Small-molecule inducers of Aβ-42 peptide production share a common mechanism of action

Abstract

The pathways leading specifically to the toxic Aβ42 peptide production, a key event in Alzheimer's disease (AD), are unknown. While searching for pathways that mediate pathological increases of Aβ42, we identified Aftin-4, a new compound that selectively and potently increases Aβ42 compared to DMSO (N2a cells: 7-fold; primary neurons: 4-fold; brain lysates: 2-fold) with an EC(50) of 30 μM. These results were confirmed by ELISA and IP-WB. Using affinity chromatography and mass spectrometry, we identified 3 proteins (VDAC1, prohibitin, and mitofilin) relevant to AD that interact with Aftin-4, but not with a structurally similar but inactive molecule. Electron microscopy studies demonstrated that Aftin-4 induces a reversible mitochondrial phenotype reminiscent of the one observed in AD brains. Sucrose gradient fractionation showed that Aftin-4 perturbs the subcellular localization of γ-secretase components and could, therefore, modify γ-secretase specificity by locally altering its membrane environment. Remarkably, Aftin-4 shares all these properties with two other "AD accelerator" compounds. In summary, treatment with three Aβ42 raising agents induced similar biochemical alterations that lead to comparable cellular phenotypes in vitro, suggesting a common mechanism of action involving three structural cellular targets.

Figure 1.

A) Structure of Aftin-4 (top panel) and roscovitine (bottom panel). The asterisk highlights the difference. B) N2a-695 cells were treated with Aftin-4 or roscovitine (100 mM) or control DMSO. Aβ40 and Aβ42 levels (ELISA) are expressed as fold change over the level of control. C) Confirmation of this effect by IP and WB analysis. Data for ≥3 experiments (means±se) were compared with the control. D) N2a-695 cells were treated with Aftin-4, fenofibrate (feno.), celecoxib (celo.), or control DMSO at the indicated concentrations. Data for ≥3 experiments (means±se) were compared (GraphPad Prism 4.0) with the control conditions (DMSO). *P < 0.05, **P < 0.01, ***P < 0.001; 1-tailed Student's t test, 95% significance level.

Figure 2.

A) Embryonic hippocampal cultures were treated with Aftin-4 (100 μM) or DMSO. Aβ40 and Aβ42 levels in the supernatants were measured (ELISA). B) Embryonic cortical cultures were treated with Aftin-4 (100 μM) or DMSO. Aβ40 and Aβ42 levels in the supernatants were measured (ELISA). C) Dose-response analysis of the effect of Aftin-4 on Aβ42 in primary cortical cultures (ELISA). Data for ≥3 experiments (means±se) were compared (GraphPad Prism 4.0) with the control conditions (DMSO). *P < 0.05, **P < 0.01, ***P < 0.001; 1-tailed Student's t test, 95% significance level.

Figure 3.

A) Brain homogenates from wild-type mice (n=3) were treated with Aftin-4 or DMSO. CT695 antibody was used to detect the level of β-CTF. NT, nontreated. B) N2a cells transfected with a plasmid encoding APP-Swedish cDNA and exposed to Aftin-4 (100 μM), fenofibrate (feno.; 100 μM) or celecoxib (celo.; 25 μM) were lysed and analyzed by WB to detect β-CTF (6E10 antibody). C) Cells were exposed to Aftin-4 (50 μM) and to DAPT or BMS-299897, as indicated. Aβ42 levels were measured after 18 h; fold increases induced by Aftin-4 compared to nontreated cells are represented. D) N2a-695 cells were treated for 3 h with various amounts of Aftin-4, and the corresponding cell lysates were analyzed by SDS-PAGE and WB using a cleaved Notch antibody, as well as a total Notch antibody. One representative experiment is shown. E) Quantification of the results of 3 independent experiments (means±se). Results are indicated as a ratio of cleaved Notch over total Notch. Normalized cleaved Notch treated with DMSO is arbitrarily set at 100. *P < 0.05, **P < 0.01; 2-tailed paired t test.

Figure 4.

A) Identification of roscovitine and Aftin-4 targets by 2-D DIGE analysis. N2a-695 cell lysates were incubated for 20 min with immobilized roscovitine or Aftin-4 on Sepharose beads. Bound proteins were stained with Cy3 for Aftin-4 (left panel) or Cy5 for roscovitine (right panel), then analyzed by 2-D DIGE. Circled blue spots are proteins that interact 2.5 times more with Aftin-4; circled red spots are proteins that interact 2.5 times more with roscovitine; circled green spots are proteins that interact with both. Rectangles indicate proteins identified mostly with Aftin-4. B) Areas used to calculate volume ratio of Aftin-4/roscovitine are shown in yellow (VDAC1: +46.48; prohibitin: +10.55; mitofilin isoforms: +6.77, +14.78, and +4.18; MAPK: −29.98). C, D) WB analysis of proteins bound to Aftin-4 and roscovitine Sepharose beads is shown in triplicate, using antibodies directed against mitofilin, VDAC1, prohibitin, and CDK5 or APP and PS1-NTF.

Figure 5.

A) N2a-695 cells treated for 3 h with Aftin-4 (100 μM) or roscovitine (100 μM) were analyzed by electron microscopy. Two representative fields for each condition are shown. B) N2a-695 cells treated 3 h with fenofibrate (100 μM) or celecoxib (25 μM) were analyzed by electron microscopy. Two representative fields for each condition are shown. C, D) N2a-695 cells were treated for 3 h with 100 μM of Aftin-4. Cellular medium was then replaced with medium supplemented with Aftin-4 (C) or DMSO (D) for an additional 3 h. Cells were fixed and analyzed by electron microscopy. Original views: ×16,000.

Figure 6.

A) Subcellular localization of various proteins was investigated by sucrose density gradient fractionation. N2a-695 cells treated with DMSO (left panel) or 100 μM Aftin-4 (right panel) were fractionated by velocity sedimentation. The distribution of proteins was analyzed by WB analysis. Cellular fractions corresponding to Golgi and mitochondria/endoplasmic reticulum (ER) are indicated. B) The Aβ42 content of each cellular fraction was also tested by ELISA assay. C) 500 μM of free Aftin-4, fenofibrate (feno.), celecoxib (celo.), or control DMSO were added to N2a-695 cell lysates for 20 min and then incubated with Aftin-4 immobilized on Sepharose beads After extensive washing, the bound proteins were analyzed by SDS-PAGE and WB analysis using VDAC1, prohibitin, or mitofilin antibodies. D) N2a-695 cells treated with DMSO, 100 μM Aftin-4, or 100 μM fenofibrate were fractionated by velocity sedimentation. Distribution of VDAC1 was analyzed by WB.