Amyloid- oligomers are inefficiently measured by enzyme-linked immunosorbent assay.код для вставкиСкачать
Amyloid-␤ Oligomers Are Inefficiently Measured by Enzyme-Linked Immunosorbent Assay Charlotte Stenh, MSc,1 Hillevi Englund, MSc,1 Anna Lord, MSc,1 Ann-Sofi Johansson, MSc,1 Claudia G. Almeida, MSc,2 Pär Gellerfors, PhD,1 Paul Greengard, PhD,3 Gunnar K. Gouras, MD, PhD,2 Lars Lannfelt, MD, PhD,1 and Lars N. G. Nilsson, PhD1 Amyloid-␤ (A␤) peptide levels are widely measured by enzyme-linked immunosorbent assay (ELISA) in Alzheimer’s disease research. Here, we show that oligomerization of A␤ results in underestimated A␤ ELISA levels. The implications are that comprehensive analysis of soluble A␤ requires either sample pretreatment at denaturing conditions or novel conformation-dependent immunoassays. Our findings might be of relevance for many neurodegenerative disorders in which soluble protein aggregates are the main neurotoxic species. Ann Neurol 2005;58:147–150 Enzyme-linked immunosorbent assay (ELISA) is widely used for measurements of amyloid-␤ (A␤) peptides, the main constituents of senile plaques in Alzheimer’s disease (AD) brain. The A␤ ELISAs have been very valuable in AD research and are today routinely used in preclinical research and anti-A␤ drug evaluation. Levels of A␤ are also measured with ELISA as a potential biomarker of AD, in which low levels of A␤42 is generally demonstrated in cerebrospinal fluid (CSF) among AD patients.1,2 In plasma, elevated levels of A␤42 are observed among familial AD cases.3 However, it has not been clarified whether A␤ levels in plasma are changed in sporadic AD. A␤, in particular, A␤42, has an inherent ability to aggregate into different soluble A␤ oligomers, for example, A␤-derived diffusible ligands (ADDLs)4 and From the 1Department of Public Health/Geriatrics, Rudbeck Laboratory, Uppsala University, Uppsala, Sweden, 2Department of Neurology and Neuroscience, Weill Medical College of Cornell University; and 3Laboratory of Molecular and Cellular Neuroscience, Rockefeller University, New York, NY. Received Feb 7, 2005, and in revised form Apr 18. Accepted for publication Apr 18, 2005. Published online Jun 27, 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.20524 Address correspondence to Dr Lannfelt, Department of Public Health/Geriatrics, Rudbeck Laboratory, Uppsala University, Dag Hammarskjölds väg 20, SE-751 85 Uppsala, Sweden. E-mail: email@example.com protofibrils,5,6 and eventually into insoluble fibrils. There is increasing evidence that soluble A␤ oligomers are present in human brain,7 are highly neurotoxic,8 –10 and possibly are the main pathogenic A␤ species in AD. Soluble A␤ oligomers might differ in their immunological properties as compared with A␤ monomers, because the C-terminal part is thought to be hidden inside the hydrophobic core of the aggregate.11 It therefore is unclear to what extent A␤ ELISAs can detect soluble A␤ oligomers, because detection is highly dependent on the conformation specificity of the antibodies used in the ELISA. A previous study drew attention to this problem by showing increased plasma A␤ signals if samples were denatured before the immunoassay.12 Previous studies of the Arctic familial Alzheimer mutation (APP E693G)13 have indeed suggested that A␤ ELISAs may underestimate A␤ levels in the presence of A␤ oligomers. The Arctic mutation leads to accelerated oligomerization of monomeric A␤ into protofibrils13 and altered amyloid precursor protein (APP) processing by favoring ␤-secretase cleavage.14,15 The latter implies that the Arctic mutation would cause increased A␤ levels. However, contrary to other pathogenic AD mutations, ELISA measurements have shown decreased levels of A␤40 and A␤42 in plasma from carriers of the Arctic mutation,13 as well as in cell culture media from transfected cells.13,14 Taken together, these findings suggest that the presence of A␤ oligomers may lead to underestimations of A␤ levels measured by ELISA. Here, we investigated if the decreased A␤ ELISA levels seen with the Arctic mutation were caused by reduced ability to measure oligomeric A␤, and if this hypothesis would be more generally applicable by extending also to wild-type A␤ oligomers. Materials and Methods Transfections HEK293 (human embryonic kidney) cells were cultured in Dulbecco’s minimum essential medium (Invitrogen, La Jolla, CA) supplemented with 10% fetal bovine serum. Cells were transfected with plasmid vector pcDNA3 (Invitrogen) containing either the APP gene with both the Arctic and Swedish mutations (APPArc-Swe) or with the Swedish mutation alone (APPSwe), using FuGENE 6 transfection reagent (Roche, Basel, Switzerland). Transfected HEK293 cells were analyzed 48 hours after transfection. Animals and Tissue Preparation cDNA clones coding for human APP695 carrying both the Swedish and the Arctic mutations (APPArc-Swe) or the Swedish mutation alone (APPSwe) were microinjected (2g/ml) into pronuclear oocytes from the hybrid mouse line C57BL/ 6-CBA-F1 and implanted into pseudopregnant foster mothers at the two-cell stage. Founder mice were identified by genotyping tail DNA with primer pairs 2577–2596 in M12379.1 and 541–522 in Y00264 as well as 1858 –1877 © 2005 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services 147 Size Exclusion Chromatography and 2299 –2281 in Y00264. Importantly, the APPArc-Swe and APPSwe founder lines used for this study both had a threefold human APP overexpression in the brain, as analyzed with Western blot. The mice used in the biochemical analyses were 2-3 months old on a C57BL/6-CBA background. The animals were anesthetized with 0.4ml Avertin (SigmaAldrich, St. Louis, MO) (50mg/ml) and perfused intracardially with 0.9% saline solution. Brains were prepared and immediately frozen in liquid nitrogen and then stored at ⫺80°C. One brain hemisphere was extracted in 1% sodium dodecyl sulfate (SDS)/0.1% Tween 20 in 1 ⫻ phosphatebuffered saline (PBS) containing Protease Inhibitor Cocktail Tablet (Roche), with an extraction ratio of 1 to 10 (wt/vol). The tissue was homogenized, sonicated for 30 seconds at a defined setting, centrifuged at 100,000g for 1 hour at ⫹4°C and the supernatant was saved. Before each SEC analysis, 1.8% Tween 20 was added to the aliquots to a final concentration of 0.6%, resulting in a final peptide concentration of 35M. The samples were centrifuged for 5 minutes, 17,900g at 16°C before injection. Samples (10l) were analyzed on a Merck (Westchester, PA) Hitachi D-7000 HPLC LaChrom system with a diod array detector. A Superdex 75 PC3.2/30 column (Amersham Biosciences, Arlington Heights, IL) was used for SEC. The samples were eluted with PBS–Tween 20 (50mM sodium phosphate, 0.15M NaCl, pH 7.4, 0.6% Tween 20) at a flow rate of 0.08ml/minute at ambient temperature. All injected samples were subjected to a wavelength scan between 200 and 400nm, and data were collected from 214nm. Peak areas were integrated using Merck Hitachi model D-7000 Chromatography Data station software. Western Blot Statistical Analysis Transfected HEK293 cells were lysed in NP lysis buffer (50mM Tris-HCl, pH 7.5, 150mM NaCl, 1% NP-40, 1 Protease Inhibitor Cocktail Tablet/25ml; Roche). Cell media and lysates from transfected HEK293 cells, SDS-soluble brain tissue extracts, and synthetic A␤1-42 wild-type aliquots were run on either 10 to 20% Tris-tricine or 4 to 20% Trisglycine SDS gels (Invitrogen) followed by electrophoretic transfer to nitrocellulose membranes (Schleicher and Schuell, Keene, NH, or Bio-Rad, Richmond, CA). Membranes were boiled in 1 ⫻ PBS for 5 minutes, blocked in 5% wt/vol nonfat dry milk, and incubated with 6E10 (Signet Laboratories, Dedham, MA), anti–␤-actin (Sigma-Aldrich), or an A␤40-specific C-terminal antibody (generously provided by Jan Näslund) for up to 72 hours at ⫹4°C. Bands were visualized with horseradish peroxidase–conjugated anti–mouse antibody (Pierce, Rockford, IL) and enzyme chemiluminescence detection. Levels of ␤-actin and A␤ were analyzed and normalized for APP levels. A␤ Enzyme-Linked Immunosorbent Assay Conditioned cell media, SDS-soluble brain tissue extracts, and synthetic A␤1-42 wild-type aliquots were analyzed for A␤1-40 and A␤1-42 levels with ELISA using Amyloid Beta 1-40 and 1-42 ELISA kits (Signet Laboratories), according to manufacturer’s instructions. To ensure equal epitope recognition between Arctic and wild-type A␤ by the antibodies used in the ELISA, we analyzed dilution series of synthetic A␤1-40 Arctic and A␤1-40 wild-type in their monomeric form with the Amyloid Beta 1-40 ELISA kit. Preparation of Synthetic Peptide Synthetic A␤1-42 wild-type (PolyPeptide Laboratories, Wolfenbüttel, Germany) was dissolved in 10mM NaOH to a final concentration of 100M, followed by 2 minutes of vortexing. This stock was diluted in 2 ⫻ PBS to a final concentration of 50M and vortexed for 1 minute. The 50M peptide solution was divided into aliquots and incubated at 37°C. Aliquots were taken for analyses with size exclusion chromatography (SEC), ELISA, or Western blot after 0, 70, 210, 280, or 350 minutes or 23 hours of incubation. 148 Annals of Neurology Vol 58 No 1 July 2005 Data were statistically analyzed with unpaired Student’s t test with or without Welch correction depending on population distribution (GraphPad InStat 3.0, GraphPad Software, San Diego, CA) and presented as mean ⫾ standard error of the mean. Results A combination of the Swedish16 and the Arctic13 APP mutation (APPArc-Swe) was expressed in cells and transgenic mice to generate high levels of oligomerized A␤ peptides. As a control, the Swedish APP mutation (APPSwe), which produces wild-type A␤ peptides, was expressed alone. A␤ levels in media from HEK293 cells transfected with APPArc-Swe or APPSwe were analyzed with both ELISA and Western blot. The ELISA demonstrated a 30% decrease in A␤1-40 levels and a 70% decrease in A␤1-42 levels in media from APPArc-Swe cells as compared with APPSwe (Fig 1A, B), similar to earlier results.14 Remarkably, Western blot (antibody 6E10) of the same samples showed a 40% increase in total A␤ levels in APPArc-Swe cells as compared with APPSwe (see Fig 1C). Furthermore, brain tissue of young APPArc-Swe and APPSwe transgenic mice was analyzed with both methods. When measured by ELISA, a 50% decrease of both A␤1-40 and A␤1-42 levels was found in APPArc-Swe transgenic mice as compared with APPSwe (see Fig 1D, E). In contrast, A␤ levels were increased by 400% when the same samples were analyzed by Western blot (see Fig 1F). There was no significant difference in APP protein levels (normalized to ␤-actin) between APPArc-Swe and APPSwe cells or between APP Arc-Swe and APPSwe transgenic mice, (see Fig 1G, H). Moreover, the A␤1-40 ELISA detected monomeric Arctic and wildtype A␤40 peptides in a comparable way (data not shown), demonstrating that the Arctic amino acid substitution does not interfere with the antibody recognition in the ELISA. Furthermore, Western blot analysis using an A␤40-specific C-terminal antibody demonstrated results similar to when using 6E10 (data not shown), suggesting that the observed differences between Western blot and ELISA is not caused by interference of the Arctic mutation with the C-terminal A␤40-specific antibody used in the ELISA. We wanted to determine whether oligomerization of wild-type A␤ also leads to underestimated A␤ levels by ELISA. Therefore, the oligomerization of wild- Fig 1. Amyloid-␤ (A␤) levels in cell culture media (A–C) and in 2- to 3-month-old transgenic mice brain (D–F). Enzymelinked immunosorbent assay (ELISA) measurements of A␤ levels in cell media showed (A) a 30% reduction of A␤1-40 and (B) a 70% reduction of A␤1-42 in APPArc-Swe cells in comparison with APPSwe. (C) In contrast, Western blot of the same samples demonstrated a 40% increased A␤ level in media from APPArc-Swe cells in comparison with APPSwe. (D) Levels of A␤1-40 and (E) A␤1-42 were decreased by 50% in APPArc-Swe mice compared with APPSwe when measured by ELISA. (F) Western blot showed a 400% increase in A␤ levels in APPArc-Swe mice as compared with APPSwe. (G, H) Representative Western blot gels showing APP and ␤-actin in APPSwe and APPArc-Swe cells (G) and transgenic mice (H). There was no significant difference in APP/␤-actin ratio between APPArc-Swe and APPSwe cells and transgenic mice indicating equal APP expression. All data were statistically analyzed with unpaired Student’s t test with or without Welch correction and presented as means ⫾ SEM. *p ⬍ 0.05, **p ⬍ 0.01 and ***p ⬍.001. n ⫽ 10 –11 (cell culture media, A–C), n ⫽ 6 –9 (transgenic mice, D–F). type A␤1-42 was analyzed in parallel with SEC, ELISA, and Western blot in a kinetic experiment. Wild-type A␤1-42 was incubated at 37°C for up to 23 hours to generate soluble A␤ protofibrils, here defined as high-molecular A␤ species that remains in Fig 2. Kinetic experiments of A␤1-42 wild-type oligomerization at 37°C. (A) Size exclusion chromatography (SEC) enables visualization of amyloid-␤ (A␤) oligomerization and fibrilization.13 At time zero, approximately 50% of the A␤1-42 wild-type peptide has oligomerized into protofibrils. Over time, the level of monomers declines as more protofibrils are formed. Moreover, the level of protofibrils starts to decline after 70 minutes as fibrillization begins. (B) At time zero, ELISA A␤1-42 wild-type levels are approximately 50% of the theoretical level, and this level continues to decline in association with protofibril formation. This demonstrates the loss of epitopes for the antibodies in the ELISA. (C) Western blot demonstrates unchanged A␤1-42 wild-type levels over time, despite the presence of oligomers. Centrifugation of the 23hour A␤1-42 wild-type sample before gel loading results in a markedly weaker band, as compared with the sample that was not centrifuged. Stenh, et al: A␤ Oligomers and ELISA 149 the supernatant after centrifugation for 5 minutes at 17,900g and elutes in the void volume of a Superdex 75 PC3.2/30 column. The SEC analysis showed a rapid aggregation of A␤ monomers into protofibrils, leading to an almost 100% protofibril content in the sample after 70 minutes of incubation. After 23 hours of incubation, nearly all A␤ peptides partitioned into the pellet upon centrifugation (Fig 2A, C). The same samples analyzed by ELISA showed a decline in A␤1-42 levels with time (see Fig 2B) but remained essentially unchanged over time when assayed by Western blot (see Fig 2C). This demonstrates that samples containing soluble A␤ oligomers need to be assayed in a denaturing environment to quantify total A␤ content. Discussion In conclusion, our findings demonstrate that oligomerization of A␤ peptides complicates interpretations of ELISA measurements, because A␤ aggregates are inefficiently detected by ELISA leading to an underestimation of A␤ levels. We believe that the decreased A␤ levels in plasma from patients with the Arctic mutation13 could be explained by this effect. Interestingly, the decreased plasma A␤ levels among Arctic mutation carriers suggests that oligomeric forms of A␤, that is, protofibrils, might exist in plasma. Recent transgenic studies with the Arctic mutation reinforces the perception that A␤ oligomerization and fibrilization are central in AD pathogenesis.15,17 Our findings highlight the great need of methods suitable to discriminate and quantify various soluble A␤ species to gain a better understanding of A␤ oligomers in AD. Protein misfolding and aggregation is a common theme among many neurodegenerative disorders, for example, Parkinson’s disease, amyotrophic lateral sclerosis, and prion disorders. Thus, novel ELISAs with conformation-specific antibodies might be useful for the analysis of neurotoxic oligomeric species in various pathogenic conditions. This work was supported by the Alzheimer Foundation (C.S., L.L.), Swedish Society for Medical Research, Hjärnfonden (L.L.), Bertil Hållstens Forskningsstiftelse (L.L.), The Swedish Research Council (2003-5546, L.L., 2004-6203, L.N.), DIADEM (QLK3-CT-200102362, L.L.), APOPIS (contract no. LSHM-CT-2003-503330, L.L.), and the NIH (National Institute of Aging, AG09464, P.G., G.K.G.). We thank H. Basun and M. Ingelsson for valuable discussions, H. van der Putten for supplying the 323-vector, and J. Näslund for supplying the A␤40-specific antibody. References 1. Jensen M, Schroder J, Blomberg M, et al. Cerebrospinal fluid A beta42 is increased early in sporadic Alzheimer’s disease and declines with disease progression. Ann Neurol 1999;45: 504 –511. 150 Annals of Neurology Vol 58 No 1 July 2005 2. Motter R, Vigo-Pelfrey C, Kholodenko D, et al. Reduction of beta-amyloid peptide42 in the cerebrospinal fluid of patients with Alzheimer’s disease. Ann Neurol 1995;38:643– 648. 3. Scheuner D, Eckman C, Jensen M, et al. Secreted amyloid beta-protein similar to that in the senile plaques of Alzheimer’s disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked to familial Alzheimer’s disease. Nat Med 1996;2:864 – 870. 4. Lambert MP, Barlow AK, Chromy BA, et al. Diffusible, nonfibrillar ligands derived from Abeta1– 42 are potent central nervous system neurotoxins. Proc Natl Acad Sci USA 1998;95: 6448 – 6453. 5. Harper JD, Wong SS, Lieber CM, Lansbury PT. Observation of metastable Abeta amyloid protofibrils by atomic force microscopy. Chem Biol 1997;4:119 –125. 6. Walsh DM, Lomakin A, Benedek GB, et al. Amyloid betaprotein fibrillogenesis. Detection of a protofibrillar intermediate. J Biol Chem 1997;272:22364 –22372. 7. Gong Y, Chang L, Viola KL, et al. Alzheimer’s disease-affected brain: presence of oligomeric A beta ligands (ADDLs) suggests a molecular basis for reversible memory loss. Proc Natl Acad Sci USA 2003;100:10417–10422. 8. Hoshi M, Sato M, Matsumoto S, et al. Spherical aggregates of beta-amyloid (amylospheroid) show high neurotoxicity and activate tau protein kinase I/glycogen synthase kinase-3beta. Proc Natl Acad Sci USA 2003;100:6370 – 6375. 9. Walsh DM, Hartley DM, Kusumoto Y, et al. Amyloid betaprotein fibrillogenesis. Structure and biological activity of protofibrillar intermediates. J Biol Chem 1999;274:25945– 25952. 10. Kayed R, Head E, Thompson JL, et al. Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis. Science 2003;300:486 – 489. 11. Roher AE, Baudry J, Chaney MO, et al. Oligomerizaiton and fibril asssembly of the amyloid-beta protein. Biochim Biophys Acta 2000;1502:31– 43. 12. Kuo YM, Emmerling MR, Lampert HC, et al. High levels of circulating Abeta42 are sequestered by plasma proteins in Alzheimer’s disease. Biochem Biophys Res Commun 1999;257: 787–791. 13. Nilsberth C, Westlind-Danielsson A, Eckman CB, et al. The ‘Arctic’ APP mutation (E693G) causes Alzheimer’s disease by enhanced Abeta protofibril formation. Nat Neurosci 2001;4: 887– 893. 14. Stenh C, Nilsberth C, Hammarbäck J, et al. The Arctic mutation interferes with processing of the amyloid precursor protein. Neuroreport 2002;13:1857–1860. 15. Cheng IH, Palop JJ, Esposito LA, et al. Aggressive amyloidosis in mice expressing human amyloid peptides with the Arctic mutation. Nat Med 2004;10:1190-1192. 16. Mullan M, Crawford F, Axelman K, et al. A pathogenic mutation for probable Alzheimer’s disease in the APP gene at the N-terminus of beta-amyloid. Nat Genet 1992;1:345–347. 17. Lord A, Kalimo H, Eckman C, et al. The Arctic Alzheimer mutation facilitates early intraneuronal A␤ aggregation and senile plaque formation in transgenic mice. Neurobiol Aging (published online Feb 17, 2005).