Mechanisms of amyloid-β34 generation indicate a pivotal role for BACE1 in amyloid homeostasis

Abstract

The beta‑site amyloid precursor protein (APP) cleaving enzyme (BACE1) was discovered due to its "amyloidogenic" activity which contributes to the production of amyloid-beta (Aβ) peptides. However, BACE1 also possesses an "amyloidolytic" activity, whereby it degrades longer Aβ peptides into a non‑toxic Aβ34 intermediate. Here, we examine conditions that shift the equilibrium between BACE1 amyloidogenic and amyloidolytic activities by altering BACE1/APP ratios. In Alzheimer disease brain tissue, we found an association between elevated levels of BACE1 and Aβ34. In mice, the deletion of one BACE1 gene copy reduced BACE1 amyloidolytic activity by ~ 50%. In cells, a stepwise increase of BACE1 but not APP expression promoted amyloidolytic cleavage resulting in dose-dependently increased Aβ34 levels. At the cellular level, a mislocalization of surplus BACE1 caused a reduction in Aβ34 levels. To align the role of γ-secretase in this pathway, we silenced Presenilin (PS) expression and identified PS2-γ-secretase as the main γ-secretase that generates Aβ40 and Aβ42 peptides serving as substrates for BACE1's amyloidolytic cleavage to generate Aβ34.

Figure 1

APP processing by β- and γ-secretases and amyloid degradation into Aβ34 at low and high BACE1/APP ratios. In the amyloidogenic pathway, sequential cleavage of APP by β-secretase and γ-secretase generates Aβ species of varying lengths including Aβ38, Aβ40 and Aβ42. In the Aβ amyloidolytic pathway, Aβ peptides resulting from the production pathway can be cleaved by β-secretase at the β34 site as part of the degradation pathway yielding the C-terminally truncated Aβ species, Aβ34. Under low BACE1/APP ratio, Aβ40 and Aβ42 levels are higher than Aβ34. In contrast, under high BACE1/APP ratio, Aβ34 levels increase as more Aβ40 and Aβ42 are degraded into Aβ34 by BACE1.

Figure 2

Aβ34 levels in AD post-mortem brain and in mouse brain tissue correlated with altered BACE1 expression and enhanced Aβ40 and Aβ42 levels. Expression of APP and BACE1, and Aβ34, Aβ40 and Aβ42 levels from post-mortem temporal brain and mouse cortices homogenates were analyzed by Western blot and MSD assays, respectively. Uncropped blots are included in a Supplementary Information file. Representative Western blot for the examination of APP and BACE1 expression levels (a). Quantification of relative protein amounts of APP (b) and BACE1 (c) of AD and non-AD. Absolute amounts of Aβ34 (d, g, and j), Aβ40 (e, h, and k) and Aβ42 (f, i, and l) determined with MSD 4-plex assays. For BACE1 knockout mice, cortices of 6 months-old 3 females and 3 males (g–i) and for London APP Transgenic mice, cortices of 6 months-old 7 females and 3 males (for WT) and 4 females and 3 males (for transgenic) (j–l) were analyzed. Data (b–f) were analyzed using unpaired Welch’s t-tests (due to violations of the normality assumption). Bars and error bars indicate mean ± s.e.m. (b) t(7) = 0.84, (c) t(13) = 3.34, (d) t(20) = 2.71, (e) t(18) = 2.54, (f) t(21) = 13.42. Data (d–f and j–l) were analyzed by unpaired t-test. Data (g–i) were analyzed by 1-WAY ANOVA and Tukey’s post-hoc tests were performed for pairwise comparisons; selected comparisons are highlighted ***p < 0.001, **p < 0.01, *p < 0.05. (g) Aβ34, 1-WAY ANOVA, F(2,15) = 10.33, p = 0.0015, (h) Aβ40, 1-WAY ANOVA, F(2,15) = 11.75, p = 0.0009, (i) Aβ42, 1-WAY ANOVA, F(2,15) = 6.637, p = 0.0086.

Figure 3

BACE1 overexpression and co-expression with APP-C99 enhanced Aβ34 production from Aβ40 and Aβ42. Expression of APP, APP-C99, and BACE1 and Aβ34 generated from endogenous levels of APP and under APP and APP-C99 overexpression conditions (wild-type APP-C99 and APP-C99 M35I mutant) were analyzed by Western blot and ELISA, respectively. Uncropped blots are included in a Supplementary Information file. HEK293T cells were transfected with indicated increasing amounts of cDNA coding for BACE1 (a) or APP695 (b) or APP-C99 and BACE1 (d) or APP-C99 M35I and BACE1 (e). Representative Western blots from 5 independent experiments for the examination of APP, BACE1, sAPPβ and sAPP_total expression (a, b, d and e). Quantification of absolute amounts of Aβ34 by ELISA (c and f). Aβ generation from BACE1 and/or APP-C99 overexpressing HEK293T cells was analyzed by ELISA, and immunoprecipitation (IP) Matrix Assisted Laser Desorption/Ionization (MALDI) mass spectrometry (MS). Cells were transfected with APP-C99, BACE1, and/or empty vector (Mock). Quantification of absolute amounts of Aβ34 (g), Aβ40 (h), and Aβ42 (i) with specific ELISAs. Aβ species were immunoprecipitated with monoclonal W02 and analyzed by MALDI-MS. Representative spectra from 3 independent experiments (j and k). Bars and error bars indicate mean ± s.e.m. Tukey’s post-hoc tests were performed for pairwise comparisons; selected comparisons are highlighted ***p < 0.001, *p < 0.05. (c) Aβ34, 1-WAY ANOVA, F(4,20) = 89.90, p < 0.0001, (f) Aβ34, 1-WAY ANOVA, F(5,24) = 28.28, p < 0.0001, (g) t-test, t(6) = 8.44, (h) t-test, t(6) = 5.16, (i) t-test, t(6) = 3.68. Linear regression was performed for the linearity test between BACE1 overexpression and Aβ34 levels (between 31.25 ng and 250 ng BACE1 DNA treatment). F(4,20) = 72.21, p < 0.0001 with the equation y = 0.3225x + 103.0.

Figure 4

Altered localization of BACE1 to the endo-lysosomal system affected Aβ34 production. Expression of BACE1 mutants and Aβ34, Aβ40 and Aβ42 levels were analyzed by Western blot and ELISA assays, respectively. Uncropped blots are included in a Supplementary Information file. Localization of BACE1 encoded by mutant constructs was analyzed by ICC. Representative Western blots of BACE1 and APP expression from 5 independent experiments with SH-SY5Y-APP (a) and SH-SY5Y-APP-C99 cells (f) transfected with different variants affecting BACE1 trafficking or mock. Absolute amounts of Aβ34 (b and g), Aβ40 (c and h), and Aβ42 (d and i). Representative ICC heatmaps of BACE1 wild-type, D495R and LL/AA in SH-SY5Y-APP (e) and SH-SY5Y-APP-C99 (j) cells from 3 independent experiments. Bars and error bars indicate mean ± s.e.m. Tukey’s post-hoc tests were performed for pairwise comparisons; selected comparisons are highlighted ***p < 0.001, **p < 0.01, *p < 0.05. (c) Aβ34, 1-WAY ANOVA, F(3,16) = 21.41, p < 0.0001, (d) Aβ40, 1-WAY ANOVA, F(3,16) = 0.2724, p = 0.8444, (e) Aβ42, 1-WAY ANOVA, F(3,16) = 0.2775, p = 0.8408, (g) Aβ34, 1-WAY ANOVA, F(3,16) = 13.20, p < 0.0001, (h) Aβ40, 1-WAY ANOVA, F(3,16) = 0.9514, p = 0.4468, (i) Aβ42, 1-WAY ANOVA, F(3,16) = 0.05190, p = 0.9838.

Figure 5

PS2 but not PS1 knockdown reduced Aβ34 levels from BACE1 overexpressing cells. Expression of PS1 and PS2 and Aβ levels were analyzed by Western blot, ELISA and MSD assays, respectively. Representative Western blots from 7 independent experiments for combinatorial PS1 and PS2 knockdown in SH-SY5Y BACE1 overexpressing cells (a). Uncropped blots are included in a Supplementary Information file. Quantification of relative amounts of PS1 (b) and of PS2 (c), and absolute amounts of Aβ34 (d), Aβ40 (e), and Aβ42 (f) in cell media by ELISA and Aβ34 (g), Aβ40 (h), and Aβ42 (i) in cell lysates by MSD. Bars and error bars indicate mean ± s.e.m. Dunnett’s post-hoc tests were performed for pairwise comparisons; selected comparisons are highlighted ***p < 0.001, **p < 0.01, *p < 0.05. (b) PS1, 1-WAY ANOVA, F(5,36) = 20.06, p < 0.0001, (c) PS2, 1-WAY ANOVA, F(5,36) = 26.37, p < 0.0001, (d) Aβ34, 1-WAY ANOVA, F(5,35) = 4.268, p < 0.005, (e) Aβ40, 1-WAY ANOVA, F(5,36) = 0.6677, p = 0.6504, (f) Aβ42, 1-WAY ANOVA, F(5,36) = 2.502, p = 0.0523, (g) Aβ34, 1-WAY ANOVA, F(5,23) = 0.8428, p = 0.5334, (h) Aβ40, 1-WAY ANOVA, F(5,24) = 2.276, p = 0.0791, (i) Aβ42, 1-WAY ANOVA, F(5,21) = 1.869, p = 0.1429.