) and neurofibrillary tangles (NFT). NFT are aggregates of the tau protein, especially hyperphosphorylated tau, which forms paired helical filaments (PHF).5 NP contain, in addition to hyperphosphorylated tau in dystrophic neuritis, primarily 38 to 43 residue long amyloid-β peptides (Aβ38 to Aβ43) produced 10.1021/pr901002t

 2010 American Chemical Society

research articles

Analysis of Human Tau in Cerebrospinal Fluid by enzymatic cleavage (secretases) of the amyloid precursor protein (APP) [6, 7]. Three biomarkers in CSF have been evaluated so far and are part of the clinical testing: Aβ42, total tau, and phospho tau.8 Tau is a microtubule-associated protein, which is located mainly in neuronal axons and plays an important role in the structure and function of the neuronal cytoskeleton. The adult human brain contains six splicing forms of tau ranging in size from 352 to 441 residues, as well as big-tau (741 residues), which also contains the 300 residue long exon 4A [9, 10]. The tau splicing forms promote the microtubuli assembly and stabilize them by binding to tubulin via three (3R-tau) or four repeat units (4R-tau).11-13 Hyperphosphorylated tau binds to tubulin at a much lower affinity, resulting in disintegration of the microtubule and axon instability. It also aggregates in the neurons, forming the intracellular NFT.14 CSF contains only very low tau quantities, typically ranging from around 100 to 1200 pg/mL, about 5 to 6 orders of magnitude below the serum albumin level. CSF-tau levels are elevated in relation to healthy subjects in many neurodegenerative diseases, including AD [15, 16]. It is considered that this elevation probably reflects the intensity of neuronal damage and degeneration.17 Different monoclonal antibodies (mAbs) have been used in enzyme-linked immunosorbent assays (ELISA) to determine tau levels in CSF, typically independent of its phosphorylation status (total-tau) [18, 19]. Despite a correlation between the CSF concentrations of Aβ and tau with AD and its progression, both biomarkers lack the desired diagnostic specificity and sensitivity, and cannot distinguish between different dementias. The CSF tau level can differentiate between AD and normal aging only with a specificity of 65 to 86% and a sensitivity of 40-86%, depending on the study design and its evaluation.19 These shortages might potentially be overcome by quantification of Aβ-versions of different lengths (38-43 residues), or phosphorylation sites specific for PHF-tau.20-23 Although phosphorylation of tau is generally considered as a likely biomarker for AD and related disorders, many unanswered questions still remain. First, it is not clear which of the forty phosphorylation sites identified in tau represent promising diagnostic and prognostic markers. Second, it has not been clarified whether the multiphosphorylated epitopes identified in PHF-tau would be better diagnostic candidates, and third, it is not clear to what extent the stages of this chronic disease, over a prolonged duration, influence the different markers. Due to the lack of specific and sensitive mAbs for most phosphorylation sites, it is impossible to screen CSF samples for each phosphorylation site. Moreover, it is not clear whether the quantities of different phosphorylated tau-versions correlate with the phosphorylation status of tau in the NFT and the neurons. To the best of our knowledge this correlation is generally assumed, but has never been proven. Here, we report on a methodology to analyze and relatively quantify all tau-versions present in CSF. First, CSF aliquots were depleted for the six most abundant proteins. This allowed (i) loading of larger CSF volumes onto the following twodimensional gel electrophoresis (2-DE) (ii) improvement of the resolution of 2-DE, and (iii) decrease in the background of the immunoblots used to visualize tau. The high sensitivity of the whole workflow enabled analysis of small CSF sample volumes of typically 1 mL containing about 750 pg of tau. The obtained 2D image for tau was very similar to tau isolated from postmortem brain samples, indicating that CSF indeed reflects

the tau-status of the human brain, with respect to both the tau splicing forms and their phosphorylation degree.

Material and Methods Preparation of Human Tau Protein. Tau was prepared from fresh bovine and frozen human brain samples (occipital cortex) with or without AD pathology according to Johnson et al.24 as reported recently.25 The total tau concentration was determined in a sandwich ELISA with anti-human tau mAb HT7 (Innogenetics GmbH, Heiden-Westfalen, Germany) to capture tau, and mAb Tau5 labeled with peroxidase (Dianova GmbH, Hamburg, Germany) to detect tau. Both mAbs recognize all tau versions independent of their phosphorylation degree. Thus, mAb HT7 (2 µg/mL) was coated in binding buffer (50 mmol/L carbonate buffer, pH 9.6) to each well of a 96-well Immuno plate MaxiSorp (Greiner Bio-One GmbH, Frickenhausen, Germany) at 4 °C overnight. The wells were washed three times with washing buffer (300 µL, 10 mmol/L phosphate buffer, pH 7.4, 0.3 mol/L sodium chloride, 0.05% (v/v) Tween 20) and blocked with blocking buffer (200 µL, 0.5% (w/v) casein in washing buffer). After 1 h the wells were washed again three times with washing buffer (300 µL) and tau was quantified with mAb Tau5-POD (1:1500). Color development was performed with 1-Step-UltraTMB (100 µL; Perbio Science Germany GmbH, Bonn, Germany) in the dark. After 15 min the reaction was stopped by addition of sulfuric acid (0.5 mol/L) and the absorption was recorded at 450 nm (620 nm as reference wavelength). Tau amounts were calculated relative to a standard curve obtained by a serial dilution of recombinant tau (Panvera, Madison, WI). CSF Samples. Control CSF samples (retain samples) from 60-80 year old nondemented persons were provided by the Universita¨tsklinikum Leipzig following the guidelines of the ethical committee. The postmortem human CSF samples were taken prior to removal of the brain with a large syringe from the lateral ventricles. The Kuopio Brain Bank sample collection includes biological material taken from subjects who themselves or their relatives had given their consent for the procedure. The permission both to store the tissue samples in the collection and to use the collected CSF samples for this study was granted by the National Authority for Medicolegal Affairs in Finland (5301/04/046/07) and by the local ethical committee at Kuopio University Hospital (99/2007). All CSF samples were stored at -80 °C. Before analysis, protease inhibitor Mix M (Serva Electrophoresis GmbH, Heidelberg, Germany) was added (10 µL/mL CSF) and the samples were filtered (0.22 µm PVDF Rotilabo filter, Carl Roth GmbH Co. KG, Karlsruhe, Germany) to avoid clogging of the depletion column. CSF-tau concentrations of the control samples were determined by a sensitive sandwich ELISA (Innogenetics, Belgium) according to the manufacturer’s instruction.18 The postmortem CSF AD samples were quantified by a novel, immunoPCR sandwich assay.26 Protein Depletion in CSF. Six abundant proteins (i.e., albumin, transferrin, IgG, IgA, haptoglobin, and antitrypsin) present in CSF were depleted with the Multiple Affinity Removal System (MARS, Agilent Technologies Sales & Services GmbH & Co. KG, Waldbronn, Germany) (4.6 × 50 mm) using eluents A and B provided together with the MARS-column and following the supplier’s protocol. Briefly, CSF (100 µL) was ¨ kta loaded with eluent A (flow rate 250 µL/min) onto an A Purifier 10 System (GE Healthcare Europe GmbH, Munich, Germany). The unretarded proteins were collected, typically in the time interval between 2 and 4 min. After 10 min, the flow rate was increased to 1 mL/min and the retarded, Journal of Proteome Research • Vol. 9, No. 3, 2010 1477

research articles dominant proteins were eluted with eluent B, typically in a broad peak between 13 and 14.5 min. The fractions from ten consecutive separations of the same CSF aliquot were combined, desalted, and concentrated in a vivaspin 500 cartridge (10 kDa cutoff; Vivascience AG, Hannover, Germany) with ammonium bicarbonate buffer (3 mmol/L, pH 7.5), and then lyophilized. Gel Electrophoresis. The lyophilized samples were dissolved in rehydration buffer (7 mol/L urea, 2 mol/L thiourea, 15 mmol/L dithiothreitol (DTT), 2% (w:v) 3-[(3-Cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS), 0.2% BioLyte 3-10, 0.001% bromophenol blue) and loaded onto immobilized pH-gradient (IPG) -strips (3-10 NL, 7 cm, Bio-Rad Laboratories GmbH, Munich, Germany) as recommended by the supplier. Active rehydration was performed at 50 V for 12 h, 250 V for 15 min, and 4,000 V for 2 h in an IEF Cell (Bio-Rad) before the proteins were focused at 4000 V for 20 000 Vh with a limiting current of maximum 50 mA per strip. The strips were equilibrated under reducing conditions (6 mol/L urea, 20% glycerol, 2% SDS, and 0.375 mol/L tris(hydroxymethyl) aminomethane (Tris), pH 8.8, 129 mmol/L DTT, 15 min), and then in the presence of iodoacetamide (6 mol/L urea, 20% glycerol, 2% SDS, 135 mmol/L iodoacetamide, 0.375 mol/L Tris, pH 8.8) to first reduce the disulfides, and then alkylate all thiol groups. The IPG-strips were embedded with sucrose on top of a vertical SDS-PAGE (12% T, 2.67% C, 1 mm thick, 7 cm × 10 cm).27 Immunoblotting. To increase the sensitivity of the immunoblot and to decrease the unspecific background caused by secondary antibodies, mAb HT7 was conjugated with peroxidase (POD) with the EZ-Link plus Activated Peroxidase Kit (Perbio Science), or digoxigenin with the DIG Protein Labeling Kit (Roche Diagnostics GmbH, Mannheim, Germany) following the supplier’s instructions.28-30 mAb HT7-Dig was probed with a mouse anti-digoxin mAb (Dianova). The proteins were electro-transferred from the 2D-gel onto a PVDF-membrane (pore size 0.45 µm, Millipore GmbH, Schwalbach, Germany) in a semi dry transfer cell (Bio-Rad, 25 V, 120 min). The membrane was blocked with milk powder (5%) in TBS (25 mmol/L Tris, pH 7.6, 150 mmol/L sodium chloride) for 1 h and then incubated with mAb HT7-POD (0.1 µg/mL) in TTBS (0.1% Tween 20 in TBS) for 1 h. The membrane was washed three times with TTBS, twice with TBS, and then incubated with a mixture of solutions A and B (1:1 by volume, 0.1 mL/cm2) provided with the ECL Advance Western blotting detection kit (GE Healthcare Europe GmbH, Munich) at room temperature (RT) for 5 min. The dried blot membranes were transferred to an X-ray film cassette (Hypercassette autoradiography cassettes, GE Healthcare), exposed to a sheet of autoradiography film (CL-X Posure film, Perbio Science) for 1 and 15 min, developed with Kodak D-19 Developer and Kodak Fixer (Sigma-Aldrich Chemie GmbH, Munich, Germany), and then scanned on an Image Scanner (GE Healthcare). The software package Tina 2.09 g (Raytest Isotopenmessgera¨t GmbH, Straubenharth, Germany) was used to determine the density of the protein bands. To probe the membranes with a second antibody, the membrane was stripped with the immunodetection removing buffer (Serva). A possible unspecific binding of mAb HT7-POD to the blocked PVDF-membrane was tested in an inhibitory immunoblot in the presence of peptide GAAPPGQK containing the HT7 recognition site, which was synthesized by Fmoc/tBustrategy and purified by reversed-phase chromatography (RPC) 1478

Journal of Proteome Research • Vol. 9, No. 3, 2010

Hanisch et al. 31

in house. The immunoblot was performed as described above, except that mAb HT7-POD (0.1 µg/mL) in TTBS was preincubated with the peptide in approximately 2-fold molar excess of the peptide (1 h, with shaking, RT), before the solution was added to the PVDF-membrane.

Results and Discussion Optimization of the Blotting Conditions. The low concentration of tau in human CSF samples, typically ranging from 100 to 1200 pg/mL, and the small CSF volumes available, require highly sensitive detection techniques, such as a sandwich-ELISA. Even higher sensitivities are required if the different tau versions present in CSF are separated using 2-DE according to their molecular weight (splicing form, terminal processing) and isoelectric points (phosphorylation). Earlier studies have shown that tau preparations from both bovine and human brain samples can be separated in about 50 to 80 different spots using 2-DE.32-34 Consequently, the immunoblot conditions, including the antibody detection system, were first optimized to detect tau with high sensitivity and to reduce cross-reactivity to other proteins present in CSF. All mAbs used in this study recognize epitopes that are homologous in human and bovine tau. Thus, the blotting conditions were optimized with a serial dilution from 26 ng to 52 pg tau prepared from calf brain and separated by SDS-PAGE. The initial detection limits were approximately 26 ng on the nitrocellulose membrane, and 5 ng on the PVDFmembrane using biotinylated mAb HT7 detected with PODlabeled ExtrAvidin and visualized with 3,3′-diaminobenzidine (DAB, Figure S1, Supporting Information). The detection limit was always determined as the lowest tau-concentration loaded onto the SDS-PAGE, where all five bands of bovine tau were detected at an intensity of at least 70% above the background. The optimal blotting time was 120 min for the PVDFmembrane. Both shorter (90 min) and longer (180 min) transfer times reduced the band intensities for all tau isoforms. Replacing the POD/DAB system by alkaline phosphatase and using NBT/BCIP (nitro blue tetrazolium, 5-bromo-4-chloro-3-indolyl) as a substrate increased the sensitivity to approximately 2 ng tau (Figure S1, Supporting Information). A similar sensitivity was obtained for fluorescence detection using avidin conjugated fluorescein isothiocyanate (FITC). As 1 mL CSF usually contains less than 1 ng tau, nine commercially available chemiluminescent substrates were tested, that is, one for alkaline phosphatase and eight for horseradish peroxidase (Table S1 and Figure S2, Supporting Information). The best sensitivities were obtained for the ECL Advanced Western Blotting Detection Kit and the SuperSignal West Femto Kit with an exposure time of 15 min, which both visualized 260 pg tau loaded onto the SDS-PAGE. Another important factor for the detection limit is the antitau mAb and the accessibility of the recognized tau sequence. Thus, the commonly used mAbs HT7, BT2 and Tau5, which recognize positions 159-163, 192-198 and 219-225 of human tau, were studied, as they are all highly sensitive and can recognize tau largely independent of its phosphorylation status. These markers were labeled with digoxigenin or peroxidase (Table S2, Supporting Information), which did not influence the epitope specificity. The sensitivities were very similar, ranging from 350 to 100 pg, with mAb HT7-POD providing the best sensitivity for an exposure time of 15 min (Table S2 and Figure S3, Supporting Information). The sensitivity was not further improved by incubating the blots with a mixture of biotinylated

Analysis of Human Tau in Cerebrospinal Fluid

research articles

Figure 1. 2-DE-immunoblots of E. coli lysates spiked with 1.5 ng of human tau prepared from control (left) and AD (right) brain samples. The proteins were loaded onto IPG-strips (7 cm, 3-10 NL), separated in the second dimension by SDS-PAGE (12% T, 2.67% C) and transferred to a PVDF-membrane. The membrane was blocked with 5% milk powder in TBS, incubated with mAb HT7-POD, and stained with the ECL Advance Western blotting detection kit.

mAbs HT7 and BT2. Preincubation of the PVDF-membranes with silver nitrate35 did also not improve the detection limit, but reduced the background significantly. It was, however, not possible to visualize the tau-bands with a second mAb after stripping due to an extremely high background. Therefore, we did not use the technique in subsequent tests, even though it provides some advantages. Other important aspects are the blocking and washing conditions after blotting the proteins on the membrane, which depend on both the mAb and any contaminating proteins present in the sample (sample matrix), that is, CSF in this study. Several commonly used blocking reagents, that is, skimmed milk powder, Tween 20, gelatin, triton X-100, nonidet-P40 and various combinations of these, were tested. With respect to the cross reactivity of mAb HT7-POD toward HSA (data not shown) and the detection limit for tau, the use of 5% (w:v) skimmed milk powder in TBS for blocking the membrane, followed by washing with TBS, offered the best compromise. The optimized immunoblot conditions following SDS-PAGE provided a detection limit of ∼70 pg for human tau prepared from control samples, and ∼50 pg for tau prepared from AD samples, using an exposure time of 15 min. The achieved detection limits would theoretically allow analysis of only 100-400 µL CSF, depending on the tau concentration. Next, we tested these optimized immunoblotting conditions in combination with 2-DE, which provides a good separation into approximately 80 tau spots based on the molecular mass of the splicing forms and their phosphorylation degree. As the IEF requires a minimum protein load of 5 µg, and in order to mimic a complex protein background, an Escherichia coli lysate (5 µg protein) was spiked with human tau preparations containing 1.5 ng tau. The protein mixture was precipitated with methanol-chloroform to remove low mass impurities (e.g., salts, lipids etc.), which were found to cause a tau loss of around 40%, as determined by a sandwich ELISA. Thus, only ∼900 pg tau were loaded onto the IPG-strip. The immunoblot with mAb HT7-POD showed the expected, well resolved tau spot pattern, including some lower mass versions (Figure 1). The 2Dimmunoblot of the pure E. coli (not spiked with tau) did not show any spots, indicating that mAb HT7-POD is not crossreactive to E. coli proteins within the detection limit. The approximately 10-fold lower sensitivity of the 2-DE immunoblot compared to the SDS-PAGE is attributed to the better resolution of the 2-DE. About 50 tau spots were detected over a wide mass and pH range, compared to the five bands used In the SDS-PAGE to determine the detection limits. Thus,

tau is distributed over a much larger membrane area, which decreases the sensitivity for tau, assuming that both immunoblots can detect the same tau concentration expressed in protein quantity per area (pg/mm2). Analysis of CSF Samples. The obtained sensitivity and the typical tau levels in CSF would require loading at least 1 mL CSF onto the IPG-strips, corresponding to a total protein amount of approximately 400 µg, which exceeds the maximum protein load. Further problems would arise from dominant proteins, especially HSA, having similar molecular weights and isoelectric points as tau. These could disturb separation and detection of the one millionfold lower tau quantities. Moreover, mAb HT7-POD was slightly cross-reactive to HSA, which is most likely attributed to an interaction between HSA and POD, rather than mAk-HT7 (data not shown). For all these reasons it was necessary to deplete CSF for abundant proteins. Two commercial affinity columns were tested for this purpose. The MARS-column removes HSA, IgG, serotransferrin, alpha-1antitrypsin, haptoglobin, and IgA. The ProteomeLab IgY 12 Kit (Beckman-Coulter GmbH, Krefeld, Germany) additionally depletes IgM, fibrinogen, alpha-2-macroglobulin, orosomucoid, apolipoprotein A-I, and apolipoprotein A-II. The MARS column reduced the total protein content of CSF by 76%, whereas the IgY 12 spin cartridge removed 85%. Thus, the depletion step would allow separation of 4-fold (MARS-column), or even 7-fold (IgY 12 spin cartridge) larger CSF volumes by 2-DE. The flow through of the IgY 12 kit, however, only contained 58 ( 5% of the original human tau amount, that is, about 40% of the CSFtau bound unspecifically to the stationary phase or the retarded (depleted) proteins. The tau recovery on the MARS-column was more favorable with 74 ( 2.5%. It should be stressed that these differences do not necessarily reflect the properties of the stationary phases, but might be attributed to the use of a HPLCcolumn (MARS) and spin cartridges (IgY 12).36 The higher unspecific binding of tau to the IgY 12 spin cartridges may not only reduce the total-tau content, but additionally discriminate some tau-versions relative to others. For this reason, and due to the easier handling of the MARS-column on an automatic HPLC-system, all consecutive CSF analyses relied on this kit. Having all steps optimized, we depleted 1 mL each of a lumbar CSF sample from a nondemented person (control, ∼750 pg tau) and two postmortem CSF samples from AD patients with confirmed AD pathology (AD117: ∼1300 pg tau and AD163: ∼1200 pg tau) in several aliquots using the MARS-column. The immunoblots probed with mAb HT7-POD contained intense spot patterns (Figure 2A and B) that were very similar to tau Journal of Proteome Research • Vol. 9, No. 3, 2010 1479

research articles

Hanisch et al.

Figure 2. Tau-immunoblots obtained from 1 mL CSF of a human control (A) and postmortem AD samples (B-D) containing ∼750 pg (A), ∼1300 pg (AD117, B and D) and ∼1200 pg tau (AD163, C). The aliquots were depleted with MARS, desalted, lyophilized, and then separated by IEF (7 cm IPG-strips, 3-10 NL) and SDS-PAGE (12% T, 2.67% C). PVDF-membranes were blocked with 5% milk powder in TBS, and incubated with mAb HT7-POD in the absence (A and B) or presence (C) of peptide GAAPPGQK, washed with TTBS, and stained with the ECL Advance Western blotting detection kit. Alternatively, mAb AT-8 (1:50,000) was used (D) and detected with a POD-conjugated secondary antibody. Ellipses indicate presumed cross-reactive spots stained with mAb HT7 (C) and tau spots recognized only by mAb AT-8 (D) and not with mAb HT7.

prepared from human brain samples (Figure 1), that is, three spot rows for tau with apparent molecular weights between 47 and 65 kDa in the neutral to basic region of the 2DImmunoblot. The control and AD117-samples showed similar spot patterns, with three spot rows in the same mass range between pH 6 and pH 10 with approximately the same absolute spot intensities (Figure 2A and B). Further spot rows in the AD117-sample were detected in the basic region below 47 kDa, as well as in the acidic region between 20 and 30 kDa. It is most likely that these spots are truncated tau-versions, as described before for total tau-prepared from brain.30 They might be more pronounced here due to the postmortem delay (protease activity) and the larger tau quantities loaded. The spots in the acidic region above 47 kDa did most likely not contain tau, as indicated by two additional experiments. When the immunoblot from AD163 was incubated with mAb HT7-POD in the presence of peptide GAAPPGQK, i.e. the epitope sequence of mAb HT7, this area was stained equally well, whereas the presumed tau spots were much weaker, or not detectable at all (Figure 2C). The expected tau spot pattern returned after stripping the membrane and treating it again with mAb HT7POD in the absence of the peptide. Furthermore, anti-R- and anti-β-tubulin mAbs stained parts of the acid region around 47 kDa (CSF from AD117), whereas the assumed tau spots were not visible. The same blot region is also stained by these mAbs for total tau preparations from brain samples.30 As the antitubulin mAbs were not tested for cross-reactivity toward other serum proteins, however, this result does not prove that these spots actually contained tubulin. The blot membrane containing tau from AD117 and developed with mAb HT7-POD was stripped to remove all mAbs, which was confirmed by addition of the ECL reagent. This blot was incubated again with mAb AT-8, a PHF-tau specific 1480

Journal of Proteome Research • Vol. 9, No. 3, 2010

antibody that recognizes a doubly phosphorylated epitope at Ser202 and Thr205.37 Although much weaker than before, most spots of the tau pattern were again visible. Moreover, three new spots were detected at a slightly higher molecular mass (∼70 kDa) that were not stained with HT7-POD, which could represent PHF-tau-like versions of tau present in this CSF sample, which possessed confirmed AD pathology (Figure 2D). This does not prove the presence of PHF-tau versions in AD patients’ lumbar CSF, as they could have been dissolved from tangles during the postmortem delay. An unlikely explanation could be a possible cross-reactivity of mAb AT-8 to another protein present in CSF. Therefore, the data should not be over interpreted for possible differences in the tau patterns of control and AD CSF. This result clearly shows that it is possible, however, to probe the phosphorylation pattern of CSF-tau with phosphorylation-dependent mAbs in 2D-immunoblots. Any differences in the spot numbers and spot intensities could also be related to individual differences among the human population, and do not necessarily indicate disease specific alterations. This could be especially true, as the brain samples and the postmortem CSF were provided from different hospitals and probably had different postmortem delay times. It was not possible to detect tau at quantities below 700 pg with reasonable spot intensities above the background level. At this level it was also difficult, and often impossible, to distinguish the tau spots from cross-reactive proteins. Thus, a tau load of 700 pg appears to be the lower detection limit. Ideally, tau quantities loaded onto the 2-DE should be between 1 and 2 ng to ensure identification of most spots of the tau pattern and to distinguish them clearly from the background and cross-reactive proteins.

Analysis of Human Tau in Cerebrospinal Fluid

Conclusion The presented strategy to deplete CSF samples for dominant proteins, and to detect tau in an immunoblot with mAb HT7POD following separation with 2-DE, will enable the study of circulating tau versions in lumbar CSF samples. In addition, it will allow these tau versions to be characterized with respect to the distribution of the splicing forms and their phosphorylation degree. This strategy will, therefore, provide a better understanding of the clearance processes of tau in both healthy and diseased tissue. It may even provide an understanding of the underlying clearance processes for tau in CSF. Furthermore, it may give insight into the changes in the phosphorylation pattern during normal aging or disease development. Different spot patterns seen in different tauopathies might hint toward the underlying pathogenetic mechanisms. Additional information may be obtained from phosphorylation-dependent antibodies, as shown here for mAb AT-8, that would show which phosphorylated tau versions exist in CSF, and whether they can be correlated to the development or progression of AD or other tauopathies. Abbreviations: 2-DE, two-dimensional gel electrophoresis; Aβ, amyloid peptide; AD, Alzheimer’s disease; APP, amyloid precursor protein; CNS, central nervous system; CSF, Cerebrospinal fluid; 2-DE, two-dimensional gel electrophoresis; DTT, dithiothreitol; ECL, enhanced chemiluminescence; ELISA, enzyme-linked immunosorbent assays; HRP, horseradish peroxidase; HSA, human serum albumin; IEF, isoelectric focusing; IPG, immobilized pH-Gradient, mAb, monoclonal antibody; MARS, multiple affinity removal system; NFT, neurofibrillary tangles; NP, neuritic plaques; PAGE, polyacrylamide gel electrophoresis; PCR, polymerase chain reaction; PHF, paired helical filaments; POD, peroxidase; SDS, sodium dodecylsulfate; Tau, microtubule-associated tau protein; TBS, tris buffered saline; TTBS, 0.1% Tween 20 in TBS.

Acknowledgment. This work was funded by the European Fund for Regional Structure Development (EFRE) by the European Union and the Free State Saxony. We thank Dr. Schulz-Scha¨ffer for providing human brain samples, Dr. Hans-Ju ¨ rgen Ku ¨ hn for providing control CSF samples, Dr. Daniela Volke for preparation of human tau, Robert Porzig for synthesizing the tau peptide, and Dr. Christina Nielsen-Marsh for proof reading the manuscript. Supporting Information Available: Figures S1-3 and Tables S1 and S2. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Hu ¨hmer, A. F.; Biringer, R. G.; Amato, H.; Fonteh, A. N.; Harrington, M. G. Protein analysis in human cerebrospinal fluid: Physiological aspects, current progress and future challenges. Disease Markers 2006, 22, 3–26. (2) Reiber, H.; Peter, J. B. Cerebrospinal fluid analysis: disease-related data patterns and evaluation programs. J. Neurol. Sci. 2001, 184 (2), 101–22. (3) Yuan, X.; Desiderio, D. M. Proteomics analysis of prefractionated human lumbar cerebrospinal fluid. Proteomics 2005, 5 (2), 541– 50. (4) McKhann, G.; et al. 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 (7), 939–44. (5) Wolozin, B.; Davies, P. Alzheimer-related neuronal protein A68: specificity and distribution. Ann. Neurol. 1987, 22 (4), 521–6. (6) Palmert, M. R.; et al. Soluble derivatives of the beta amyloid protein precursor in cerebrospinal fluid: alterations in normal aging and in Alzheimer’s disease. Neurology 1990, 40 (7), 1028–34.

research articles (7) Sennvik, K.; et al. Levels of alpha- and beta-secretase cleaved amyloid precursor protein in the cerebrospinal fluid of Alzheimer’s disease patients. Neurosci. Lett. 2000, 278 (3), 169–72. (8) Hampel, H.; Goernitz, A.; Buerger, K. Advances in the development of biomarkers for Alzheimer’s disease: from CSF total tau and Abeta(1-42) proteins to phosphorylated tau protein. Brain Res. Bull. 2003, 61 (3), 243–53. (9) Couchie, D.; et al. Primary structure of high molecular weight tau present in the peripheral nervous system. Proc. Natl. Acad. Sci. U.S.A. 1992, 89 (10), 4378–81. (10) Goedert, M.; Spillantini, M. G.; Crowther, R. A. Cloning of a big tau microtubule-associated protein characteristic of the peripheral nervous system. Proc. Natl. Acad. Sci. U.S.A. 1992, 89 (5), 1983–7. (11) Goedert, M.; et al. Multiple isoforms of human microtubuleassociated protein tau: sequences and localization in neurofibrillary tangles of Alzheimer’s disease. Neuron 1989, 3 (4), 519–26. (12) Goedert, M. Tau protein and the neurofibrillary pathology of Alzheimer’s disease. Trends Neurosci. 1993, 16 (11), 460–5. (13) Cleveland, D. W.; Hwo, S. Y.; Kirschner, M. W. Physical and chemical properties of purified tau factor and the role of tau in microtubule assembly. J. Mol. Biol. 1977, 116 (2), 227–47. (14) Lindwall, G.; Cole, R. D. Phosphorylation affects the ability of tau protein to promote microtubule assembly. J. Biol. Chem. 1984, 259 (8), 5301–5. (15) Khatoon, S.; Grundke-Iqbal, I.; Iqbal, K. Brain levels of microtubule-associated protein tau are elevated in Alzheimer’s disease: a radioimmuno-slot-blot assay for nanograms of the protein. J. Neurochem. 1992, 59 (2), 750–3. (16) Vigo-Pelfrey, C.; et al. Elevation of microtubule-associated protein tau in the cerebrospinal fluid of patients with Alzheimer’s disease. Neurology 1995, 45 (4), 788–93. (17) Andreasen, N.; et al. Cerebrospinal fluid tau protein as a biochemical marker for Alzheimer’s disease: a community based follow up study. J. Neurol. Neurosurg. Psychiatry 1998, 64 (3), 298–305. (18) Vandermeeren, M.; 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 (5), 1828–34. (19) Blennow, K.; et al. , Tau protein in cerebrospinal fluid: a biochemical marker for axonal degeneration in Alzheimer disease. Mol. Chem. Neuropathol. 1995, 26 (3), 231–45. (20) Davies, L.; 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 (11), 1688–93. (21) Blennow, K.; Hampel, H. CSF markers for incipient Alzheimer’s disease. Lancet Neurol. 2003, 2 (10), 605–13. (22) Hampel, H.; et al. Core biological marker candidates of Alzheimer’s disease - perspectives for diagnosis, prediction of outcome and reflection of biological activity. J. Neural Transm. 2004, 111 (3), 247–72. (23) Tapiola, T.; et al. Cerebrospinal fluid {beta}-amyloid 42 and tau proteins as biomarkers of Alzheimer-type pathologic changes in the brain. Arch Neurol. 2009, 66 (3), 382–9. (24) Johnson, G. V.; Jope, R. S.; Binder, L. I. Proteolysis of tau by calpain. Biochem. Biophys. Res. Commun. 1989, 163 (3), 1505–11. (25) Singer, D.; et al. Neighbored phosphorylation sites as PHF-tau specific markers in Alzheimer’s disease. Biochem. Biophys. Res. Commun. 2006, 346 (3), 819–28. (26) Singer, D., et al., Immuno-PCR-based quantification of multiple phosphorylated tau-epitopes linked to Alzheimer’s disease. Anal. Bioanal. Chem., 2009. (27) Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970, 227 (5259), 680–5. (28) Porstmann, B.; et al. Which of the commonly used marker enzymes gives the best results in colorimetric and fluorimetric enzyme immunoassays: horseradish peroxidase, alkaline phosphatase or beta-galactosidase. J. Immunol. Methods 1985, 79 (1), 27–37. (29) Imagawa, M. Characteristics and evaluation of antibody- horseradish peroxidase conjugates prepared by using a maleimide compound, glutaraldehyde and periodate. J. Appl. Biochem. 1982, 4, p. 41–57. (30) Nakane, P. K.; Kawaoi, A. Peroxidase-labeled antibody. A new method of conjugation. J. Histochem. Cytochem. 1974, 22 (12), 1084–91. (31) Singer, D.; Volke, D.; Hoffmann, R. Characterization of Phosphorylation Dependent Antibodies To Study the Phosphorylation Status of the Tau Protein. Int. J. Pept. Res. Ther. 2005, 11 (4), 279– 289. (32) Volke, D.; Hoffmann, R. Purification of bovine Tau versions by affinity chromatography. Protein Expr. Purif. 2006, 50 (1), 37–42.

Journal of Proteome Research • Vol. 9, No. 3, 2010 1481

research articles (33) Janke, C.; et al. Analysis of the molecular heterogeneity of the microtubule-associated protein tau by two-dimensional electrophoresis and RT-PCR. Brain Res. Brain Res. Protoc. 2000, 5 (3), 231–42. (34) Volke, D.; Hoffmann, R. Characterization of bovine tau-preparations by two-dimensional gel electrophoresis. Protein Pept. Lett. 2006, 13 (6), 617–25. (35) Duchesne, L.; Fernig, D. G. Silver and gold nanoparticle-coated membranes for femtomole detection of small proteins and peptides by Dot and Western blot. Anal. Biochem. 2007, 362 (2), 287–9.

1482

Journal of Proteome Research • Vol. 9, No. 3, 2010

Hanisch et al. (36) Echan, L. A.; et al. Depletion of multiple high-abundance proteins improves protein profiling capacities of human serum and plasma. Proteomics 2005, 5 (13), 3292–303. (37) Porzig, R.; Singer, D.; Hoffmann, R. Epitope mapping of mAbs AT8 and Tau5 directed against hyperphosphorylated regions of the human tau protein. Biochem. Biophys. Res. Commun. 2007, 358 (2), 644–9. (38) Anderson, N. L.; Anderson, N. G. The human plasma proteome: history, character, and diagnostic prospects. Mol. Cell. Proteomics 2002, 1 (11), 845–67.

PR901002T