|Home | About | Journals | Submit | Contact Us | Français|
Cerebral β-amyloidosis and associated pathologies can be exogenously induced by the intracerebral injection of small amounts of pathogenic Aβ-containing brain extract into young β-amyloid precursor protein (APP) transgenic mice. The probable β-amyloid-inducing factor in the brain extract has been identified as a species of aggregated Aβ that is generated in its most effective conformation or composition in vivo. Here we report that Aβ in the brain extract is more proteinase K- (PK) resistant than is synthetic fibrillar Aβ, and that this PK-resistant fraction of the brain extract retains the capacity to induce β-amyloid deposition upon intracerebral injection in young, pre-depositing APP23 transgenic mice. After ultra-centrifugation of the brain extract, less than 0.05% of the Aβ remained in the supernatant fraction, and these soluble Aβ species were largely PK-sensitive. However, upon intracerebral injection, this soluble fraction accounted for up to 30% of the β-amyloid induction observed with the un-fractionated extract. Fragmentation of the Aβ seeds by extended sonication increased the seeding capacity of the brain extract. In summary, these results suggest that multiple Aβ assemblies, with various PK sensitivities, are capable of inducing β-amyloid aggregation in vivo. The finding that small and soluble Aβ seeds are potent inducers of cerebral β-amyloidosis raises the possibility that such seeds may mediate the spread of β-amyloidosis in the brain. If they can be identified in vivo, soluble Aβ seeds in bodily fluids also could serve as early biomarkers for cerebral β-amyloidogenesis and eventually Alzheimer´s disease.
Alzheimer´s disease (AD) is an age-related neurodegenerative disorder and the most common form of senile dementia. Characteristic pathological hallmarks are proteinaceous deposits of amyloid-β (Aβ) (senile plaques) and hyperphosphorylated tau (neurofibrillary tangles), accompanied by an inflammatory response and damage to neurons and synapses (Duyckaerts et al., 2009; Holtzman et al., 2011). According to the amyloid cascade hypothesis, the misfolding and aggregation of Aβ is a fundamental event in the pathogenesis of AD (Hardy and Selkoe, 2002). While a wealth of studies have characterized Aβ aggregation in vitro (Harper and Lansbury, 1997; Roychaudhuri et al., 2009; Du et al., 2011), much less is known about the in vivo trigger for cerebral β-amyloidosis.
We and others have shown that intracerebral injections of minute amounts of Aβ-containing extract from AD brains or aged APP-transgenic (tg) mouse brains can induce β-amyloidosis and associated pathologies (activation of microglia and astrocytes; dystrophic neurites) in the brains of young APP-tg mice (Kane et al., 2000; Meyer-Luehmann et al., 2006; Eisele et al., 2009; Watts et al., 2011). β-amyloid deposition also can be induced in APP-tg mice by intraperitoneal infusion of Aβ-containing brain extracts (Eisele et al., 2010), indicating that the seeds can translocate from the periphery to the brain. The probable β-amyloid-inducing factor in the brain extract has been identified as a species of aggregated Aβ because (i) brain extracts from wild-type mice or pre-depositing APP-tg mice were ineffective in seeding β-amyloid; and (ii) Aβ-immunodepletion or formic acid-denaturation of the extract abolished the induction of β-amyloidosis (Meyer-Luehmann et al., 2006). The finding that synthetic Aβ preparations fail to show robust β-amyloid-inducing activity indicates that the misfolding of Aβ in vivo results in a different conformation and/or that brain-specific cofactors are needed for effective seeding (Meyer-Luehmann et al., 2006).
The exogenous induction of cerebral β-amyloidosis, either by intracerebral or intraperitoneal application of Aβ seeds, is reminiscent of the induction of prion disease in mice (Aguzzi and Rajendran, 2009; Colby and Prusiner, 2011). Prions consist of PrPSc, an abnormally folded isoform of the cellular prion protein (PrPc). Upon inoculation of mice, PrPSc can induce the templated misfolding of PrPc and thereby cause fatal neurodegeneration (Wadsworth et al., 2010). Infectious prions are multimers of PrPSc, and their size, solubility, resistance to proteinase K (PK) treatment, and ease of fragmentation govern their infectivity and neurotoxicity (Aguzzi et al., 2007; Collinge and Clarke, 2007).
The objective of the present study was to further characterize the β-amyloid-inducing factor in the Aβ-containing extract from aged APP-tg mouse brains. Guided by the prion findings, we have tested the PK resistance, solubility and fragmentation of the amyloidogenic seeds. Our results bolster the assumption that the β-amyloid-inducing factor is Aβ, and indicate that, rather than consisting of a single type of multimer, β-amyloid seeds comprise a range of Aβ aggregates of various sizes and PK-sensitivities.
For all experiments, 3–4 month-old male or female APP23 transgenic mice were used (Sturchler-Pierrat et al., 1997). The mice have been backcrossed with C57BL/6J mice for 20 generations (C57BL/6J-Tg(Thy1-APPK670N;M671L)23). APP23 mice currently bred at the Hertie Institute for Clinical Brain Research first develop individual β-amyloid plaques in the neocortex at 7 months of age. In the hippocampus, β-amyloid deposition starts at 8–10 months of age (females 8–9 months; males 9–10 months). All mice were kept under specific pathogen-free conditions. The experimental procedures were undertaken in accordance with the veterinary office regulations of Baden-Württemberg (Germany) and approved by the local Animal Care and Use Committees.
Brain extracts were derived from aged (22–28 month-old) β-amyloid-depositing male or female APP23 transgenic mice and from age-matched non-transgenic mice. After removal of the cerebellum and lower brainstem, the forebrain was immediately fresh-frozen on dry ice and stored at −80°C until use. Tissue was then homogenized (Ultra Turrax T8, IKA®-Werke Staufen, Germany) at 10% (w/v) in sterile, phosphate-buffered saline (PBS, Lonza, Switzerland), vortexed, sonicated three times for 5 seconds (LabSonic, B. Braun Biotech International GmbH, Melsungen, Germany; 0.5 mm diameter sonotrode, cycle 1, amplitude 80%) and centrifuged at 3,000 × g for 5 minutes as previously described (Meyer-Luehmann et al., 2006; Eisele et al., 2009). The supernatant was aliquoted and immediately frozen. For all experiments, the 10% (w/v) extract was used unless otherwise stated.
Aβ 1–40 and Aβ 1–42 (American Peptide, Sunnyvale, CA, USA) were dissolved at equimolar concentrations in PBS to reach a final concentration of 100 µM total Aβ. To generate fibrils, the peptides were incubated on a rotator for 5 days at 37°C as previously described (Meyer-Luehmann et al., 2006). Fibrils were stored at −80°C until use.
Brain extracts and synthetic Aβ preparations were thawed on ice and subsequently treated with 50µg/ml proteinase K (PK) (Roche Diagnostics GmbH, Mannheim, Germany) for up to 2 hours at 37°C. After 30 minutes, 1 hour or 2 hours, samples were boiled for 5 min at 95°C to heat-inactivate the PK.
Brain extracts were thawed on ice and centrifuged in a Beckman Centrifuge at 100,000 × g for 1 hour at 4°C in Protein LoBind Tubes® (Eppendorf AG, Hamburg, Germany). Supernatant was transferred to a new tube and the pellet was resuspended in sterile PBS to obtain the same volume as the supernatant. Samples were stored on ice until use.
Brain extracts (10% (w/v) in PBS; 3,000 × g supernatant) were thawed on ice and sonicated with a LabSonic (B. Braun Biotech International GmbH, Melsungen, Germany) three times for 20 seconds (0.5 mm diameter sonotrode, cycle 1, amplitude 80%).
Aβ levels (Aβx–38, Aβx–40 and Aβx–42) in extracts were determined with an electrochemiluminescence-linked immunoassay using the MSD® 96-well MULTI-SPOT® Human (6E10) Aβ Triplex Assay (Meso Scale Discovery, Gaithersburg, MD, USA). To this end, the extracts were first treated with formic acid (final concentration: 70%) (Sigma, St. Louis, MO, USA). Samples were sonicated for 30 seconds on ice and centrifuged at 25,000 × g for 1 h at 4°C. Supernatants were equilibrated in neutralization buffer (1M Tris base, 0.5M Na2HPO4, 0.05% NaN3). The 100,000 × g supernatant was measured directly. Aβ-detection was further conducted according to the manufacturer’s instructions. In brief, 96-well plates pre-spotted with capture antibodies against Aβx–38, Aβx–40, Aβx–42 and bovine serum albumin were blocked for 1 h with 1% Blocker A solution and then washed three times with 1× Tris buffer. In a second step, samples were diluted 1:100 in 1% Blocker A solution, except supernatant of ultracentrifugation, and co-incubated with the SULFO-TAG 6E10 detection antibody solution on the plate for 2 h. After washing, MSD Read Buffer T was added and the plate was read immediately on a Sector® Imager 6000. Data analysis used MSD® DISCOVERY WORKBENCH® software 2.0. For total Aβ, Aβx–38, Aβx–40 and Aβx–42 were combined.
Total protein of the brain extract was quantified with a microplate bicinchoninic acid (BCA) assay (Pierce Biotechnology, Rockford, IL, USA). Samples were measured on an ELISA plate reader (Mithras, Berthold Technologies, Bad Wildberg, Germany).
Brain extracts and synthetic Aβ preparations were analyzed on NuPage® Bis-Tris mini gels using NuPage® LDS sample buffer and MES running buffer (Invitrogen, Carlsbad, CA, USA). To stain for total protein, silver staining was performed according to a previously established protocol (Nesterenko et al., 1994) with the following modifications. Gels were fixed with a solution of 50% acetone, 1% trichloroacetic acid (TCA) and 0.015% formaldehyde (HCHO) for 5 minutes, washed three times with double-distilled water (ddH2O), incubated again with ddH2O for 5 minutes and washed again thrice. For pre-treatment, gels were incubated in pure acetone for 5 minutes and afterwards 1 minute in 1.4 mM dithionite solution. After washing with ddH2O, the staining solution (15.7mM silver nitrate in ddH2O with 0.18% HCOH) was added and incubated for 8 minutes. This step was followed by incubation in a developing solution (70mM soda-decahydrate, 0.015% HCHO and 0.004% sodium thiosulfate) and the staining was stopped afterwards with 3% glacial acetic acid. After washing with ddH2O, the gel was incubated in a gel-drying solution (5% glycerol and 30% methanol in ddH2O) for 30min. Staining of the gel was documented with a scanner, and afterwards the gel was dried with a Gel Dryer (BioRad, München, Germany).
Brain extracts and synthetic Aβ preparations were analyzed on NuPage® Bis-Tris mini gels using NuPage® LDS sample buffer and MES running buffer (Invitrogen, Carlsbad, CA, USA). For western blotting, samples were wet-blotted onto a nitrocellulose membrane, probed with 6E10 antibody (Covance Research Products, Dedham, MA, USA) and visualized with chemiluminescence using SuperSignal West Pico (Thermo Scientific, Rockford, IL, USA). Densitometric values of band intensities were analyzed using the public domain software ImageJ, version 1.34 (Rasband, 1997–2011).
APP23 host mice were anaesthetized with a mixture of ketamine (100mg/kg body weight) and xylazine (10mg/kg body weight) in saline. Bilateral stereotactic injections of 2.5µl brain extract were made with a Hamilton syringe into the hippocampus (AP −2.5mm, L +/− 2.0mm, DV −1.8mm). Injection speed was 1.0µl/minute and the needle was kept in place for an additional 2 minutes before it was slowly withdrawn. The surgical area was cleaned with sterile saline, the incision was sutured, and the mice were monitored until recovery from anaesthesia.
Brains were removed and immersion-fixed for 48h in 4% paraformaldehyde in PBS, then cryoprotected in 30% sucrose in PBS for an additional 2 days. After freezing, serial, 25µm-thick coronal sections were cut through the brains using a freezing-sliding microtome. The sections were collected in 0.1M Tris-buffered saline (pH 7.4) and stained immunohistochemically according to previously published protocols (Stalder et al., 2005). Polyclonal antibody CN3 (Eisele et al., 2010) was used for immunostaining of Aβ. Adjacent sections were stained with Congo red according to standard protocols and viewed under cross-polarized light. The following additional antibodies were used: rabbit polyclonal antibody against cow glial fibrillary acidic protein (GFAP) (Dako, Glostrup, Denmark); rabbit polyclonal antibody against ionized calcium binding adapter molecule 1 (Iba1) (Wako; Richmond,VA); and rabbit polyclonal antibody A4CT against APP (Calhoun et al., 1999) (gift of K. Beyreuther, Heidelberg, Germany).
Aβ load was quantified on an Aβ-immunostained set of every 12th systematically sampled, serial, coronal section throughout the entire hippocampus. Researchers who were blinded to the inoculation groups performed the analysis. Stereological analysis of CN3-positive staining was performed using a microscope equipped with a motorized x-y-z stage coupled to a video-microscopy system and the Stereo Investigator software (MicroBrightField, Inc., Williston, VT) as previously described (Bondolfi et al., 2002). The β-amyloid load (percentage) was determined by calculating the areal fraction occupied by CN3-positive immunostaining in two-dimensional sectors at a single focal plane (20×/0.45 objective). The percentage of Aβ deposits that were also Congo red-positive was determined manually on an adjacent set of systematically-sampled sections that were double-stained by the anti-Aβ-antibody and Congo red. Quantification was performed using a 20× objective and a Zeiss Axioskop 2 microscope (Zeiss, Oberkochen, Germany).
Brain extracts from aged APP23 transgenic mice (Tg extract) and preparations of synthetic Aβ fibrils in PBS or spiked into brain extract from aged non-transgenic (wildtype) mice (Wt extract) were incubated with 50 µg proteinase K (PK)/ml for up to 2 hours. Subsequent Aβ-immunoblotting revealed that Aβ in the Tg extract was to a high degree PK-resistant (80% remaining after 30min PK treatment; 70% remaining after 2h PK treatment), whereas synthetic Aβ fibrils were almost completely digested after 30 minutes (Fig. 1A–C). The spiking of the synthetic Aβ fibrils into Wt extract made the fibrils more PK-resistant (50% remaining after 30min PK treatment; 35% remaining after 2h PK treatment), although they were still significantly less resistant than were the Aβ aggregates in the Tg extract (Fig. 1A–C).
To test whether PK-treated Tg extracts harbor β-amyloid-inducing activity in vivo, Tg extracts with and without PK treatment were injected into the hippocampus of 4-month old pre-depositing APP23 mice and compared with the injection of Wt extracts (Fig. 1D–G). Inactivation of PK was achieved by boiling the samples (including the controls) prior to injection. Boiling has been found to reduce, but not abolish, the β-amyloid-inducing activity of the brain extracts (Meyer-Luehmann et al., 2006). Therefore, to optimize the seeding signal, an incubation time of 5 months was chosen. Immunohistochemical analysis revealed robust β-amyloid deposition in the hippocampus of mice receiving either the Tg extract or the PK-treated Tg extract, whereas β-amyloid was not induced by the Wt extracts (Fig. 1D–F), i.e., the effect of the Wt extract was not different from that of the PBS injection (Aβ load: 0.06 ± 0.03% vs. 0.03 ± 0.01%; n=5–6 per group, mean ± SEM, t = 1.893; p>0.05). β amyloid induced by the PK-treated Tg extract was both congophilic and diffuse (51.1 ± 8.1% of the Aβ deposits had a detectable Congo red-positive core). All congophilic β-amyloid plaques were surrounded by intensely stained hypertrophic GFAP-positive astrocytes, activated Iba1-positive microglia, as well as APP-positive dystrophic neurites, as previously described (Meyer-Luehmann et al., 2006). Quantification of the induced β-amyloid load indicated, however, that the PK-treated Tg extracts yielded only 55% of the β-amyloid-inducing activity of the Tg extract without PK treatment (Fig. 1G), suggesting a partial digestion of the β-amyloid-inducing activity by PK. Histological analysis of the brains one day after injection of the PK-treated Tg extract revealed no Aβ deposits, ruling out the possibility that the induced Aβ deposition represents the injected material itself (n=3; data not shown), consistent with previous studies (Meyer-Luehmann et al., 2006; Eisele et al., 2009).
Tg extract was ultracentrifuged (100,000 × g for 1 hour at 4°C) and the supernatant and pellet fractions were subsequently assessed for total protein content and Aβ levels. Measurements of total protein concentration (by BCA assay) revealed 3.6 ± 0.2 µg/µl and 2.8 ± 0.3 µg/µl protein in the supernatant and in the pellet fraction, corresponding to approximately 60% and 40% of total protein of the original extract, respectively. This is consistent with the silver staining of acrylamide gels (Fig. 2A). In contrast, Aβ-immunoblotting and Aβ-ELISA revealed that the great majority (>99.9%) of Aβ was in the pellet fraction (Fig. 2A; Table 1).
To investigate the β-amyloid-inducing activity of these fractions, 3–4-month-old, pre-depositing APP23 mice were given intracerebral injections of the material. Histological analysis was performed 4 months later (Fig. 2B–D). Both the supernatant and pellet fractions induced β-amyloid deposition. The β-amyloid induction by the pellet fraction achieved 95% of the activity observed with unfractionated, total Tg extract. Strikingly, the supernatant induced almost 30% of the β-amyloid load observed with the total Tg extract (Fig. 2E). Given that this fraction harbors less than 0.05% of the Aβ in the total Tg extract, these observations indicate a very high seeding activity of the 100,000 × g soluble Aβ assemblies in the brain extract (Table 1). The β-amyloid deposits induced by the pellet fraction were morphologically similar to those in the original Tg extract, while the β-amyloid induced by the supernatant was mostly diffuse and Congo red-negative (only 2.4 ± 1.0% of the compact Aβ deposits had a Congo red-positive core in comparison to 39.8 ± 6.5% with the unfractionated, total extract) after the 4-month incubation period (Fig. 2C). After a 6-month incubation period, however, the compact deposits induced by the supernatant fraction also were increasingly Congo red-positive (29.8 ± 1.2%) (Fig. 2F). Again, all of the Congo red-positive Aβ deposits were accompanied by cytopathologic changes, i.e. activation of microglia and astrocytes, as well as dystrophic neurites (Fig. 2G), all identical in appearance to those induced by the total extract and in normal aged APP23 mice (Meyer-Luehmann et al., 2006).
Next we assessed the PK-sensitivity of the soluble β-amyloid-inducing fraction (Fig. 3). Results demonstrated that less than 5% of the Aβ in the 100,000 × g supernatant fraction was resistant to 30 min PK treatment, while the original Tg extract again revealed 80% PK resistance (Fig. 3A, B). The pellet fraction resulting from ultracentrifugation showed PK resistance similar to that of the untreated Tg extract (data not shown). The in vivo β-amyloid-inducing activity of the PK-digested supernatant fraction was then evaluated by intracerebral injections into 3 month-old pre-depositing APP23 mice, and compared with the activity of the undigested supernatant fraction 6 months after injection. Inactivation of PK was again achieved by boiling the samples (including the controls) prior to injection. Results revealed the expected β-amyloid induction by the supernatant fraction, whereas PK treatment almost completely abolished the β-amyloid-inducing activity of this fraction (Fig. 3C–E), suggesting that the potent, β-amyloid-inducing soluble Aβ assemblies are PK-sensitive.
To determine whether fragmentation of aggregates can augment the seeding capacity of brain extracts, the impact of extended sonication on the β-amyloid-inducing activity of the Tg extracts was assessed. Immunoblot analysis of the original (briefly sonicated) Tg extract and Tg extract with extended sonication revealed no change in total Aβ concentration (Fig. 4A). However, ultracentrifugation of the extracts revealed a higher amount of soluble Aβ in the supernatant of the Tg extract that had undergone extended sonication, suggesting the breakage of insoluble Aβ assemblies into smaller, 100,000 × g soluble units (Fig. 4A). The additional Aβ released into the supernatant fraction after extended sonication showed PK sensitivity similar to that of the Aβ in the supernatant of the untreated Tg extract (data not shown).
To test the β-amyloid-inducing activity of the Tg extract with and without extended sonication, extracts were injected intracerebrally into 4-month-old APP23 mice, and the brains of the injected animals were analyzed 4 months later. Aβ-immunostaining revealed increased induction of β-amyloid deposits in the mice with the extra-sonicated brain extract compared to the mice injected with the original seeding extract (Fig. 4B–D). Morphologically, the original seeding extract induced a more filamentous and dense pattern of β-amyloid aggregation, whereas the deposits induced by the extra-sonicated seeding extract appeared smaller and more punctate. In both groups, a subset of the induced Aβ deposits were Congo red-positive (Fig. 4B, C).
Converging experimental data implicate corruptive protein templating in the induction and spread of proteinaceous lesions in a number of neurodegenerative diseases (Walker et al., 2006; Aguzzi and Rajendran, 2009; Colby and Prusiner, 2011). Pathologic similarities with prion disease suggest that these other disorders, though not known to be infectious, could involve the prion-like propagation of proteopathic seeds (Jucker and Walker, 2011). While it is now evident that the aggregation of Aβ can be exogenously induced in vivo by brain extracts containing aggregated Aβ (Kane et al., 2000; Meyer-Luehmann et al., 2006; Eisele et al., 2009; Eisele et al., 2010; Watts et al., 2011), the seeds themselves have not been rigorously characterized. The present findings, together with previous work (Meyer-Luehmann et al., 2006), indicate that the β-amyloid-inducing factor in brain extracts from aged, APP-tg mice consists of a range of aggregated Aβ species of different sizes and of various sensitivities to PK digestion.
We previously found that synthetic Aβ fibrils – in PBS or mixed with wildtype brain extract - do not effectively seed Aβ deposition in vivo (Meyer-Luehmann et al., 2006). Our present results reveal that aggregated Aβ from β-amyloid-laden brains is more PK-resistant than are synthetic Aβ fibrils. Inasmuch as the amino acid sequence is the same, this finding indicates that Aβ fibrils in vivo either assume a multi-dimensional conformation that confers PK resistance compared to that of synthetic Aβ fibrils, or that post-translational modifications and/or auxiliary factors in the brain inhibit the interaction of proteinase K with Aβ. Conformation has been shown to influence the resistance of Aβ to proteases, presumably by determining the accessibility of putative cleavage sites (Nordstedt et al., 1994), and differences in the structural properties of brain-derived Aβ fibrils and synthetic Aβ fibrils have been reported (Paravastu et al., 2009). In addition to these conformational differences, co-factors in the brain may protect Aβ fibrils from protease cleavage (Soderberg et al., 2005; Tennent et al., 1995), which may explain our finding of increased PK resistance of synthetic Aβ fibrils that are mixed with wildtype brain extract. Alternatively, the increased PK resistance could be an indirect effect of the increase in total protein when synthetic Aβ is mixed with wildtype brain extract.
Misfolded proteins with increased protease resistance may escape proteolytic degradation and become proteopathic seeds (Lansbury, 1997; Powers et al., 2009). However, PK resistant Aβ-seeds are not the sole inductive species in our assay. PK-treated extracts induced β-amyloid deposition at a magnitude of only ~55% of that of untreated extract (Fig. 1), arguing that PK-sensitive seeds must exist, and that they are highly efficient β-amyloid-inducing agents. This conclusion is supported by our ultracentrifugation experiments. Although less than 0.05% of the Aβ was in the 100,000 × g supernatant fraction, this soluble fraction was responsible for up to 30% of the seeding activity and was PK-sensitive (Fig. 2, ,3).3). Similarly, PK resistance was initially thought to be an important feature of infectious prions (McKinley et al., 1983), but more recent studies reveal that PK-sensitive PrP aggregates also are pathogenic, suggesting a range of infectious PrP species (Safar et al., 1998; Pastrana et al., 2006; Gambetti et al., 2008; Colby et al., 2010).
Small Aβ aggregates provide relatively more molecular interfaces for templated misfolding and efficient propagation compared to larger aggregates or fibrils. Thus, a higher fibrillogenic activity of small, soluble Aβ species compared to larger, insoluble assemblies would be predicted (Lee et al., 2007; Shankar and Walsh, 2009; Keshet et al., 2010). This view is supported by the present finding that extended sonication of Aβ-containing brain extract, i.e. fragmentation of larger β-amyloid assemblies into numerous smaller seeds, increases the seeding activity of the extract, in line with numerous in vitro studies of amyloidogenic proteins (Soto et al., 2002; Knowles et al., 2009; Xue et al., 2009). Similarly, for prions, non-fibrillar small particles were found to be the most infectious agent, and sonication has been reported to increase prion infectivity (Silveira et al., 2005).
In AD research, a variety of pathogenic Aβ species have been described, ranging from Aβ dimers to larger oligomers, protofibrils and fibrils (for review see (Haass and Selkoe, 2007; Glabe, 2008). In most studies of AD postmortem brain samples, the soluble pools consisting of small Aβ aggregates were found to correlate highly with AD severity and progression (Lue et al., 1999; McLean et al., 1999; Wang et al., 1999; Tomic et al., 2009; Mc Donald et al., 2010). In the present study, the morphological appearance of the induced β-amyloid deposits differed between the soluble and insoluble fractions. The β-amyloid deposits induced by the insoluble and PK-treated brain extracts were rather large and frequently congophilic aggregates that appeared to be nonuniformly distributed throughout the injected area. In contrast, the soluble Aβ species from brain extracts induced smaller, more often Congo red-negative patches of Aβ aggregates with a more homogeneous distribution throughout the injected region. Similarly, fragmentation of seeds by extended sonication induced smaller and more punctate β-amyloid deposits compared to the original, briefly sonicated extract. While these findings imply that the size of the seeds influences the nature of the induced deposits, the distributional and morphological differences of the induced lesions may also represent a difference in diffusion/transport of the larger, bulkier, insoluble Aβ seeds versus the smaller soluble seeds. In all cases, and independent of the morphological appearance, the congophilic aggregates were similarly surrounded by activated microglia, hypertrophic astrocytes, and dystrophic neurites, as previously described (Meyer-Luehmann et al., 2006). The relatively delayed emergence of Congo-red positive plaques in mice receiving the soluble material suggests either a different time-course of initiation following seeding, and/or that these lesions eventually develop from the diffuse plaques.
In summary, the β-amyloid-inducing activity of brain extracts containing aggregated Aβ can be attributed to multiple entities with varying degrees of PK resistance or sensitivity. While insoluble pools of Aβ clearly are capable of seeding β-amyloidosis in vivo, brain extracts also contain soluble Aβ assemblies that we found to be super-proportionally active in inducing β-amyloidosis. The existence of soluble, bioactive Aβ seeds raises the possibility that these agents may mediate the spread of β-amyloidosis within the brain, and that they could serve as diagnostic biomarkers in biological fluids such as CSF or plasma. There is increasing evidence that Aβ deposition in the brain precedes the clinical symptoms of AD by many years (Holtzman et al., 2011). Sensitive and reliable identification of the earliest forms of pathogenic Aβ assemblies has the potential to expand the window for preventive interventions, and to refine current therapeutic strategies specifically targeting the initial and most pathogenic Aβ-seeds.
We would like to thank Lisa Münter, Gerd Multhaup (Berlin), Heinke Schieb, Hans Klafki, Jens Wiltfang (Essen), Christoph Schall, Thilo Stehle (Tübingen), Ulrike Obermüller, Jörg Odenthal, Stephan Kaeser and all the other members of our department for experimental help and comments on this manuscript. This work was supported by grants from the Competence Network on Degenerative Dementias (BMBF-01GI0705), and the BMBF in the frame of ERA-Net NEURON (MIPROTRAN) and NIH RR-00165.