DC11: a novel monoclonal antibody revealing Alzheimer??s disease-specific tau epitope

NEUROREPORT

GENETICS OF NERVOUS SYSTEM DISEASES

DC11: a novel monoclonal antibody revealing Alzheimer’s disease-speci¢c tau epitope Lubica Vechterova,1,2 Eva Kontsekova,1 Norbert Zilka,1 Miroslav Ferencik,2 Rivka Ravid3 and Michal Novak1,2,CA 1

Axon Neuroscience, Renweg 95b, A-1030 Vienna, Austria; 2Institute of Neuroimmunology, Slovak Academy of Sciences, 842 45 Bratislava, Slovak Republic; 3The Netherlands Brain Bank, Meibergdreef 33,1105 Amsterdam, The Netherlands CA,2

Corresponding Author and Address

Received 20 October 2002; accepted18 November 2002 DOI: 10.1097/01.wnr.0000053064.88427.50

Using tau protein extracts from Alzheimer’s disease (AD) brain tissue, we generated a monoclonal antibody (mAb DC11) which decorated neuro¢brillary pathology in brain derived from AD patients on immunohistochemistry, and which lacked reactivity with healthy brain tissue. The same pattern of DC11 speci¢city was observed on Western blot. The main constituent of structures decorated by DC11 is microtubule-associated protein tau. In Western blot, DC11 recognized neither native healthy tau nor its full length

recombinant counterpart. However, the mAb showed strong immunoreactivity with truncated tau (residues t151^ 421), thus indicating the requirement for a conformational epitope. Importantly, the DC11 epitope was phosphorylation independent. The immunochemical parameters of mAb show that DC11 could represent a novel structural probe with the speci¢city for conformation of pathological tau present in AD brains. NeuroReport c 2003 Lippincott Williams & Wilkins. 14:87^91


Key words: Alzheimer’s disease; Conformation; Monoclonal antibody; Neuro¢brillary pathology; Tau protein

INTRODUCTION Alzheimer’s disease (AD) is the most frequent cause of premature irreversible cognitive decline in adult life. There is a correlation between the degree of dementia and the content of pathological lesions: neurofibrillary tangles, neuritic plaques and neuropil threads. The common term for these pathological structures is neurofibrillary pathology. Molecular studies of lesions have demonstrated that their constituents are paired helical filaments (PHF) [1,2]. The major protein subunit of PHF is the microtubuleassociated tau protein [3]. However, the conformation of tau involved in neurofibrillary pathology is distinct from that of normal healthy tau. Consequently, much effort was devoted to the generation of Alzheimer’s tau specific monoclonal antibodies (mAbs). Currently, two groups of anti-tau mAbs are broadly exploited for analysis of AD-tau: phosphorylation-dependent antibodies and phosphorylation-independent antibodies. The first group of antibodies (e.g. AT8, AT100, PHF1, AP422) recognizes phosphorylated epitopes (phosphoepitopes) on tau protein [4–7], while the second group (e.g. mAb 423, ALZ-50 family) is specific for unphosphorylated tau epitopes [8,9]. In spite of great effort, at present there are no mAbs available for immunohistochemical diagnosis which are able to decorate exclusively neurofibrillary pathology in AD brains. All commercial phosphorylation-dependent anti-tau antibodies react with AD-associated phosphoepitopes, which could be demon-

c Lippincott Williams & Wilkins 0959- 4965


strated not only on diseased tau in AD brains but also on normal tau in cell cultures [10]. Phosphorylation-independent mAb ALZ50 exhibits in Western blot cross-reactivity with healthy human tau [11], and affinity parameters of mAb 423 are not optimal for histochemical application. Therefore the aim of our study was to develop monoclonal antibody with genuine specificity for neurofibrillary pathology in AD brains, which could be effectively used in immunohistochemistry and Western blot. Here we describe the preparation and characterization of a novel anti-tau mAb designated DC11.

MATERIALS AND METHODS Immunogen: Preparation of immunogen from human AD brains was based on the method described by Ko¨pke et al. [12], with the following modifications. Brain tissues from hippocampus of three AD pure (non-mixed) dementia patients (stage VI according to Braak’s classification [13]) were homogenized (10% w/v homogenate) in 20 mM Tris pH 8, 0.32 mM sucrose, 10 mM b-mercaptoethanol, 5 mM EGTA, 10 mM EDTA, 5 mM MgSO4, 1 mM phenylmethylsulfonyl fluoride, 50 mM sodium fluoride, 5 mM benzamidine, 5 mg/ml leupeptin, 1.5 mg/ml pepstatin, 2 mg/ml aprotinin in a Heidolph DIAX 900 homogenizer for 10 min at 41C. The homogenates were pooled and centrifuged at 27 000  g for 30 min at 41C to remove cellular debris. The

Vol 14 No 1 2 0 January 2003

87

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

NEUROREPORT

L.VECHTEROVA ETAL.

pellet was resuspended in 100 mM Tris, pH 6.8, 20% glycerol, 4% sodium dodecylsulphate (SDS), 10 mM bmercaptoethanol, centrifuged at 27 000  g for 30 min at 41C and the supernatant (AD brain extract) was dialysed against phosphate-buffered saline (PBS) overnight at 41C. AD brain extracts were used for immunization of mice and further in Western blot analysis. The same method was used for preparation of brain extracts from control healthy agematched human brains. Immunization and fusion: Six-week-old female BALB/c mice (16 in total) were immunized with five i.p. injections of 50 ml AD brain extract emulsified in Freund’s complete adjuvant (1:1) in the first dose and Freund’s incomplete adjuvant (1:1) in the subsequent doses. Injections were done at 3–4 week intervals. The mice were boosted by an i.v. injection of the same amount of antigen without adjuvant. Harvested immune spleen cells were fused with mouse myeloma cell line NS0, following the fusion protocol described previously [14]. Growing hybridomas were screened by immunohistochemistry for production of monoclonal antibodies with specificity to decorate neurofibrillary pathology in AD brain tissue. Positive clones were recloned in soft agar. Hybridomas were grown in high glucose (4.5 g/l) Dulbecco’s modified minimal essential medium (Sigma) supplemented with 10% horse serum and their culture supernatants used for experiments. Isotypes of mAbs were determined by an enzyme-linked immunosorbent assay (ELISA) using mouse Ig-isotyping kit (ISO-2, Sigma). Recombinant human tau proteins: The cDNAs coding for six human tau isoforms were kindly provided by Dr M. Goedert (MRC-LMB Cambridge, UK). The cDNA for double truncated tau (residues t151–421) was derived from mRNA isolated from the hippocampus of an AD patient by PCR using specific primers. The numbering of amino acids is that of the longest human tau isoform, t40 [15]. All DNA constructs were cloned in pET17b vector (Novagen) through NdeI–EcoRI restriction sites. The integrity of each construct was verified by DNA sequence analysis (ABI Prism 377 DNA Sequencer, PerkinElmer). Tau cDNAs were expressed

Table 1.

in Escherichia coli and recombinant tau proteins purified from bacterial lysates according to methods described elsewhere [16,17]. Immunohistochemistry: Brain tissues from AD patients and age-matched healthy persons were obtained from Slovak Brain Bank, Bratislava and Netherlands Brain Bank, Amsterdam (Table 1). Tissues from entorhinal cortex and hippocampus were fixed in 4% paraformaldehyde for 2 days, placed in 20% sucrose until they settled and then quickly frozen. The 50 mm thin sections were cut from frozen brains on Leica cryostat. Non-specific binding was blocked by preincubation of sections in a solution containing 5% normal horse serum and 0.3% Triton in PBS. Sections were incubated with mAb DC11 or with anti-tau phosphorylation dependent antibody AT8 (Innogenetics; 1:100 dilution in 0.3% Triton, PBS, 5% horse serum) overnight at 41C. After washing, sections were incubated with horse anti-mouse Ig biotinylated antibody (Vector Labs) for 60 min, followed by avidin–biotin complex (Vector Elite, Burlingame, CA) and visualized with VIP solution (Vector VIP kit, Burlingame, CA). Sections were mounted on gelatin-coated slides, dried, treated with graded alcohols, xylene and coverslipped with Entellan (Merck, Darmstadt, Germany). For fluorescent labeling, sections were pretreated with 1% NaHBO4 for 30 min to reduce brain tissue autofluorescence. Then stained with the monoclonal antibody DC11 and incubated with a horse anti-mouse biotinylated antibody (1:1000, Vector Lab., CA), followed by streptavidin-Alexa 488 (1:1000, Molecular Probes). Fluorescence was observed using the Olympus BX 51 fluorescent microscope. Controls without primary antibodies were included for each experiment. Immunocytochemistry: For immunocytochemical analysis, cultures of hippocampal neurons were prepared from SHR rats (spontaneously hypertensive rat) at gestational day 18. Cells were plated on laminin-precoated glass slides (5  105 cells/slide) and cultivated for 5 days. Cultures were then fixed with 4% paraformaldehyde (Sigma) for 10 min at room temperature and cell membranes were permeabilized by incubation in 0.1% Triton X-100 in PBS for 4 min at room temperature. Cells were incubated with mAb DC11 or mAb

Summary of human material used for biochemical and histological analysis of mAb DC11.

Case number

PMD (h)

Gender

Age (years)

Duration (years)

Clinics

Cause of death

Brain weight (g)

CERAD

Braak staging

b4 amyloid

XV XVI XIII XVIII I II III IV V VII VIII

9 6 7 6 2.5 2.5 6 4 4 4 2.5

Male Male Male Male Female Female Male Female Female Female Female

53 55 86 57 80 87 79 78 75 81 76

^ ^ ^ ^ 5 5 6 8 3 4 5

Control Control Pre-AD Pre-AD AD AD AD AD AD AD AD

Infarct myocardis Tumor pulmonis Haemorhagic shock Oedema cerebri Cardio-pulmonary failure Pulmonary tromboembolus Cardio-pulmonary failure Pulmonary embolus Cardio-pulmonary failure Cardio-pulmonary failure Cardio-pulmonary failure

1260 1360 1230 1430 1200 1020 1200 1100 1150 1080 1000

NonAD NonAD PossAD ProbAD DefAD DefAD DefAD DefAD DefAD DefAD DefAD

^ ^ IV. IV. VI. IV. V. V. V. IV. VI.

  þ þ þ þ þ þ þ þ þ

The tissue samples were obtained from the Slovak Brain Bank and Netherlands Brain Bank. All cases were classi¢ed according to Braak’s staging procedure [13]. Brain tissues used in this study had post mortem delay of 2.5^9.0 h. PMD, post-mortem delay; AD, Alzheimer’s disease; pre-AD, preclinical AD without cognitive impairment; CERAD, Consortium to Establish a Registry for Alzheimer’s Disease.

88

Vol 14 No 1 2 0 January 2003

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

ALZHEIMER’S DISEASE-SPECIFIC TAU EPITOPE

NEUROREPORT

AT8 for 1 h at RT. After washing, cells were incubated with biotinylated secondary anti-mouse Ig serum (Vector Labs) and then with streptavidin-ALEXA 488 (Molecular Probes). Preparations were examined with a confocal microscope Olympus IX 70. The same procedures were used for immunocytochemical study with BDNF (brain derived neurotrophic factor) differentiated human neuronal cell line SH-SY5Y [18]. Western blot analysis: Extracts from human control brains and AD brains were diluted in SDS-sample buffer, boiled and loaded onto 5–20% SDS polyacrylamide gel and electrophoresed in Tris–glycine–SDS running buffer for B40 min at 25 mA. Transfer of proteins to a polyvinylidene fluoride (PVDF) membrane was performed for 1 h at 150 mA in 10 mM 3-(cyclohexylamino) propanesulfonic acid (CAPS), pH 12. After blocking with 3% non-fat dried milk in PBS for 1 h at room temperature, the membrane were incubated for 90 min with mAb DC11, Tau 1 (generous gift from Professor L. Binder) and pan-tau monoclonal antibody MN7.51 [19] (culture supernatants) respectively, followed by three washes with large volume of PBS for 5 min. Peroxidase conjugated goat anti-mouse Ig (DAKO) diluted 1:4000 with PBS were used as secondary antibody. Incubation (1 h at room temperature) was followed by washing (three times with 0.2% Igepal in PBS). Positive reaction was detected by enhanced chemiluminescence (ECL) method (Amersham).

Fig. 1. Monoclonal antibody DC11 is able to recognize neuro¢brillary tangles (arrow a), neuropil threads (arrow b) and neuritic plaques (arrow c) in AD brain tissues. Note the large number of neuro¢brillary tanglebearing pyramidal neurons stained in the hippocampal area CA 1. Bar ¼ 50 mm.

RESULTS AND DISCUSSION The aim of our work was to develop a novel monoclonal antibody with potential application for neurofibrillary pathology analysis in AD brain tissues. Specifically, we looked for an antibody recognizing the AD-specific and not only AD-associated epitope. A monoclonal antibody with such immunochemical parameters has not been prepared yet. For generation of the required antibody we immunized mice with hippocampal extracts from AD brains. Following the 16 fusions, 5467 growing hybridomas were screened by immunohistochemical assay for production of mAbs with capacity to stain neurofibrillary pathology in AD brain tissues. In the second round of screening, 83 positive antibodies were tested for potential cross-reactivity with normal brain tissues from healthy persons. From those only one hybridoma clone secreted monoclonal antibody with AD-specific reactivity thus qualifying for further analyzis in Western blot. The clone was designated DC11 and produced antibody was of IgG1-isotype, which gave specific results in both assays. Recloning twice in soft agar stabilized antibody secretion of this hybridoma. Culture supernatant containing monoclonal antibodies was used for further analysis. In immunochemistry, the antibody DC11 stained intensively and in highly specific manner neurofibrillary tangles, neuropil threads, ghost tangles and neuritic plaques which are present in AD brains (Fig. 1). No DC11 reactivity was observed with brain tissue form age-matched healthy persons (Fig. 2a,b). It is known that the major constituent of neurofibrillary pathology is tau protein [1], therefore we supposed that the DC11 epitope could result from ADspecific pathological conformation of tau. To prove this tempting possibility, we compared the immunoreactivity of mAb DC11 with the commercially available mAb AT8. The

Fig. 2. Monoclonal antibody DC11identi¢ed neuro¢brillary pathology in the enthorinal cortex in AD brain (a) but not in healthy age-matched control brain (b). Moreover there was a signi¢cant di¡erence in staining patterns between monoclonal antibody DC11 (c) and AT8 (d) on serial sections from AD entorhinal cortex. Normal tau in primary rat hippocampal cultures was not recognized with DC11 (e); however, AT8 immunoreactivity appeared mainly in axons and to the less extent in the perinuclear cytoplasm (f ). Bar ¼100 mm (a,b), 20 mm (c^ f).

Vol 14 No 1 2 0 January 2003

89

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

NEUROREPORT

L.VECHTEROVA ETAL.

monoclonal antibody AT8 is directed against the phosphorylation-dependent epitope (phosphoepitope) on tau protein and has also ability to decorate neurofibrillary pathology [5]. In our study, both antibodies recognized neurofibrillary lesions in AD brain tissues albeit with different staining pattern (Fig. 2c,d). It was suggestive that the antigenic structure against which DC11 was raised could be formed on diseased tau protein. Further detailed analysis showed that, there is a significant difference in the specificity between the novel mAb DC11 and mAb AT8. The latter antibody, as well as other phosphoepitope-specific mAbs (AT100, PHF1, AP422) routinely used for staining of AD neurofibrillary lesions, binds the AD-associated phosphoepitope formed on pathologically hyperphosphorylated tau, but the same epitopes could be demonstrated also on normal phosphorylated tau [7,10,20]. This difference is evident from comparison of DC11 immunorectivity with that of AT8 on rat primary cultures (Fig. 2e,f). Whereas AT8 strongly reacted with the normal phosphorylated tau, DC11 showed no cross-reactivity with rat neurons. This type of cross-reactivity with normal rat tau we observed also with mAbs AT100 and PHF1 (not shown). The lack of DC11 reactivity with rat tissue provided a direct proof that DC11epitope is distinct from epitopes of other anti-tau mAbs compared in this study. In order to exclude possible speciesspecificity of DC11 we have tested its reactivity with human neuronal cell line SH-SY5Y, with the same negative result. Based on our findings, we suggest that the DC11-epitope is Alzheimer’s tau specific and is not present on healthy tau protein. To confirm this suggestion, we have used for further DC11 epitope identification brain tau extracts from normal healthy age-matched controls; AD brain tissues and from six human recombinant tau isoforms. Indeed, Western blot analysis confirmed immunohistochemical findings and showed that mAb DC11 is able to discriminate between AD and healthy brain tissues (Fig. 3; Lane 1 and 2). The antibody showed high specificity for tau protein derived from AD brains and did not recognize the tau protein

62 47.5

DC11 1

2

MN 7.51

TAU 1 3

1

2

3

1

2

3

Fig. 3. In Western blot mAb DC11 recognized AD tau extracted from AD brains but did not react with tau extracted from healthy age-matched controls and with six human tau isoforms produced in bacteria. Two widely used anti-tau mAbs, Tau 1 and MN 7.51, stained healthy (lane1), diseased (lane 2) and recombinant (lane 3) tau proteins. Lanes: 1, tau extract from human healthy brain; 2, tau extract from AD brain; 3, six human recombinant tau isoforms.

90

175

83 62 47.5 32.5 25 16.5 6.5

1

2

3

1

2

3

Fig. 4. Western blot analysis of full length recombinant tau40 (residues t1^ 441) and tau fragment comprising residues t151^ 421 demonstrated that only tau truncation could produce a conformation required by DC11. A control blot reacted with pan-tau mAb 7.51. Lanes: 1, six human recombinant tau isoforms; 2, full length recombinant tau 40 (residues t1^ 441); 3, double truncated recombinant tau fragment composed of residues t151^ 421.

extracted from a normal human brain or recombinant human tau isoforms. Nevertheless, the unambiguous evidence that DC11 is able to react with tau protein, though structurally modified, was still missing. It has been suggested that to acquire AD-specific conformations, tau protein should undergo truncation [8]. Therefore we tested immunoreactivity of DC11 in Western blot with double truncated tau protein containing residues t151–421 (Fig. 4). Surprisingly, while lacking reactivity with full length tau, the antibody showed strong binding with the truncated tau. Further, this observation shows that the epitope recognized by DC11 is conformational and its formation requires truncation of tau. Importantly, the reactivity with bacterially produced recombinant tau protein demonstrated that the epitope for DC11 was phosphorylation independent. This characteristics clearly distinguished DC11 from a group of phosphoepitope-specific anti-tau antibodies. In addition, it is interesting to note, that two phosphorylation-independent mAbs recognizing AD-tau (mAb 423, ALZ-50), have both conformational epitopes, as well. However, in contrast to DC11, mAb 423 recognizes only tau C-terminally truncated at position Glu391 [8,19], whereas for binding of Alz50family the N-terminal residues 2–9 of normal tau are necessary [9]. Thus it is evident that the novel antigenic structure defined on AD-tau by DC11 is distinct also from those identified by other phosphorylation-independent antibodies.

CONCLUSIONS We have produced the first-generation mAb, DC11 that can specifically discriminate between AD specific neurofibril-

Vol 14 No 1 2 0 January 2003

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

ALZHEIMER’S DISEASE-SPECIFIC TAU EPITOPE

lary changes and normal healthy brain tissues. Furthermore, biochemical and immunological analysis of tau protein extracts from normal healthy brains, AD diseased brains and recombinant human tau isoforms proved that mAb DC11 is recognizing conformationally modified Alzheimer’s disease tau not present in normal healthy tau proteins. Identification of a novel antigenic structure on Alzheimer’s tau will open new possibilities for detailed analysis of tau modifications specific to Alzheimer’s disease. Furthemore, DC11 could be instrumental in dissecting the sequence of events leading to conformational transition of normal tau into AD-specific variant.

REFERENCES 1. Wischik CM, Novak M and Thogersen T. Proc Natl Acad Sci USA 85, 4506–4510 (1988). 2. Kondo J, Honda T, Mori H et al. Neuron 1, 827–834 (1988). 3. Kosik KS, Joachim CL and Selkoe D. Proc Natl Acad Sci USA 83, 4044–4048 (1986).

NEUROREPORT 4. Goedert M, Jakes R and Vanmechelen E. Neurosci Lett 189, 167–169 q(1995). 5. Zheng-Fischhofer Q, Biernat J, Mandelkow EM et al. Eur J Biochem 252, 542–552 (1998). 6. Otvos L, Feiner L, Lang E et al. J Neurosci Res 39, 669–673 (1994). 7. Hasegawa M, Jakes R, Crowther RA et al. FEBS Lett 384, 25–30 (1996). 8. Novak M, Kabat J and Wischik C. EMBO J 12, 365–370 (1993). 9. Jicha GA, Berenfeld B and Davies P. J Neurosci Res 55, 713–723 (1999). 10. Lesort M and Johnson GVW. Neuroscience 99, 305–316 (2000). 11. Greenberg SG and Davies P. Proc Natl Acad Sci USA 87, 5827–5831 (1990). 12. Ko¨pke E, Tung YC, Shaikh S et al. J Biol Chem 268, 24374–24384 (1993). 13. Braak E, Braak H and Mandelkow EM. Acta Neuropathol 87, 554–567 (1994). 14. Kontsekova´ E, Novak M, Kontsek P et al. Folia Biol 34, 18–22 (1988). 15. Goedert M, Spillantini MG, Jakes R et al. Neuron 3, 519–526 (1989). 16. Goedert M and Jakes R. EMBO J 9, 4225–4230 (1990). 17. Kontsekova E, Cattaneo A and Novak M. J Immunol Methods 185, 245–248 (1995). 18. Encinas M, Iglesias M, Liu Y et al. J Neurochem 75, 991–1003 (2000). 19. Novak M, Jakes R, Edwards PC et al. Proc Natl Acad Sci USA 88, 5837–5841 (1991). 20. Ga¨rtner U, Janke C, Holzer M et al. Neurobiol Aging 19, 535–543 (1998).

Acknowledgements:This work was supported by grant from Axon Neuroscience.Brain tissues were kindly supplied by Slovak Brain Bank, Bratislava.

Vol 14 No 1 2 0 January 2003

91

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.