Alzheimer amyloid beta inhibition of Eg5/kinesin 5 reduces neurotrophin and/or transmitter receptor function

Abstract

The mechanism by which amyloid beta (Aβ) causes neuronal dysfunction and/or death in Alzheimer's disease (AD) is unclear. Previously, we showed that Aβ inhibits several microtubule-dependent kinesin motors essential for mitosis and also present in mature neurons. Here, we show that inhibition of kinesin 5 (Eg5) by Aβ blocks neuronal function by reducing transport of neurotrophin and neurotransmitter receptors to the cell surface. Specifically, cell-surface NGF/NTR(p75) and NMDA receptors decline in cells treated with Aβ or the kinesin 5 inhibitor monastrol, or expressing APP. Aβ and monastrol also inhibit NGF-dependent neurite outgrowth from PC12 cells and glutamate-dependent Ca++ entry into primary neurons. Like Aβ, monastrol inhibits long-term potentiation, a cellular model of NMDA-dependent learning and memory, and kinesin 5 activity is absent from APP/PS transgenic mice brain or neurons treated with Aβ. These data imply that cognitive deficits in AD may derive in part from inhibition of neuronal Eg5 by Aβ, resulting in impaired neuronal function and/or survival through receptor mislocalization. Preventing inhibition of Eg5 or other motors by Aβ may represent a novel approach to AD therapy.

Keywords: Abeta peptide; Alzheimer's disease; Down syndrome; Eg5; Kinesin 5; Microtubules; NMDA receptor; Neurite outgrowth; Neurodegeneration; Neurotransmitter receptor; Neurotrophin receptor; p75.

Figure 1

Aβ or monastrol reduces cell surface localization and function of neurotrophin p75 receptors. Confocal immune-microscopy revealed fewer p75 receptors (green) on the cell surface of non-permeabilized H4 cells treated with the Eg5/kinesin 5 inhibitor monastrol for 24 hrs and of non-permebilized H4APP cells overexpressing Aβ (A,B), whereas the total receptors staining did not change compared to control when the cells were permeabilized (C). WGA is shown by red. Treatment of PC12 cells with either Aβ1–40, Aβ 1–42, or monastrol also resulted in reduced cell surface expression of p75 receptors (D), which could be partly restored by over-expression of Eg5 from a transfected plasmid (E). PC12 cells treated with Aβ40, Aβ42 or monastrol grew none or a few, very short processes in response to NGF, as opposed to cells exposed to Aβ-scrambled peptide or to untreated control cells, which grew evident processes (F). Quantitation of process formation under different conditions is shown at 24 hr (G) and 48 hr (H). Blue: round cells with zero processes; Red: round cell with one or two processes shorter than cell body; Green: round or elongated cell with one or more processes longer than the cell body; Purple: star shaped, flattened cell with more processes.

Figure 2

Inhibition of Eg 5/kinesin5 by Aβ or monastrol treatment causes reduced cell surface localization of NMDA receptors in H4 cells and disrupts the microtubule network in neurons. Confocal immuno-microscopy showed reduced numbers of cell surface NMDA/NR1 (green) in H4APP cells or monastrol treated H4 cells (A, B). WGA is represented by red. When the cells were permeabilized, there was no significant difference in the relative fluorescence intensity between the treatment groups (C). Similarly, an antibody to a cytosol-facing epitope of the NMDAR2B subunit showed a significantly-reduced level of receptors in permeabilized H4APP and in H4 cells treated with monastrol (D,E).

Figure 3

Active Eg5/kinesin5 is present in primary mouse brain neurons. Immunocytochemistry confirms the presence of Eg5/kinesin 5 in cell bodies and processes of E18 neurons (A). FITC-labeled Aβ peptide is taken up into both the cell bodies and processes of primary neurons (B). Active Eg5 can be recovered from extracts of primary neuron cultures by immunoprecipitation but not from cells pre-treated with Aβ, indicating that Aβ taken up by neurons enters the cytoplasm where it can inhibit Eg5/kinesin 5 activity (C). To determine whether Aβ or monastrol reduced cell viability, a live/dead assay was performed after 48 hours of incubation (D). Neither 1 μM Aβsc (P=0.693) nor Aβ42 (p=0.419) significantly reduced viability, while 100 μM monastrol reduced viability by only 5% (p=0.007). In contrast, treatment of primary neurons with Aβ or monastrol severely disrupted the microtubule network shown by immunocytochemistry of α-tubulin (E).

Figure 4

Treatment of primary neurons with Aβ or monastrol reduces cell surface localization of NMDA and p75 receptors. Confocal immuno-microscopy showed fewer cell surface NMDA/NR1 receptor subunits on non-permeabilized primary mouse E18 neurons after exposure to Aβ 1–40, Aβ 1–42, or monastrol (A; also see S1), with some reduction also seen in permeabilized cells (B). Cross section confocal microscopy images show that the NMDA/NR1 receptors (green) are located both inside the cell and on the surface of the control cells (C–E), while the receptors are mainly inside the cells and are missing from the cell surface in Aβ42 and monastrol treated neurons (F–H).

Figure 5

Top views of NMDA/NR1 receptor subunits with confocal immunoflourscence microscopy also revealed receptor mis-localization in primary neurons treated with Aβ42 (A–D). NMDA/NR1 receptor (red) is localized both in the cell body and in the processes after treatment with scrambled Aβ peptide (A,B) for 24hr, while only aggregated, mainly peri-nuclear receptors were found after treatment with Aβ42 (C,D). Similarly, total NMDAR2B receptor subunits (red) observed in permeabilized cell bodies and processes of primary neurons treated with scrambled Aβ42 peptide were reduced and became perinuclear after treatment of the neurons with Aβ or monastrol (E). Quantitation of NMDAR2B by confocal microscopy (F) and flow cytometry (G) confirmed a reduction in receptors induced by Aβ or monastrol. Finally, p75 receptors were significantly reduced in non-permeabilized (H) but not in permeabilized primary neurons (I) after 48h treatment with Aβ42 or monastrol.

Figure 6

Glutamate-induced Ca⁺⁺ entry and LTP/neuroplasticity are reduced by Aβ or monastrol; Eg5 activity is absent in APP/PS AD mice. Glutamate induced Ca⁺⁺ entry into primary neurons was reduced by treatment with Aβ or monastrol by quantitative flourescence microscopic detection of Fura-2 (A). Hippocampal slices from non-transgenic mice were incubated in the presence of monastrol or buffer alone for 30 minutes. Subsequently, the slices were subjected to a high frequency (2 trains, 100pulses, 100Hz; HFS) stimulation, and their LTP responses were measured and found to be completely inhibited by monastol (B). Immunoprecipitation of Eg5 from extracts of brain showed recoverable activity in normal mice but no activity in APP/PS mice (C).