- Email: [email protected]
Quantification of tau phosphorylated at threonine 181 in human cerebrospinal fluid: a sandwich ELISA with a synthetic phosphopeptide for standardization
Neuroscience Letters 285 (2000) 49±52
www.elsevier.com/locate/neulet
Quanti®cation of tau phosphorylated at threonine 181 in human cerebrospinal ¯uid: a sandwich ELISA with a synthetic phosphopeptide for standardization E. Vanmechelen a,*, H. Vanderstichele a, P. Davidsson b, E. Van Kerschaver a, B. Van Der Perre a, M. SjoÈgren c, N. Andreasen d, K. Blennow b, e a Innogenetics NV, Industriepark Zwijnaarde 7, Box 4, B-9052 Gent, Belgium Department of Clinical Neuroscience, Unit of Neurochemistry, University of GoÈteborg, Sahlgren's University Hospital, MoÈlndal, Sweden c Department of Clinical Neuroscience, Unit of Psychiatry, University of GoÈteborg, Sahlgren's University Hospital, MoÈlndal, Sweden d Department of Rehabilitation, Pitea River Valley Hospital, Pitea, Sweden e The Medical Research Council, Sweden
b
Received 20 January 2000; received in revised form 17 March 2000; accepted 17 March 2000
Abstract Hyperphosphorylation of the microtubule-associated protein tau is speci®cally found in those brain cells affected in several tauopathies. Tau has also been consistently found to be present in the cerebrospinal ¯uid (CSF). Here we report the quanti®cation in CSF of tau phosphorylated at Thr 181 using an immunoassay with a synthetic peptide for standardization. The choice of the peptide was based on ®ne mapping of a phospho-dependent antibody, AT270 (P176PAPKTpP132)and a human speci®c tau antibody, HT7 (P159PGQK163). CSF-phospho-tau levels were increased in Alzheimer patients (23.5 ^ 10.1 pM, P , 0:01) compared with age-matched controls (15.9 ^ 5.7 pM), while decreased in patients with frontotemporal dementia (8.6 ^ 3.9 pM; P , 0:01). In every diagnostic group, a highly signi®cant correlation was found between total tau and phospho-tau (Alzheimer's disease, r 2 0:73; frontotemporal dementia, r 2 0:43; Control, r 2 0:42), suggesting that the degree of phosphorylation of CSF-tau changes in different clinical conditions. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Biological markers; Tauopathies; Cerebrospinal ¯uid; Tau phosphorylation; Phosphopeptides; Immunoassay
The role of tau hyperphosphorylation in the pathology of several age-related neurodegenerative conditions such as Alzheimer's disease (AD), frontotemporal dementia (FTD) and other so-called tauopathies is presently not well understood. On the one hand, studies in AD strongly suggest that tau phosphorylation is an early event [2] and that abnormal tau phosphorylation in tauopathies is restricted to those brain cells/regions that are affected [4]. On the other hand, in vitro studies have shown that hyperphosphorylation of tau does not lead to precipitation of tau per se. In fact, polyanions such as heparin are able to induce aggregation in the complete absence of tau phosphorylation [7]. Circulating concentrations of tau in cerebrospinal ¯uid * Corresponding author. Tel.: 132-9-2410711; fax: 132-92410907. E-mail address: [email protected] (E. Vanmechelen).
(CSF) are low and, thus, phospho-tau is only quanti®able by sensitive immunoassays using phospho-speci®c antibodies. Previously, we have shown that CSF-tau is phosphorylated, and that absolute phospho-tau levels were 3- to 4-fold higher than total tau levels; in vitro phosphorylated recombinant tau was used for standardization [1]. Since it is dif®cult to determine accurately the degree of phosphorylation of speci®c phospho-sites concentrated in the proline-rich region of tau, we decided to use a synthetic phospho-peptide for standardization. Here we report the development of a sandwich ELISA speci®c for a proline-directed phosphosite, Thr 181, and its application in CSF of patients with two different tauopathies: AD and FTD. Two polymerase chain reaction (PCR) primers (a primer contain the starting methionine codon -CATGGCTGAGCCCCGCCAGGAGTTCGAAGTGATGG (21 to 34) and the reverse primer around the stop codon CCTGATCACAAACCCTGCTTGGCCAGGGAGGC) were used to amplify the smallest form of human tau from human brain
0304-3940/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 0) 01 03 6- 3
50
E. Vanmechelen et al. / Neuroscience Letters 285 (2000) 49±52
cDNA (Clontech, Westburg, Leusden, Netherlands). The identity of the PCR product was con®rmed by sequencing. Changes were only observed at the third basepair in codons: at Pro 182 CCG instead of CCA, Ala 227 GCG instead of GCA, and Asn 251 AAC instead of AAT. All numbering with respect to the amino acid sequence refers to the longest tau isoform: hTau40 [9]. The PCR fragment was inserted into a PL-based expression system for mapping studies of the antibodies on Escherichia coli-expressed tau proteins. Frameshift mutants of tau protein were constructed based on ®lling in the unique SacII sites (amino acid position 154± 155) and PstI site (position 242±243) that allows the positioning of the epitopes of antibodies to three regions: the Nterminal region (1±154), the proline-rich region (155±242) and the C-terminal region, containing the repeat domain and the C-tail (243±441). Automated peptide synthesis was performed on a Millipore 9050 synthesizer, usually as N-terminally biotinylated peptides. Phosphorylated peptides were synthesized using phosphorylated Thr derivatives (N-a-O-benzyl-l-phosphoThr; Calbiochem-Novabiochem AG, San Diego, CA). Large-scale synthesis of peptide Ac-P154RGAAPPGQKGQANATRIPAKTPPAPKT(p)PPSSGE187-NH2 for sandwich ELISA was done in-house and by Neosystems (Strasbourg, France). After detachment from the solid support, peptides and phospho-peptides were puri®ed by reverse-phase highpressure liquid chromatography (RP-HPLC). Quality control includes RP-HPLC (.99% pure), mass spectrometry analysis (average MW 3454.8), and amino acid analysis (net peptide content 84.3%). For epitope mapping, peptides were synthesized manually on derivatized pins (Multiple Peptide Systems, San Diego, CA) or on paper. Details of the isolation and characterization of antibodies have been described for AT120 [16], HT7 [14], BT2, and AT270 [6]. To quantify peptide antibody interactions by capturing assay, streptavidin (Boehringer Mannheim, Brussels, Belgium) was coated on immunoplates (Nunc, Roskilde, Denmark) and biotinylated peptides were added after blocking. A second antibody coupled to horseradish peroxidase was used to quantify the monoclonal antibodies bound to the peptide. The research version of the INNOTEST phospho-tau (181P) was designed essentially in the same format as the INNOTEST hTau-Ag (Innogenetics) [1,16]. Exact concentrations of recombinant tau (calculated molecular weight 41065) and synthetic peptides were determined using amino acid analysis [1]. PHF-tau was selectively precipitated from brain extracts using 1% Nlauroylsarcosinate [8]. Tissue from temporal cortex (PHFA-D) or hippocampus (PHF-E) of AD patients was obtained from the Born-Bunge Brain Bank (Dr P. Cras, Antwerp, Belgium). Clinical diagnosis and CSF sample-pretreatment were carried out as described previously in detail [1]. A total of 41 patients (13 males/28 females; mean age 74 ^ 6 years) with a diagnosis of probable AD were included. They did not differ demographically from the non-neurological
control group (n 17, age 72 ^ 4 years; four male/13 female). The 18 FTD patients (®ve male/13 female) were, however, signi®cantly younger than the control group (66 ^ 8 years). For statistical comparisons between two groups, the Mann±Whitney U-test was used. Data are expressed as mean ^ SD. The Spearman rank correlation coef®cient was used for correlations. In order to map those tau antibodies recognizing recombinant tau to regions of the tau molecule, speci®c deletion mutants, based on the unique restriction sites SacII and PstI, were constructed in a PL-expression system and produced in E. coli. The antibodies tested, i.e. HT7, AT120, BT2 and tau1, all map to the proline-rich region (position 154±242 on hTau40, results not shown). To further delineate the epitopes, small overlapping peptides were synthesized. In a ®rst step, 48 peptides, nine amino acids long and overlapping eight amino acids were synthesized on pins. The sequence ranged from 155 until 208. AT120, HT7, BT2 and tau1 were tested. Three of the four antibodies could be mapped: the minimal epitope of HT7 was P159PGQK163, while the reactivities of BT2 and tau-1 were indistinguishable: D193 RSGYS198 (Fig. 1a). Since AT120 could not be mapped on these peptides, a new set of peptides were synthesized on paper, covering the sequence from 206±232. A total of 16 peptides, 12 amino acids long, and overlapping with 11 amino acids, were needed to cover this region. The minimal epitope of AT120 was de®ned by the sequence P218PTREPK224 (Fig. 1b). The speci®city of the phospho-dependent antibody,
Fig. 1. Fine mapping of tau antibodies, HT7, BT2, AT120 and AT270 on overlapping synthetic peptides. Only immunoreactive peptides are shown. (A) Mapping on peptides synthesized on pins. Forty-eight nonapeptides overlapping eight amino acids were used to cover the tau region from 155 to 208. (B) Peptide synthesis on paper. Sixteen overlapping peptides, 12 amino acids long de®ne the AT120 epitope in the region 206±232. (C) Mapping of the AT270 antibody on biotinylated phosphopeptides covering the region 166 until 196.
E. Vanmechelen et al. / Neuroscience Letters 285 (2000) 49±52
AT270, was con®rmed on synthetic phospho-peptides covering 22 phosphorylation sites on tau. The sequences of these peptides are summarized in Table 1. Non-phosphorylated peptides corresponding to 12 of these sites were analyzed in parallel (Table 1). AT270 only reacted with phospho-peptides containing phospho-Thr 181 or phospho-Thr 175. Since in this capturing assay there is an excess amount of peptides, the speci®city of AT270 was further de®ned by dilution of the reacting phospho-peptides (phospho-Thr 181 and phospho-Thr 175). AT270 was 18fold less reactive, on a molar basis, with the peptide containing phospho-Thr 175 as compared with phospho-Thr 181 (results not shown). Finally, the minimal epitope was de®ned using biotinylated phosphorylated peptides, 15 amino acids-long, covering the region 166±196. Immunoreactive peptides are shown in Fig. 1c, and the minimal epitope of AT270 was P176PAPKT(p)P182. Using this peptide information, a phospho-peptide was synthesized covering the epitopes of HT7 (159±163) and AT270 (176±182), and an additional ®ve amino acids Nand C-terminal to these epitopes. The range of the ELISA is between 5 and 300 pM, and intra-assay and inter-assay coef®cients of variation were below 10%. To determine how the degree of phosphorylation relates to absolute levels of phospho-tau (181) and total tau, ®ve different PHF-tau preparations were simultaneously quanti®ed for total tau and phospho-tau. Absolute levels of phospho-tau were always higher than total tau in the PHF-tau preparations as shown in Table 2. Levels of phospho-tau (181) and total tau were determined Table 1 Sequences of phosphorylated peptides used to determine speci®city of AT270 for phospho-Thr 181 Name
Sequence
Non-phospho peptide
153 175 181 198 199 202 205 208 210 212 214 217 231 235 262 396 400 403 404 409 412 422
KGADGKTKIATpPRGAAPPGQK QANATRIAPKTpPPAPKTPPSS RIPAKTPPAPKTpPPSSGEPPKS PPKSGDRSGYSPSGSPGTPGSR PKSGDRSGYSSpGSPGTPGSRS SGDRSGYSSGSPPGTPGSRSRT RSGYSSGSPGTPPGSRSRTPSL YSSGSPGTPGSPRSRTPSLPTP SGSPGTPGSRSPRTPSLPTPTR SPGTPGSRSRTPPSLPTPTREP GTPGSRSRTPSPLPTPTREPKK GSRSRTPSLPTPPTREPKKVAV REPKKVAVVRTPPKSPSSAKS KKVAVVRTPKSPPSSAKSRLQ VKSKIGSPTENLK TDHGAEIVYKSPPVVSDTSPRH AEIVYKSPVVSPDTSPRHLSNV IVYKSPVVSDTPSPRHLSNVSS YKSPVVSDTSPPRHLSNVSST VVSDTSPRHLSPNVSSTGSIDM DTSPRHLSNVSPSTGSIDMVDS SSTGSIDMVDSPPQLATLADEV
1 1 1
1 1 1 1 1 1 1 1 1
51
in AD and FTD patients, and compared with age-matched controls. Highly signi®cant differences in tau levels were found between AD (20.1 ^ 7.6 pM; P , 0:01) and controls (8.3 ^ 2.8 pM), whereas in FTD patients CSF-tau levels were normal (9.3 ^ 2.9 pM). The CSF-tau levels in AD were signi®cantly higher than in the FTD group (P , 0:01). CSF-phospho-tau levels were markedly reduced in FTD (8.6 ^ 3.9 pM; P , 0:01) compared with controls (15.9 ^ 5.7 pM), while increased in AD (23.5 ^ 10.1 pM, P , 0:01). In all groups, a highly statistical relevant correlation was found between tau and phospho-tau (AD: y 1.14x 1 0.24, r 2 0:73; FTD: y 0.89x 2 0.089, r 2 0:43; Control: y 1.31x 1 4.95, r2 0:42). Tau circulating in CSF is phosphorylated. To determine how the degeneration of brain cells containing hyperphosphorylated tau in AD is re¯ected in the CSF, tau phosphorylated at Thr 181 was quanti®ed using an immunoassay. This particular site was selected because (1) it is a relatively isolated phosphorylation site in the proline-rich region, (2) it is preferentially phosphorylated by proline-directed kinases, and (3) the size of the synthetic peptide used for standardization is small. The design of this peptide was based on ®ne-mapping of phospho-tau and tau antibodies, recognizing all tau isoforms. The minimal epitope of the phospho-dependent antibody, AT270, was de®ned as P176PAPKT(p)P182 and its speci®city was con®rmed on 22 different phospho-peptides, including neuro®lament crossreactive sites, such as Thr 231 and Ser 396. The epitope of AT120 lies in a region where other antibodies have been mapped, such as tau5, TG5 [3] and 5E2 [12]. HT7, an antibody that does not cross-react with rat or mouse tau [14], maps to the same region as another human-speci®c antibody, tau-14 [10,12]. Finally the minimal epitope of BT2 and tau-1, speci®c for non-phosphorylated tau [14], are identical. Optimal recognition of BT2/tau-1 for their minimal epitope, however, requires a C-terminal tail [5]. Absolute levels of phospho-tau were clearly higher than total tau levels (Table 2). Therefore, additional studies are required to de®ne how levels in immunoassays are related to the phosphorylation status of the tau protein. Taking into account that PHF-tau is phosphorylated close to 100% [6] and PHF-tau preparation B has the highest ratio of phosphotau to total tau (Table 2), the amount of phospho-tau is 4.3 Table 2 Levels of phospho-tau and total tau as determined by INNOTEST phospho-tau (181P) and INNOTEST hTau-Ag. PHF-tau was prepared from human temporal cortex (PHF-A-D) and hippocampal formation (PHF-E) according to Goedert et al. [8] and dissolved according to the amount of tissue used for isolation PHF-tau prep
Phospho-tau (181P) (pM)
Total tau (pM)
PHF-A PHF-B PHF-C PHF-D PHF-E
101 108 61 137 219
36 25 18 35 87
52
E. Vanmechelen et al. / Neuroscience Letters 285 (2000) 49±52
higher than the level of total tau. Using this factor as a correction, the degree of phosphorylation of CSF-tau in controls can be estimated: 15.9/(8.3 £ 4.3) 45%. This is in close agreement with the phosphorylation status of threonine 181 in human or rat brain-derived tau under normal conditions (20±30% [13,17]). Levels of CSF-phospho-tau are signi®cantly increased in AD, while decreased in FTD patients. Since CSF-phosphotau is closely correlated to total CSF-tau levels, it suggests that the degree of phosphorylation is differentially affected in several clinical conditions, as suggested [11]. A standardized assay for phospho-tau, as presented here, is essential for further validation of phospho-tau as a potential biomarker for tau pathology in different tauopathies, such as AD, FTD, corticobasal degeneration, progressive supranuclear palsy and others. Since Lewy body dementia requires expert treatment management [15] and is a special form of AD without tau pathology, phospho-tau as a biomarker might have therapeutic implications. This work was supported by a grant from the `Vlaams Instituut voor de bevordering van het Wetenschappelijk Technologisch onderzoek in de industrie' (Grant 990220) and the Medical Research Council, Sweden, Grant No. 12103 and 11560. Fred Shapiro is thanked for critical review and editing of the manuscript and Bob DeLeys for the initial experiments on the epitope mapping. [1] Blennow, K., Wallin, A., Agren, H., Spenger, C., Siegfried, J. and Vanmechelen, E., Tau protein in cerebrospinal ¯uid: a biochemical marker for axonal degeneration in Alzheimer disease? Mol. Chem. Neuropathol., 26 (1995) 231±245. [2] Braak, E., Braak, H. and Mandelkow, E.M., A sequence of cytoskeleton changes related to the formation of neuro®brillary tangles and neuropil threads, Acta Neuropathol. (Berlin), 87 (1994) 554±567. [3] Carmel, G., Mager, E.M., Binder, L.I. and Kuret, J., The structural basis of monoclonal antibody Alz50's selectivity for Alzheimer's disease pathology, J. Biol. Chem., 271 (1996) 32789±32795. [4] Delacourte, A., Sergeant, N., Wattez, A., Gauvreau, D. and Robitaille, Y., Vulnerable neuronal subsets in Alzheimer's and Pick's disease are distinguished by their tau isoform distribution and phosphorylation, Ann. Neurol., 43 (1998) 193±204. [5] DeLeys, R., Hendrickx, G., Dekeyser, F., Demol, H., Raymackers, J., Brasseur, R., Borremans, F., Vanmechelen, E. and Van de Voorde, A., Mapping and sequence requirements of the phosphorylation-sensitive epitopes recognized by the monoclonal antibodies Tau1, BT2, and AT8, In C.H. Schneider (Ed.), Peptides in Immunology, Wiley, New York, 1996, pp. 239±244.
[6] Goedert, M., Jakes, R., Crowther, R.A., Cohen, P., Vanmechelen, E., Vandermeeren, M. and Cras, P., Epitope mapping of monoclonal antibodies to the paired helical ®laments of Alzheimer's disease: identi®cation of phosphorylation sites in tau protein, Biochem. J., 301 (1994) 871±877. [7] Goedert, M., Jakes, R., Spillantini, M.G., Hasegawa, M., Smith, M.J. and Crowther, R.A., Assembly of microtubuleassociated protein tau into Alzheimer-like ®laments induced by sulphated glycosaminoglycans, Nature, 383 (1996) 550±553. [8] Goedert, M., Spillantini, M.G., Cairns, N.J. and Crowther, R.A., Tau proteins of Alzheimer paired helical ®laments: abnormal phosphorylation of all six brain isoforms, Neuron, 8 (1992) 159±168. [9] Goedert, M., Spillantini, M.G., Jakes, R., Rutherford, D. and Crowther, R.A., Multiple isoforms of human microtubuleassociated protein tau: sequences and localization in neuro®brillary tangles of Alzheimer's disease, Neuron, 3 (1989) 519±526. [10] GoÈtz, J., Probst, A., Spillantini, M.G., SchaÈfer, T., Jakes, R., Byne, W. and Goedert, M., Somatodendritic localization and hyperphosphorylation of tau protein in transgenic mice expressing the longest human brain tau isoform, EMBO J., 14 (1995) 1304±1313. [11] Ishiguro, K., Ohno, H., Arai, H., Yamaguchi, H., Urakami, K., Park, J.M., Sato, K., Kohno, H. and Imahori, K., Phosphorylated tau in human cerebrospinal ¯uid is a diagnostic marker for Alzheimer's disease, Neurosci. Lett., 270 (1999) 91±94. [12] Kosik, K.S., Orecchio, L.D., Binder, L., Trojanowski, J.Q., Lee, V.M. and Lee, G., Epitopes that span the tau molecule are shared with paired helical ®laments, Neuron, 1 (1988) 817±825. [13] Matsuo, E.S., Shin, R.W., Billingsley, M.L., Van deVoorde, A., O'Connor, M., Trojanowski, J.Q. and Lee, V.M., Biopsyderived adult human brain tau is phosphorylated at many of the same sites as Alzheimer's disease paired helical ®lament tau, Neuron, 13 (1994) 989±1002. [14] Mercken, M., Vandermeeren, M., Lubke, U., Six, J., Boons, J., Vanmechelen, E., Van de Voorde, A. and Gheuens, J., Af®nity puri®cation of human tau proteins and the construction of a sensitive sandwich enzyme-linked immunosorbent assay for human tau detection, J. Neurochem., 58 (1992) 548±553. [15] Rojas-Fernandez, C.H. and MacKnight, C., Dementia with Lewy bodies: review and pharmacotherapeutic implications, Pharmacotherapy, 19 (1999) 795±803. [16] Vandermeeren, M., Mercken, M., Vanmechelen, E., Six, J., Van de Voorde, A., Martin, J.J. and Cras, P., Detection of tau proteins in normal and Alzheimer's disease cerebrospinal ¯uid with a sensitive sandwich enzyme-linked immunosorbent assay, J. Neurochem., 61 (1993) 1828±1834. [17] Watanabe, A., Hasegawa, M., Suzuki, M., Takio, K., Morishima-Kawashima, M., Titani, K., Arai, T., Kosik, K.S. and Ihara, Y., In vivo phosphorylation sites in fetal and adult rat tau, J. Biol. Chem., 268 (1993) 25712±25717.
www.elsevier.com/locate/neulet
Quanti®cation of tau phosphorylated at threonine 181 in human cerebrospinal ¯uid: a sandwich ELISA with a synthetic phosphopeptide for standardization E. Vanmechelen a,*, H. Vanderstichele a, P. Davidsson b, E. Van Kerschaver a, B. Van Der Perre a, M. SjoÈgren c, N. Andreasen d, K. Blennow b, e a Innogenetics NV, Industriepark Zwijnaarde 7, Box 4, B-9052 Gent, Belgium Department of Clinical Neuroscience, Unit of Neurochemistry, University of GoÈteborg, Sahlgren's University Hospital, MoÈlndal, Sweden c Department of Clinical Neuroscience, Unit of Psychiatry, University of GoÈteborg, Sahlgren's University Hospital, MoÈlndal, Sweden d Department of Rehabilitation, Pitea River Valley Hospital, Pitea, Sweden e The Medical Research Council, Sweden
b
Received 20 January 2000; received in revised form 17 March 2000; accepted 17 March 2000
Abstract Hyperphosphorylation of the microtubule-associated protein tau is speci®cally found in those brain cells affected in several tauopathies. Tau has also been consistently found to be present in the cerebrospinal ¯uid (CSF). Here we report the quanti®cation in CSF of tau phosphorylated at Thr 181 using an immunoassay with a synthetic peptide for standardization. The choice of the peptide was based on ®ne mapping of a phospho-dependent antibody, AT270 (P176PAPKTpP132)and a human speci®c tau antibody, HT7 (P159PGQK163). CSF-phospho-tau levels were increased in Alzheimer patients (23.5 ^ 10.1 pM, P , 0:01) compared with age-matched controls (15.9 ^ 5.7 pM), while decreased in patients with frontotemporal dementia (8.6 ^ 3.9 pM; P , 0:01). In every diagnostic group, a highly signi®cant correlation was found between total tau and phospho-tau (Alzheimer's disease, r 2 0:73; frontotemporal dementia, r 2 0:43; Control, r 2 0:42), suggesting that the degree of phosphorylation of CSF-tau changes in different clinical conditions. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Biological markers; Tauopathies; Cerebrospinal ¯uid; Tau phosphorylation; Phosphopeptides; Immunoassay
The role of tau hyperphosphorylation in the pathology of several age-related neurodegenerative conditions such as Alzheimer's disease (AD), frontotemporal dementia (FTD) and other so-called tauopathies is presently not well understood. On the one hand, studies in AD strongly suggest that tau phosphorylation is an early event [2] and that abnormal tau phosphorylation in tauopathies is restricted to those brain cells/regions that are affected [4]. On the other hand, in vitro studies have shown that hyperphosphorylation of tau does not lead to precipitation of tau per se. In fact, polyanions such as heparin are able to induce aggregation in the complete absence of tau phosphorylation [7]. Circulating concentrations of tau in cerebrospinal ¯uid * Corresponding author. Tel.: 132-9-2410711; fax: 132-92410907. E-mail address: [email protected] (E. Vanmechelen).
(CSF) are low and, thus, phospho-tau is only quanti®able by sensitive immunoassays using phospho-speci®c antibodies. Previously, we have shown that CSF-tau is phosphorylated, and that absolute phospho-tau levels were 3- to 4-fold higher than total tau levels; in vitro phosphorylated recombinant tau was used for standardization [1]. Since it is dif®cult to determine accurately the degree of phosphorylation of speci®c phospho-sites concentrated in the proline-rich region of tau, we decided to use a synthetic phospho-peptide for standardization. Here we report the development of a sandwich ELISA speci®c for a proline-directed phosphosite, Thr 181, and its application in CSF of patients with two different tauopathies: AD and FTD. Two polymerase chain reaction (PCR) primers (a primer contain the starting methionine codon -CATGGCTGAGCCCCGCCAGGAGTTCGAAGTGATGG (21 to 34) and the reverse primer around the stop codon CCTGATCACAAACCCTGCTTGGCCAGGGAGGC) were used to amplify the smallest form of human tau from human brain
0304-3940/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 0) 01 03 6- 3
50
E. Vanmechelen et al. / Neuroscience Letters 285 (2000) 49±52
cDNA (Clontech, Westburg, Leusden, Netherlands). The identity of the PCR product was con®rmed by sequencing. Changes were only observed at the third basepair in codons: at Pro 182 CCG instead of CCA, Ala 227 GCG instead of GCA, and Asn 251 AAC instead of AAT. All numbering with respect to the amino acid sequence refers to the longest tau isoform: hTau40 [9]. The PCR fragment was inserted into a PL-based expression system for mapping studies of the antibodies on Escherichia coli-expressed tau proteins. Frameshift mutants of tau protein were constructed based on ®lling in the unique SacII sites (amino acid position 154± 155) and PstI site (position 242±243) that allows the positioning of the epitopes of antibodies to three regions: the Nterminal region (1±154), the proline-rich region (155±242) and the C-terminal region, containing the repeat domain and the C-tail (243±441). Automated peptide synthesis was performed on a Millipore 9050 synthesizer, usually as N-terminally biotinylated peptides. Phosphorylated peptides were synthesized using phosphorylated Thr derivatives (N-a-O-benzyl-l-phosphoThr; Calbiochem-Novabiochem AG, San Diego, CA). Large-scale synthesis of peptide Ac-P154RGAAPPGQKGQANATRIPAKTPPAPKT(p)PPSSGE187-NH2 for sandwich ELISA was done in-house and by Neosystems (Strasbourg, France). After detachment from the solid support, peptides and phospho-peptides were puri®ed by reverse-phase highpressure liquid chromatography (RP-HPLC). Quality control includes RP-HPLC (.99% pure), mass spectrometry analysis (average MW 3454.8), and amino acid analysis (net peptide content 84.3%). For epitope mapping, peptides were synthesized manually on derivatized pins (Multiple Peptide Systems, San Diego, CA) or on paper. Details of the isolation and characterization of antibodies have been described for AT120 [16], HT7 [14], BT2, and AT270 [6]. To quantify peptide antibody interactions by capturing assay, streptavidin (Boehringer Mannheim, Brussels, Belgium) was coated on immunoplates (Nunc, Roskilde, Denmark) and biotinylated peptides were added after blocking. A second antibody coupled to horseradish peroxidase was used to quantify the monoclonal antibodies bound to the peptide. The research version of the INNOTEST phospho-tau (181P) was designed essentially in the same format as the INNOTEST hTau-Ag (Innogenetics) [1,16]. Exact concentrations of recombinant tau (calculated molecular weight 41065) and synthetic peptides were determined using amino acid analysis [1]. PHF-tau was selectively precipitated from brain extracts using 1% Nlauroylsarcosinate [8]. Tissue from temporal cortex (PHFA-D) or hippocampus (PHF-E) of AD patients was obtained from the Born-Bunge Brain Bank (Dr P. Cras, Antwerp, Belgium). Clinical diagnosis and CSF sample-pretreatment were carried out as described previously in detail [1]. A total of 41 patients (13 males/28 females; mean age 74 ^ 6 years) with a diagnosis of probable AD were included. They did not differ demographically from the non-neurological
control group (n 17, age 72 ^ 4 years; four male/13 female). The 18 FTD patients (®ve male/13 female) were, however, signi®cantly younger than the control group (66 ^ 8 years). For statistical comparisons between two groups, the Mann±Whitney U-test was used. Data are expressed as mean ^ SD. The Spearman rank correlation coef®cient was used for correlations. In order to map those tau antibodies recognizing recombinant tau to regions of the tau molecule, speci®c deletion mutants, based on the unique restriction sites SacII and PstI, were constructed in a PL-expression system and produced in E. coli. The antibodies tested, i.e. HT7, AT120, BT2 and tau1, all map to the proline-rich region (position 154±242 on hTau40, results not shown). To further delineate the epitopes, small overlapping peptides were synthesized. In a ®rst step, 48 peptides, nine amino acids long and overlapping eight amino acids were synthesized on pins. The sequence ranged from 155 until 208. AT120, HT7, BT2 and tau1 were tested. Three of the four antibodies could be mapped: the minimal epitope of HT7 was P159PGQK163, while the reactivities of BT2 and tau-1 were indistinguishable: D193 RSGYS198 (Fig. 1a). Since AT120 could not be mapped on these peptides, a new set of peptides were synthesized on paper, covering the sequence from 206±232. A total of 16 peptides, 12 amino acids long, and overlapping with 11 amino acids, were needed to cover this region. The minimal epitope of AT120 was de®ned by the sequence P218PTREPK224 (Fig. 1b). The speci®city of the phospho-dependent antibody,
Fig. 1. Fine mapping of tau antibodies, HT7, BT2, AT120 and AT270 on overlapping synthetic peptides. Only immunoreactive peptides are shown. (A) Mapping on peptides synthesized on pins. Forty-eight nonapeptides overlapping eight amino acids were used to cover the tau region from 155 to 208. (B) Peptide synthesis on paper. Sixteen overlapping peptides, 12 amino acids long de®ne the AT120 epitope in the region 206±232. (C) Mapping of the AT270 antibody on biotinylated phosphopeptides covering the region 166 until 196.
E. Vanmechelen et al. / Neuroscience Letters 285 (2000) 49±52
AT270, was con®rmed on synthetic phospho-peptides covering 22 phosphorylation sites on tau. The sequences of these peptides are summarized in Table 1. Non-phosphorylated peptides corresponding to 12 of these sites were analyzed in parallel (Table 1). AT270 only reacted with phospho-peptides containing phospho-Thr 181 or phospho-Thr 175. Since in this capturing assay there is an excess amount of peptides, the speci®city of AT270 was further de®ned by dilution of the reacting phospho-peptides (phospho-Thr 181 and phospho-Thr 175). AT270 was 18fold less reactive, on a molar basis, with the peptide containing phospho-Thr 175 as compared with phospho-Thr 181 (results not shown). Finally, the minimal epitope was de®ned using biotinylated phosphorylated peptides, 15 amino acids-long, covering the region 166±196. Immunoreactive peptides are shown in Fig. 1c, and the minimal epitope of AT270 was P176PAPKT(p)P182. Using this peptide information, a phospho-peptide was synthesized covering the epitopes of HT7 (159±163) and AT270 (176±182), and an additional ®ve amino acids Nand C-terminal to these epitopes. The range of the ELISA is between 5 and 300 pM, and intra-assay and inter-assay coef®cients of variation were below 10%. To determine how the degree of phosphorylation relates to absolute levels of phospho-tau (181) and total tau, ®ve different PHF-tau preparations were simultaneously quanti®ed for total tau and phospho-tau. Absolute levels of phospho-tau were always higher than total tau in the PHF-tau preparations as shown in Table 2. Levels of phospho-tau (181) and total tau were determined Table 1 Sequences of phosphorylated peptides used to determine speci®city of AT270 for phospho-Thr 181 Name
Sequence
Non-phospho peptide
153 175 181 198 199 202 205 208 210 212 214 217 231 235 262 396 400 403 404 409 412 422
KGADGKTKIATpPRGAAPPGQK QANATRIAPKTpPPAPKTPPSS RIPAKTPPAPKTpPPSSGEPPKS PPKSGDRSGYSPSGSPGTPGSR PKSGDRSGYSSpGSPGTPGSRS SGDRSGYSSGSPPGTPGSRSRT RSGYSSGSPGTPPGSRSRTPSL YSSGSPGTPGSPRSRTPSLPTP SGSPGTPGSRSPRTPSLPTPTR SPGTPGSRSRTPPSLPTPTREP GTPGSRSRTPSPLPTPTREPKK GSRSRTPSLPTPPTREPKKVAV REPKKVAVVRTPPKSPSSAKS KKVAVVRTPKSPPSSAKSRLQ VKSKIGSPTENLK TDHGAEIVYKSPPVVSDTSPRH AEIVYKSPVVSPDTSPRHLSNV IVYKSPVVSDTPSPRHLSNVSS YKSPVVSDTSPPRHLSNVSST VVSDTSPRHLSPNVSSTGSIDM DTSPRHLSNVSPSTGSIDMVDS SSTGSIDMVDSPPQLATLADEV
1 1 1
1 1 1 1 1 1 1 1 1
51
in AD and FTD patients, and compared with age-matched controls. Highly signi®cant differences in tau levels were found between AD (20.1 ^ 7.6 pM; P , 0:01) and controls (8.3 ^ 2.8 pM), whereas in FTD patients CSF-tau levels were normal (9.3 ^ 2.9 pM). The CSF-tau levels in AD were signi®cantly higher than in the FTD group (P , 0:01). CSF-phospho-tau levels were markedly reduced in FTD (8.6 ^ 3.9 pM; P , 0:01) compared with controls (15.9 ^ 5.7 pM), while increased in AD (23.5 ^ 10.1 pM, P , 0:01). In all groups, a highly statistical relevant correlation was found between tau and phospho-tau (AD: y 1.14x 1 0.24, r 2 0:73; FTD: y 0.89x 2 0.089, r 2 0:43; Control: y 1.31x 1 4.95, r2 0:42). Tau circulating in CSF is phosphorylated. To determine how the degeneration of brain cells containing hyperphosphorylated tau in AD is re¯ected in the CSF, tau phosphorylated at Thr 181 was quanti®ed using an immunoassay. This particular site was selected because (1) it is a relatively isolated phosphorylation site in the proline-rich region, (2) it is preferentially phosphorylated by proline-directed kinases, and (3) the size of the synthetic peptide used for standardization is small. The design of this peptide was based on ®ne-mapping of phospho-tau and tau antibodies, recognizing all tau isoforms. The minimal epitope of the phospho-dependent antibody, AT270, was de®ned as P176PAPKT(p)P182 and its speci®city was con®rmed on 22 different phospho-peptides, including neuro®lament crossreactive sites, such as Thr 231 and Ser 396. The epitope of AT120 lies in a region where other antibodies have been mapped, such as tau5, TG5 [3] and 5E2 [12]. HT7, an antibody that does not cross-react with rat or mouse tau [14], maps to the same region as another human-speci®c antibody, tau-14 [10,12]. Finally the minimal epitope of BT2 and tau-1, speci®c for non-phosphorylated tau [14], are identical. Optimal recognition of BT2/tau-1 for their minimal epitope, however, requires a C-terminal tail [5]. Absolute levels of phospho-tau were clearly higher than total tau levels (Table 2). Therefore, additional studies are required to de®ne how levels in immunoassays are related to the phosphorylation status of the tau protein. Taking into account that PHF-tau is phosphorylated close to 100% [6] and PHF-tau preparation B has the highest ratio of phosphotau to total tau (Table 2), the amount of phospho-tau is 4.3 Table 2 Levels of phospho-tau and total tau as determined by INNOTEST phospho-tau (181P) and INNOTEST hTau-Ag. PHF-tau was prepared from human temporal cortex (PHF-A-D) and hippocampal formation (PHF-E) according to Goedert et al. [8] and dissolved according to the amount of tissue used for isolation PHF-tau prep
Phospho-tau (181P) (pM)
Total tau (pM)
PHF-A PHF-B PHF-C PHF-D PHF-E
101 108 61 137 219
36 25 18 35 87
52
E. Vanmechelen et al. / Neuroscience Letters 285 (2000) 49±52
higher than the level of total tau. Using this factor as a correction, the degree of phosphorylation of CSF-tau in controls can be estimated: 15.9/(8.3 £ 4.3) 45%. This is in close agreement with the phosphorylation status of threonine 181 in human or rat brain-derived tau under normal conditions (20±30% [13,17]). Levels of CSF-phospho-tau are signi®cantly increased in AD, while decreased in FTD patients. Since CSF-phosphotau is closely correlated to total CSF-tau levels, it suggests that the degree of phosphorylation is differentially affected in several clinical conditions, as suggested [11]. A standardized assay for phospho-tau, as presented here, is essential for further validation of phospho-tau as a potential biomarker for tau pathology in different tauopathies, such as AD, FTD, corticobasal degeneration, progressive supranuclear palsy and others. Since Lewy body dementia requires expert treatment management [15] and is a special form of AD without tau pathology, phospho-tau as a biomarker might have therapeutic implications. This work was supported by a grant from the `Vlaams Instituut voor de bevordering van het Wetenschappelijk Technologisch onderzoek in de industrie' (Grant 990220) and the Medical Research Council, Sweden, Grant No. 12103 and 11560. Fred Shapiro is thanked for critical review and editing of the manuscript and Bob DeLeys for the initial experiments on the epitope mapping. [1] Blennow, K., Wallin, A., Agren, H., Spenger, C., Siegfried, J. and Vanmechelen, E., Tau protein in cerebrospinal ¯uid: a biochemical marker for axonal degeneration in Alzheimer disease? Mol. Chem. Neuropathol., 26 (1995) 231±245. [2] Braak, E., Braak, H. and Mandelkow, E.M., A sequence of cytoskeleton changes related to the formation of neuro®brillary tangles and neuropil threads, Acta Neuropathol. (Berlin), 87 (1994) 554±567. [3] Carmel, G., Mager, E.M., Binder, L.I. and Kuret, J., The structural basis of monoclonal antibody Alz50's selectivity for Alzheimer's disease pathology, J. Biol. Chem., 271 (1996) 32789±32795. [4] Delacourte, A., Sergeant, N., Wattez, A., Gauvreau, D. and Robitaille, Y., Vulnerable neuronal subsets in Alzheimer's and Pick's disease are distinguished by their tau isoform distribution and phosphorylation, Ann. Neurol., 43 (1998) 193±204. [5] DeLeys, R., Hendrickx, G., Dekeyser, F., Demol, H., Raymackers, J., Brasseur, R., Borremans, F., Vanmechelen, E. and Van de Voorde, A., Mapping and sequence requirements of the phosphorylation-sensitive epitopes recognized by the monoclonal antibodies Tau1, BT2, and AT8, In C.H. Schneider (Ed.), Peptides in Immunology, Wiley, New York, 1996, pp. 239±244.
[6] Goedert, M., Jakes, R., Crowther, R.A., Cohen, P., Vanmechelen, E., Vandermeeren, M. and Cras, P., Epitope mapping of monoclonal antibodies to the paired helical ®laments of Alzheimer's disease: identi®cation of phosphorylation sites in tau protein, Biochem. J., 301 (1994) 871±877. [7] Goedert, M., Jakes, R., Spillantini, M.G., Hasegawa, M., Smith, M.J. and Crowther, R.A., Assembly of microtubuleassociated protein tau into Alzheimer-like ®laments induced by sulphated glycosaminoglycans, Nature, 383 (1996) 550±553. [8] Goedert, M., Spillantini, M.G., Cairns, N.J. and Crowther, R.A., Tau proteins of Alzheimer paired helical ®laments: abnormal phosphorylation of all six brain isoforms, Neuron, 8 (1992) 159±168. [9] Goedert, M., Spillantini, M.G., Jakes, R., Rutherford, D. and Crowther, R.A., Multiple isoforms of human microtubuleassociated protein tau: sequences and localization in neuro®brillary tangles of Alzheimer's disease, Neuron, 3 (1989) 519±526. [10] GoÈtz, J., Probst, A., Spillantini, M.G., SchaÈfer, T., Jakes, R., Byne, W. and Goedert, M., Somatodendritic localization and hyperphosphorylation of tau protein in transgenic mice expressing the longest human brain tau isoform, EMBO J., 14 (1995) 1304±1313. [11] Ishiguro, K., Ohno, H., Arai, H., Yamaguchi, H., Urakami, K., Park, J.M., Sato, K., Kohno, H. and Imahori, K., Phosphorylated tau in human cerebrospinal ¯uid is a diagnostic marker for Alzheimer's disease, Neurosci. Lett., 270 (1999) 91±94. [12] Kosik, K.S., Orecchio, L.D., Binder, L., Trojanowski, J.Q., Lee, V.M. and Lee, G., Epitopes that span the tau molecule are shared with paired helical ®laments, Neuron, 1 (1988) 817±825. [13] Matsuo, E.S., Shin, R.W., Billingsley, M.L., Van deVoorde, A., O'Connor, M., Trojanowski, J.Q. and Lee, V.M., Biopsyderived adult human brain tau is phosphorylated at many of the same sites as Alzheimer's disease paired helical ®lament tau, Neuron, 13 (1994) 989±1002. [14] Mercken, M., Vandermeeren, M., Lubke, U., Six, J., Boons, J., Vanmechelen, E., Van de Voorde, A. and Gheuens, J., Af®nity puri®cation of human tau proteins and the construction of a sensitive sandwich enzyme-linked immunosorbent assay for human tau detection, J. Neurochem., 58 (1992) 548±553. [15] Rojas-Fernandez, C.H. and MacKnight, C., Dementia with Lewy bodies: review and pharmacotherapeutic implications, Pharmacotherapy, 19 (1999) 795±803. [16] Vandermeeren, M., Mercken, M., Vanmechelen, E., Six, J., Van de Voorde, A., Martin, J.J. and Cras, P., Detection of tau proteins in normal and Alzheimer's disease cerebrospinal ¯uid with a sensitive sandwich enzyme-linked immunosorbent assay, J. Neurochem., 61 (1993) 1828±1834. [17] Watanabe, A., Hasegawa, M., Suzuki, M., Takio, K., Morishima-Kawashima, M., Titani, K., Arai, T., Kosik, K.S. and Ihara, Y., In vivo phosphorylation sites in fetal and adult rat tau, J. Biol. Chem., 268 (1993) 25712±25717.