EPPS rescues hippocampus-dependent cognitive deficits in APP/PS1 mice by disaggregation of amyloid-β oligomers and plaques

Abstract

Alzheimer's disease (AD) is characterized by the transition of amyloid-β (Aβ) monomers into toxic oligomers and plaques. Given that Aβ abnormality typically precedes the development of clinical symptoms, an agent capable of disaggregating existing Aβ aggregates may be advantageous. Here we report that a small molecule, 4-(2-hydroxyethyl)-1-piperazinepropanesulphonic acid (EPPS), binds to Aβ aggregates and converts them into monomers. The oral administration of EPPS substantially reduces hippocampus-dependent behavioural deficits, brain Aβ oligomer and plaque deposits, glial γ-aminobutyric acid (GABA) release and brain inflammation in an Aβ-overexpressing, APP/PS1 transgenic mouse model when initiated after the development of severe AD-like phenotypes. The ability of EPPS to rescue Aβ aggregation and behavioural deficits provides strong support for the view that the accumulation of Aβ is an important mechanism underlying AD.

Figure 1. EPPS ameliorates Aβ-induced memory deficits in mice.

(a) Time course of the experiments. (b) Intracerebroventricular (i.c.v.) injection site brain schematic diagram. (c) Pretreated effects of EPPS on Aβ-aggregate-induced memory deficits observed by the % alternation on the Y-maze. EPPS, 0 (n=10), 30 (n=9) or 100 mg kg⁻¹ per day (n=10), was orally given to 8.5-week-old ICR male mice for 1 week; then, vehicle (10% DMSO in PBS, n=10) or Aβ aggregates (50 pmol per 10% DMSO in PBS; Supplementary Fig. 1A) were injected into the intracerebroventricular region (P=0.022). (d) Co-treated effects of EPPS on Aβ-aggregate-induced memory deficits observed by the % alternation on the Y-maze. Male, 8.5-week-old ICR mice received an injection of vehicle (n=9) or Aβ aggregates into the intracerebroventricular region, and then EPPS, 0 (n=10), 30 (n=10) or 100 mg kg⁻¹ per day (n=10), was orally given to these mice for 5 days. From the top, P=0.003, 0.006, 0.015. The error bars represent the s.e.m. One-way analysis of variance followed by Bonferroni's post-hoc comparisons tests were performed in all statistical analyses. (*P<0.05, **P<0.01, ***P<0.001; other comparisons were not significant).

Figure 2. EPPS rescues hippocampus-dependent cognitive deficits.

(a) Time course of behavioural tests. EPPS, 0 (TG(−), male, n=15), 10 (TG(+), male, n=11) or 30 mg kg⁻¹ per day (TG(++), male, n=8), was orally given to 10.5-month-old APP/PS1 mice for 3.5 months and their behavioural changes were compared with age-matched WT mice (WT(−), male, n=16). (b) Pre-EPPS treatment evaluation of cognitive deficits; 10.5-month-old WT mice (male, n=18) and age-matched APP/PS1 TG mice (male, n=13) were used in Y-maze tests, to obtain the % alternation before the administration of EPPS. The data indicated cognitive deficits in 10.5-month-old APP/PS1 mice (P<0.0001). (c–i) Y-maze, fear-conditioning and Morris water maze tests on 14-month-old APP/PS1 mice after EPPS administration for 3.5 months total. (c) Per cent alternation on Y-maze. From the top, P=0.000, 0.020, 0.008, 0.010. (d) Total entry number into each arm of the Y-maze test. (e) Per cent total freezing from contextual fear conditioning. From the top, P=0.000, 0.036, 0.038. (f) Per cent total freezing in the cued task, P=0.025. (g) Hidden platform test (significances, see Supplementary Table 3) and (h) the probe test in the Morris water maze, P=0.003, 0.033. (i) Swim speeds of the probe test (crossing number of located hidden platform analysis, see Supplementary Fig. 2I). (j) Dose-dependent evaluation of EPPS-induced memory alterations. EPPS was orally given to 12-month-old APP/PS1 TG male mice in 0, 0.1, 1 or 10 mg kg⁻¹ per day (n=7–9) dosages for 3 months. Per cent alternation on Y-maze (P=0.005, 0.041). The error bars represent the s.e.m. One-way analysis of variance followed by Bonferroni's post-hoc comparisons tests were performed in all statistical analyses (*P<0.05, **P<0.01, ***P<0.001; other comparisons were not significant).

Figure 3. EPPS does not affect synaptic plasticity in mice.

EPPS, 0 (EPPS−) or 30 mg kg^-1 per day (EPPS++), was orally given to WT (n=4) and APP/PS1 TG (n=4) mice for 5 days. (a–d) LTP, measured by % field EPSP% fEPSP slope, in CA1 of hippocampal slices (three slices per mouse) from WT mice. (a) Upper trace, a representative trace of fEPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus in non-treated mice. Lower trace, a time course of EPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus in non-treated mice. (b) Upper trace, a representative trace of EPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus (three slices per mouse) in EPPS-treated mice. (c) Averaged time course of EPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus in non-treated and EPPS-treated mice. (d) Quantification of the effect of EPPS on LTP. (e–h) LTP, measured by the % fEPSP slope, in CA1 of hippocampal slices from TG mice. (e) Upper trace, a representative trace of EPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus in non-treated mice. Lower trace, a time course of EPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus in non-treated mice. (f) Upper trace, a representative trace of EPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus in EPPS-treated mice. (g) Averaged time course of EPSP before and after inducing LTP by pairing stimuli in CA1 of the hippocampus in saline-treated and EPPS-treated mice. (h) Quantification of EPPS effect on LTP in TG mice. LTP was induced by theta burst stimulation (TBS), represented by the arrow. The error bars represent the s.e.m. Student's unpaired t-tests were performed in statistical analyses; comparisons were not significant (P>0.05).

Figure 4. EPPS disaggregates Aβ plaques and oligomers in APP/PS1 mice.

APP/PS1 mice and WTs from the aforementioned behavioural tests were killed and subjected to brain analyses. EPPS, 0 (TG(−), male, n=15), 10 (TG(+), male, n=11) or 30 mg kg^-1 per day (TG(++), male, n=8), was orally given to 10.5-month-old APP/PS1 for 3.5 months and their brains were compared with age-matched WT brains (WT(−), male, n=16). (a) ThS-stained Aβ plaques in whole brains (scale bars, 1 mm) and the hippocampal region (scale bars, 200 μm) of each group. The mouse brain schematic diagram was created by authors (green and red boxes: regions of brain images, a and f, respectively). (b) Number or area of plaques normalized (%) to the level in 10.5-month-old TG mice. Plaque number: P-values compared with TG (male, 10.5-month-old) are all <0.0001 (#). P-values compared with TG(−) (male, 14-month-old) are all <0.0001 (*). Plaque area: P-values compared with TG (male, 10.5-month-old) are all <0.0001 (#). P-values compared with TG(−) (male, 14-month-old) are all <0.0001 (*). (c–e) Aβ-insoluble and -soluble fractions analyses from brain lysates. (c) Sandwich ELISA of Aβ-insoluble fractions. Hippocampus: all P<0.0001; cortex: P=0.004, 0.046. (d) Sandwich ELISA of Aβ-soluble fractions. (e) Dot blotting of the total Aβ (anti-Aβ: 6E10, also recognizes APP) and oligomers (anti-amyloidogenic protein oligomer: A11). (f) Histochemical analyses of Aβ deposition. Aβs were stained with the 6E10 antibody and ThS. Aβ plaques (first row): green; all Aβs (second row): red; 4,6-diamidino-2-phenylindole (DAPI): blue (as a location indicator). The third and bottom rows show merged images of plaques and Aβs, and plaques and Aβs with DAPI staining. Scale bars, 50 μm. (g) Western blotting analyses of APP expression in hippocampal and cortical lysates (detected at ∼100 kDa by 6E10 antibody). Densitometry (see Supplementary Fig. 3A). Full version (see Supplementary Fig. 7). The error bars represent the s.e.m. One-way analysis of variance followed by Bonferroni's post-hoc comparisons tests were performed in all statistical analyses (*P<0.05, **P<0.01, ***P<0.001, ^#P<0.05, ^##P<0.01, ^###P<0.001; other comparisons were not significant).

Figure 5. EPPS lowers inflammation and glial GABA release.

APP/PS1 and WT mice from the aforementioned behavioural tests were killed and subjected to brain analyses. EPPS, 0 (TG(−), male, n=15), 10 (TG(+), male, n=11) or 30 mg kg⁻¹ per day (TG(++), male, n=8), was orally given to 10.5-month-old APP/PS1 for 3.5 months and their brains were compared with age-matched WT brains (WT(−), male, n=16). (a) Western blotting of phosphorylation of cyclic AMP response element-binding protein (pCREB), pJNK, GFAP and Iba-1 (densitometry, see Supplementary Fig. 3B, C). Full version (see Supplementary Fig. 8). (b–e) Histochemical analyses of Aβ deposition, GFAP, Iba-1 and GABA. (b) Aβ plaques stained with ThS (first row): blue; GFAP (second row): green; Iba-1 (third row): red; and 4,6-diamidino-2-phenylindole (DAPI; fourth row): blue (as a location indicator). The bottom row shows merged images of plaques, GFAP and Iba-1 with DAPI staining. (c) Aβ plaques stained with ThS (first row): green; GFAP (second row): blue; and GABA (third row): violet. Scale bars, 50 μm. (d) Quantification of GABA in confocal images. Each dot represents the number of GFAP-positive cells with GABA; a.u., arbitrary unit. (e) GABA average from the previous panel (P<0.0001 for all). Values refer to GFAP-positive GABA. The error bars represent the s.e.m. One-way analysis of variance followed by Bonferroni's post-hoc comparison tests were performed in the statistical analysis (***P<0.001; other comparisons were not significant). The mouse brain schematic diagram was created by authors (green box: region of brain images).

Figure 6. EPPS disaggregates Aβ aggregates by selective binding.

(a–c) Preformed Aβ42 aggregates (25 μM, Day 0) were incubated with EPPS. (a) EPPS concentration-dependent (200, 20, 2, 0.2, 0.02, 0.002, 0.0002 or 0 mM of EPPS, 3-day treatment) and incubation time-dependent (20 mM EPPS for 1, 2, 3 and 7 days) ThT assays. Fluorescence intensity was normalized to preformed Aβ aggregates (100%, Day 0). Statistical comparisons were made with day 0 (n=4, Student's t-test; from the left: P=0.034, 0.004, 0.005, 0.000, 0.000). (b) Transmission electron microscopic images of EPPS-induced Aβ fibril disassembly. Scale bars, 200 nm. The inset shows the chemical structure of EPPS. (c) Silver staining for the SDS–PAGE analysis of PICUP cross-linked Aβ aggregates and the densitometry analysis in the ratio of monomer to fibril (HMW, high molecular weight; LMW, low molecular weight). Full-length version (see Supplementary Fig. 9). (d) SEC analysis. Size markers: BSA (yellow) and thioredoxin (Trx, pink). Control: Aβ42 monomer (green). a.u., arbitrary unit. (e) Surface plasmon resonance analyses. Dose-dependent kinetics of EPPS targeting Aβ40 oligomers and the corresponding fitting curve from the saturated region of the sensorgram. (f) MTT assays. Aβ42: 2.5 μM Aβ42 aggregates, Aβ42(7d): 2.5 μM Aβ42 aggregates were incubated for 7 days with/without EPPS (2 mM). HT-22 cells were treated with the prepared samples for 24 hours. Cell viability was normalized to that of the non-treated cells (100%). All P-values were <0.0001 (n=5). The error bars represent the s.e.m. of independent triplicate measurements. One-way analysis of variance followed by Bonferroni's post-hoc comparison tests were performed in the statistical analyses (*P<0.05, **P<0.01, ***P<0.001; other comparisons were not significant).