Quantitative Attribution of the Protective Effects of Aminosterols against Protein Aggregates to Their Chemical Structures and Ability to Modulate Biological Membranes

Abstract

Natural aminosterols are promising drug candidates against neurodegenerative diseases, like Alzheimer and Parkinson, and one relevant protective mechanism occurs via their binding to biological membranes and displacement or binding inhibition of amyloidogenic proteins and their cytotoxic oligomers. We compared three chemically different aminosterols, finding that they exhibited different (i) binding affinities, (ii) charge neutralizations, (iii) mechanical reinforcements, and (iv) key lipid redistributions within membranes of reconstituted liposomes. They also had different potencies (EC₅₀) in protecting cultured cell membranes against amyloid-β oligomers. A global fitting analysis led to an analytical equation describing quantitatively the protective effects of aminosterols as a function of their concentration and relevant membrane effects. The analysis correlates aminosterol-mediated protection with well-defined chemical moieties, including the polyamine group inducing a partial membrane-neutralizing effect (79 ± 7%) and the cholestane-like tail causing lipid redistribution and bilayer mechanical resistance (21 ± 7%), linking quantitatively their chemistry to their protective effects on biological membranes.

Figure 1

Structural formulas of the three AMs investigated in this work, and mode of membrane insertion and perturbation for TRO. (A) Chemical structures of TRO (blue), SQ (red), and ENT-03 (green). (B) Schematic representation of the insertion and localization of TRO within biological membranes, as previously determined experimentally. The 55° angle refers to the whole molecule, rather than the steroid group or polyamine group only. The 14.8 ± 0.2 and 5.6 ± 0.2 Å distances refer to the space occupied by the molecule and portion sticking out of the membrane along the normal to the bilayer plane, respectively. (C) Schematic representation of the three major physico-chemical effects on cell membranes induced by the insertion of TRO, all possibly mediating the TRO-induced protection against the toxicity of misfolded protein oligomers.

Figure 2

Changes in fluorescence anisotropy and emission of the three fluorescently labeled AMs in the presence of LUVs. (A, B) Fluorescence anisotropy (r) values at 570 nm for BODIPY (A) and at 617 nm for A594 (B), of 10 μM L-Arg (gray), TRO (blue), SQ (red), and ENT-03 (green) labeled with BODIPY (A) and A594 (B), obtained in the absence and presence of 0.5 mg/mL LUVs. (C, D) Fluorescence emission corresponding to the integrated area between 550–650 nm for BODIPY (C) and 600–700 nm for A594 (D), of 10 μM L-Arg (gray), TRO (blue), SQ (red), and ENT-03 (green) labeled with BODIPY (C) and A594 (D), obtained in the absence and presence of 0.5 mg/mL LUVs. Bars indicate mean ± SEM (n = 5. n.s., nonsignificant; ***, p < 0.001 relative to corresponding values in the absence of LUVs (Student’s t-test).

Figure 3

Binding of the three AMs to LUVs. Binding plots reporting the fluorescence emission at 572 nm (A) and 612 nm (B), of 10 μM BODIPY (A) and A594 (B) labeled TRO (blue), SQ (red), ENT-03 (green), and L-Arg (gray) versus LUV concentration. The lines through the data points represent the best fits to eq 6. Each graph reports the obtained K_D value in units of mg/mL and μM of total lipids. Experimental errors are SEM (n = 5).

Figure 4

Light scattering intensity of LUVs in the presence of increasing concentrations of the three AMs. Plots reporting the square root of the light scattering intensity of LUVs with (I_AM) and without (I_NO) AM, respectively, versus AM concentration, representing the increase in LUV mass due to TRO (A), SQ (B), and ENT-03 (C) incorporation. The indicated AM concentrations correspond to the values reported on the x axis at saturation points. Experimental errors are SEM (n = 3).

Figure 5

The three AMs increase the transition temperatures of LUVs. (A) Zeta potential (ζ) and (B) fitted curves (left axis) and first derivative curves (right axis) of ζ values as a function of temperature for naked LUVs (black square), and AM-containing LUVs: TRO (blue), SQ (red), and ENT-03 (green). Experimental errors are standard deviations (n = 5). In each case, the T_m corresponds to the minimum of the first derivative curve.

Figure 6

The three AMs increase the mechanical resistance of lipid bilayers to indentation or BTF. Breakthrough force (BTF) distributions measured on SLBs formed from LUVs prepared in the absence (A, gray) and in the presence of 5 μM TRO (B, blue), SQ (C, red), and ENT-03 (D, green). Distributions were obtained from at least six independent force maps. The statistically significant difference between AMs was calculated using a Kruskal-Wallis test, which resulted in p < 0.001, and a Dunn test, which highlighted a difference between each group with p < 0.05.

Figure 7

The three AMs redistribute CHOL and GM1 molecules in LUVs. (A–D) Fluorescence spectra of LUVs containing the indicated D-labeled lipid (green), A-labeled CHOL (orange), and both (blue) in the absence (A) and in the presence of TRO (B), SQ (C), and ENT-03 (D). (E) FRET efficiency (E) values obtained for the various pairs using eq 12 in the absence (gray) and presence of TRO (blue), SQ (red), and ENT-03 (green). Experimental errors are SEM (n = 5). The symbols *** refer to p values of <0.001 relative to r values obtained in the absence of AMs. (F) Mean shortest distances (r) between the indicated lipid-D and CHOL-A in absence and presence of AMs obtained from FRET E values reported in panel E using eq 13.

Figure 8

Far UV CD analysis of α-synuclein displacement exerted by the three AMs. (A–C) Far-UV CD spectra of αS in the absence and presence of DMPS LUVs incubated with increasing concentrations of TRO (A), SQ (B), and ENT-03 (C). Spectra were blank subtracted and normalized using eq 14. (D, E) Mean residue ellipticity at 222 nm (D) and 192 nm (E) of αS incubated with DMPS LUVs and increasing concentrations of TRO (blue), SQ (red), and ENT-03 (green).

Figure 9

The three AMs bind to the plasma membrane of SH-SY5Y cells and prevent the increase of intracellular Ca²⁺ levels induced by ADDLs (global fitting analysis). (A) Representative confocal microscopy images (median planes parallel to the coverslip) of SH-SY5Y cells incubated for 30 min at room temperature with 5 μM of L-Arg-BODIPY, TRO-BODIPY, SQ-BODIPY, or ENT-03-BODIPY (probe:molecule of 1:10). Blue and red fluorescences indicate Hoechst-labeled nuclei and AM-BODIPY, respectively. (B) Representative confocal scanning microscopy images of free Ca²⁺ levels in untreated SH-SY5Y cells or in cells treated for 15 min with 1 μM ADDLs in the absence or presence of 1 μM AMs. (C) Intracellular free Ca²⁺-derived fluorescence in untreated SH-SY5Y cells or in cells treated for 15 min with ADDLs in the absence or presence of the indicated concentrations of AMs. Experimental errors are SEM (n = 4). *** symbols refer to p values <0.001 relative to untreated cells. §§ and §§§ symbols refer to p values <0.01 and < 0.001, respectively, relative to ADDLs without AMs. (D) Normalized dose–response curve obtained from Ca²⁺-derived fluorescence data of all AMs in panel C and fitted to eq 2. (E) Plot reporting theoretical versus experimental response values obtained from eq 2. (F–H) Normalized dose–response curves for TRO (blue), SQ (red), and ENT-03 (green) obtained from Ca²⁺-derived fluorescence data in panel C. In each plot, the lines through the data do not represent independent fitting procedures using eq 2, but result from global fitting using only eq 4, in all cases with corresponding values of ζ potential, BTF, and FRET data. Experimental errors are SEM (n = 4). (I) Plot reporting theoretical versus experimental response values obtained from eq 4.

Figure 10

Contributions of the various membrane alterations and chemical groups of AMs to their potency in this experimental setting. (A) Contributions of the various types of membrane alterations and (B) of the chemical groups of AMs to the EC₅₀ parameter in this experimental setting. The numbers reported in panel B refer to the contributions to the AM potency in membrane protection of SH-SY5Y cultured cells against misfolded protein oligomers of Aβ (ADDLs, 1 μm monomer equivalents) causing Ca²⁺ influx. The chemical formula in the image refers to TRO. (C) Representative AMs with their potencies (EC₅₀ values) predicted in our experimental setting using the values reported in panel B. The EC₅₀ values predicted for TRO, SQ, and ENT-03 are in good agreement with those observed experimentally.