Endocytic uptake of monomeric amyloid-β peptides is clathrin- and dynamin-independent and results in selective accumulation of Aβ(1–42) compared to Aβ(1–40)

Intraneuronal accumulation of amyloid-β (Aβ) peptides represent an early pathological feature in Alzheimer’s disease. We have therefore utilized flow cytometry and confocal microscopy in combination with endocytosis inhibition to explore the internalisation efficiency and uptake mechanisms of Aβ(1–40) and Aβ(1–42) monomers in cultured SH-SY5Y cells. We find that both variants are constitutively internalised via endocytosis and that their uptake is proportional to cellular endocytic rate. Moreover, SH-SY5Y cells internalise consistently twice the amount of Aβ(1–42) compared to Aβ(1–40); an imaging-based quantification showed that cells treated with 1 µM peptide for 8 h contained 800,000 peptides of Aβ(1–42) and 400,000 of Aβ(1–40). Both variants co-localised to >90% with lysosomes or other acidic compartments. Dynasore and chlorpromazine endocytosis inhibitors were both found to reduce uptake, particularly of Aβ(1–42). Overexpression of the C-terminal of the clathrin-binding domain of AP180, dynamin2 K44A, or Arf6 Q67L did however not reduce uptake of the Aβ variants. By contrast, perturbation of actin polymerisation and inhibition of macropinocytosis reduced Aβ(1–40) and Aβ(1–42) uptake considerably. This study clarifies mechanisms of Aβ(1–40) and Aβ(1–42) uptake, pinpoints differences between the two variants and highlights a common and putative role of macropinocytosis in the early accumulation of intraneuronal Aβ in AD.


Fluorescence emission intensity of HF488 labelled Aβ peptides
The emission intensity was measured on 1 µM Aβ solutions to confirm equal emission intensities of the two fluorophore-conjugated Aβ variants at the same peptide concentration (Fig. S1). Spectra were recorded in a 60 µl, 3 mm path length, quartz micro-cuvette (Hellma Analytics, Müllheim, Switzerland) using a Varian Eclipse fluorimeter (Agilent Technologies, Santa Clara, CA, US). The samples were excited at 470 nm and the emission was collected between 480-700 nm. Figure S1. Emission spectra of HF488 labelled  and  at a peptide concentration of 1 µM. The data is presented as average fluorescence emission intensity of two aliquots each of  and Aβ(1-42).

Calibration curve for quantification of amyloid-β in SH-SY5Y cells
Quantification of intracellular Aβ(1-40) and Aβ(1-42) was performed using confocal imaging, through establishment of a calibration curve using Aβ solutions with known concentrations ranging from 0-10 μM. The procedures are described in the Methods section of the main text.
The calculated average pixel intensity as a function of peptide concentration is shown in Fig.   S2. The images were recorded with a pixel size of 152 nm*152 nm and a probe depth of 1.85 μm (corresponding to 2.5 axial voxels). Using these numbers, the average pixel intensity was translated to a mean intensity per Aβ molecule of 5.67. This value was thereafter used to quantify cellular uptake as displayed in Fig. 1D and explained in the main text.

Identification of the live cell cluster for flow cytometry analysis
In the quantitative peptide uptake experiments analysed by flow cytometry it was advantageous to only include viable cells in the analysis, as dead cells could have compromised membranes that may influence the extent of uptake, thereby biasing the results. It was also important to establish an experimental treatment window for endocytosis perturbation, whereby the inhibitors had significant effects without being highly cytotoxic. Supplementary

Optimisation of protocols for perturbation of endocytosis
The endocytosis inhibitors used in this study were toxic at high concentrations and/or prolonged incubation, therefore protocols needed to be optimized to find conditions where endocytosis was perturbed without concomitant cytotoxic effects. Optimisation was performed using the live/dead gates defined in Supplementary Fig. S3. Supplementary Fig. S5 shows FSC/SSC dotplots corresponding to the treatment conditions applied for endocytosis perturbation (Fig. 2-5).

Control for the transferrin pulse addition following endocytosis perturbation
The timespans of cell treatment with transferrin (Trf) (5 min) and Aβ (60 min) during endocytosis perturbation are different and effect the total time that cells are exposed to endocytosis inhibitors. We hypothesized that this may influence relative uptake as prolonged inhibitor exposure might lead to toxic effects. To test this, Trf was added in a 5 min pulse immediately after a 30 min pre-treatment period with each inhibitor, or following a 60 min pretreatment as schematically depicted in Supplementary Fig. S6A. We compared Trf uptake in control samples (0.08% DMSO), in cells treated with the macropinocytosis inhibitor IPA-3 (Trf uptake should be unaffected) and in cells treated with dynamin inhibitor dynasore (Trf uptake should be affected) ( Supplementary Fig. S6B). None or only minor effects could be observed, suggesting that toxicity due to the inhibitor does not develop during the 60 min incubation time window used for peptide incubations in this study. It was hence concluded Trf can be used as control for inhibition specificity, regardless of the pre-treatment duration.

Optimisation of CPZ concentration and inhibition of clathrin mediated endocytosis including control samples
Since CPZ was found to be rapidly degraded in aqueous solutions, we optimised the concentration of this inhibitor by adding the Trf 5-min-pulse to cells which had already been treated with Aβ for 1 h, and thereafter followed the protocol for Trf treated cells as described in the Methods section. In doing so, we avoided the risk of the inhibitor being degraded during the experiment day. The results are depicted in Supplementary Fig. S7. Based on the plateauing effect in the reduction of Trf uptake, a concentration of 5 µg/ml CPZ was chosen for further experiments.