Requirement of aggregation propensity of Alzheimer amyloid peptides for neuronal cell surface binding

Abstract

Background: Aggregation of the amyloid peptides, Abeta40 and Abeta42, is known to be involved in the pathology of Alzheimer's disease (AD). Here we investigate the relationship between peptide aggregation and cell surface binding of three forms of Abeta (Abeta40, Abeta42, and an Abeta mutant).

Results: Using confocal microscopy and flow cytometry with fluorescently labelled Abeta, we demonstrate a correlation between the aggregation propensity of the Alzheimer amyloid peptides and their neuronal cell surface association. We find that the highly aggregation prone Abeta42 associates with the surface of neuronal cells within one hour, while the less aggregation prone Abeta40 associates over 24 hours. We show that a double mutation in Abeta42 that reduces its aggregation propensity also reduces its association with the cell surface. Furthermore, we find that a cell line that is resistant to Abeta cytotoxicity, the non-neuronal human lymphoma cell line U937, does not bind either Abeta40 or Abeta42.

Conclusion: Taken together, our findings reveal that amyloid peptide aggregation propensity is an essential determinant of neuronal cell surface association. We anticipate that our approach, involving Abeta imaging in live cells, will be highly useful for evaluating the efficacy of therapeutic drugs that prevent toxic Abeta association with neuronal cells.

Figure 1

Comparable aggregation properties of TMR-labelled and unlabelled Aβ. (A) Similar β-sheet conformation of TMR-labelled Aβ40 (closed circles) and unlabelled Aβ40 (open circles) was observed using circular dichroism. (B) Similar β-sheet conformation of TMR-labelled Aβ42 (closed squares) and unlabelled Aβ42 (open squares) was found using circular dichroism. Unstructured conformation of TMR-labelled mutant peptide (closed triangles) is also shown. (C) Thioflavin-T fluorescence indicated similar pH-dependent aggregation profiles for both TMR-labelled Aβ40 (solid line, closed circles) and unlabelled Aβ40 (dashed line, open circles). (D) Thioflavin-T fluorescence also indicated similar aggregation profiles for both TMR-labelled Aβ42 (solid line, closed squares) and unlabelled Aβ42 (dashed line, open squares) as well as very little aggregation produced by the TMR-labelled mutant peptide (dotted line, closed triangles). Electron microscopy images of TMR-labelled Aβ40 and TMR-labelled Aβ42 amyloid fibrils are shown in (E) and (F), respectively, with 100 nm scale bars. The high similarity between labelled and unlabelled peptides suggests that the addition of the fluorescent label has no observable effect on the aggregation profile of Aβ.

Figure 2

Comparison of aggregation propensities of different TMR-labelled Aβ. (A) Dynamic light scattering analysis indicated that very small particles (likely to be monomers) were present in the mutant peptide solution, two distributions of oligomers were present for Aβ40, and larger oligomers were present for Aβ42. (B) Aggregation marked by thioflavin-T fluorescence at pH 6 and pH 5 illustrated weak fluorescence for the mutant peptide (Mut), moderate aggregation for Aβ40, and the highest aggregation for Aβ42. The thioflavin-T alone sample represents the total emission peak area for a control sample not containing any Aβ. The data from these techniques indicates that the rank order of aggregation propensity is Aβ42 > Aβ40 > mutant.

Figure 3

Confocal microscopy images of various cell lines treated for 12 hours with TMR-labelled peptides. All scale bars are 20 μm in length. The fluorescent emissions of TMR-labelled peptides are indicated in red. Aβ42 displayed the greatest cell surface binding among the peptides to the differentiated neuronal cell lines (PC12, N2A and SH-SY5Y). The mutant peptide was not observed to associate with any of the cell lines tested. None of the peptides tested associated with the U937 cells.

Figure 4

Flow cytometry histogram plots showing association kinetics of TMR-labelled peptides with live NGF differentiated PC12 cells. The thick black line represents untreated control cell autofluorescence. Aβ42 showed rapid cell association, whereas mutant peptide did not appear to significantly associate with the live cells. It is interesting to note the bimodal distribution that develops over time.

Figure 5

Average kinetics of TMR-labelled peptide association with live NGF differentiated PC12 cells. Aβ42 (solid line) revealed an initial rapid binding phase within the first hour, followed by a slower binding phase over the next 24 hours. Aβ40 (dashed line) did not display the initial rapid binding phase, but was otherwise similar to Aβ42. The mutant peptide (dotted line) did not interact with the live NGF differentiated PC12 cells. This finding suggests a correlation between aggregation propensity and cell association for Aβ.

Figure 6

Flow cytometry fluorescence results indicating trypsin significantly reduces Aβ cell surface binding. (A) Cell surface treatments show that 10 μg/mL trypsin significantly decreased Aβ42 association to NGF differentiated PC12 cells, while treatments that target cell surface lipids or carbohydrates had no effect. SBTI-inactivated trypsin had little effect on Aβ42 association. (B) Concentration dependence of trypsin treated NGF differentiated PC12 cells with 0, 5, 10 and 100 μg/mL trypsin as indicated, followed by 1 hour Aβ42 treatment. (C) 10 μg/mL trypsin treatment of NGF differentiated PC12 cells demonstrate that Aβ42 association with the cell surface recovers over time.

Figure 7

Cell surface aggregation of Aβ. All scale bars are 20 μm in length. (A) Schematic diagram indicating the order of cell surface treatments on NGF differentiated PC12 cells, separated by wash steps indicated by the arrows. (B) Confocal microscopy image of cells treated with 6E10 and secondary Alexa fluor 488 labelled antibody. The absence of staining in this control indicates the primary and secondary antibodies do not non-specifically bind to the cell surface. (C) Images of cells treated with 5 μM unlabelled Aβ42 followed by immunostaining with 6E10 and secondary Alexa fluor 488 labelled antibody. (D) Images of cells treated with 5 μM unlabelled Aβ42 followed by immunostaining, and treatment with 1.5 μM TMR-Aβ42 using 488 nm laser excitation. (E) Images of cells treated with 5 μM unlabelled Aβ42 followed by immunostaining, and treatment with 1.5 μM TMR-Aβ42 using 543 nm laser excitation. (F) Merge confocal microscopy image indicating regions of 6E10 antibody only in green, TMR-labelled Aβ42 in red and co-localized regions in yellow.