Detection of a novel intraneuronal pool of insoluble amyloid beta protein that accumulates with time in culture

Abstract

The amyloid-beta peptide (Abeta) is produced at several sites within cultured human NT2N neurons with Abeta1-42 specifically generated in the endoplasmic reticulum/intermediate compartment. Since Abeta is found as insoluble deposits in senile plaques of the AD brain, and the Abeta peptide can polymerize into insoluble fibrils in vitro, we examined the possibility that Abeta1-40, and particularly the more highly amyloidogenic Abeta1-42, accumulate in an insoluble pool within NT2N neurons. Remarkably, we found that formic acid extraction of the NT2N cells solubilized a pool of previously undetectable Abeta that accounted for over half of the total intracellular Abeta. Abeta1-42 was more abundant than Abeta1-40 in this pool, and most of the insoluble Abeta1-42 was generated in the endoplasmic reticulum/intermediate compartment pathway. High levels of insoluble Abeta were also detected in several nonneuronal cell lines engineered to overexpress the amyloid-beta precursor protein. This insoluble intracellular pool of Abeta was exceptionally stable, and accumulated in NT2N neurons in a time-dependent manner, increasing 12-fold over a 7-wk period in culture. These novel findings suggest that Abeta amyloidogenesis may be initiated within living neurons rather than in the extracellular space. Thus, the data presented here require a reexamination of the prevailing view about the pathogenesis of Abeta deposition in the AD brain.

Figure 1

FA extraction of NT2N neurons reveals a large pool of insoluble intracellular Aβ. (A) 10-cm dishes of NT2N cells (replate 2, 4 wk old) were either sequentially extracted in RIPA followed by FA, or lysed directly into FA. Aβ1-40 and Aβ1-42 levels in the RIPA and FA samples were quantified by sandwich ELISA. Mean results and standard errors are shown (six separate experiments, each done with duplicate samples). (B) 1-42/1-40 ratios were calculated for the RIPA soluble pool of Aβ, the RIPA insoluble (FA soluble) pool of Aβ, and the total intracellular pool of Aβ for these NT2N neurons, demonstrating that the insoluble intracellular pool of Aβ consists mainly of Aβ1-42, while the soluble pool contains similar amounts of Aβ1-40 and Aβ1-42.

Figure 2

FA extraction of a variety of cell lines reveals the presence of varying levels of insoluble intracellular Aβ. (A) Uninfected and SFV-APPwt-infected NT2 cells, NT2N cells (replate 2, 4 wk old), CHO Pro5 cells, CHO-695 cells (uninfected only), and BHK-21 cells were sequentially extracted in RIPA followed by FA. Aβ 1-40 and 1-42 levels in the RIPA and FA samples were quantified by sandwich ELISA. Means and standard errors (four separate experiments done in triplicate) of Aβ levels are shown. (B) Samples from RIPA cell lysates of each of these cell lines (both uninfected and SFV-APPwt infected) were resolved on a 7.5% Tris-glycine acrylamide gel, and immunoblotted with Karen antibody; bands were detected by PhosphorImager after using an I¹²⁵-labeled secondary antibody.

Figure 2

FA extraction of a variety of cell lines reveals the presence of varying levels of insoluble intracellular Aβ. (A) Uninfected and SFV-APPwt-infected NT2 cells, NT2N cells (replate 2, 4 wk old), CHO Pro5 cells, CHO-695 cells (uninfected only), and BHK-21 cells were sequentially extracted in RIPA followed by FA. Aβ 1-40 and 1-42 levels in the RIPA and FA samples were quantified by sandwich ELISA. Means and standard errors (four separate experiments done in triplicate) of Aβ levels are shown. (B) Samples from RIPA cell lysates of each of these cell lines (both uninfected and SFV-APPwt infected) were resolved on a 7.5% Tris-glycine acrylamide gel, and immunoblotted with Karen antibody; bands were detected by PhosphorImager after using an I¹²⁵-labeled secondary antibody.

Figure 3

Aβ can be recovered from RIPA-soluble and RIPA- insoluble (FA-solubilized) fractions of cell lysate. SFV-APPwt– infected NT2N cells and CHO Pro5 cells were metabolically labeled for 12 h before being lysed in either FA or RIPA, followed by extraction of insoluble material by FA. RIPA and FA samples were subjected to immunoprecipitation with 6E10, and were resolved on a 10/16.5% step gradient Tris-tricine gel. Molecular weight standards and Aβ bands are labeled.

Figure 4

Insoluble Aβ1-42 can be produced by APPΔKK expressing NT2N and CHO cells. (A) N2TN cells and (B) CHO Pro5 cells were infected with SFV-APPwt or SFV-APPΔKK for 12 h before sequential extraction in RIPA, followed by FA. Aβ in each sample was quantified by sandwich-ELISA. Means and standard errors (two separate experiments each done in triplicate) are shown.

Figure 5

Insoluble intracellular Aβ accumulates over time within NT2N cells. 10-cm dishes of Replate 2 NT2N cells were harvested each week 3–7 wk after replating. Cells were lysed in RIPA buffer, and insoluble material was resuspended in FA. Aβ1-40 (A) and 1-42 (B) in the soluble and insoluble intracellular pools were quantified by sandwich-ELISA. Data were normalized for total protein present (in mg). The experiment was repeated three times, and means and standard errors (three to five samples per time point) are shown for a representative experiment.

Figure 6

The insoluble pool of intracellular Aβ is stable over 24 h, while the soluble pool of cellular Aβ turns over more rapidly. 10-cm dishes of Replate 2 NT2N cells were treated with 150 μg/ml cyclohexamide for the times indicated before being sequentially extracted in RIPA and FA. Aβ1-40 and 1-42 in the soluble (A) and insoluble intracellular pools (B) were quantified by sandwich-ELISA. The experiment was repeated three times, and means and standard errors (three samples per time point) are shown for a representative experiment.