CIP2A-promoted astrogliosis induces AD-like synaptic degeneration and cognitive deficits

Abstract

Reactive astrogliosis and early synaptic degeneration are 2 characteristic hallmarks in Alzheimer's disease (AD) brains, but a direct link between the 2 events has not been established. Here, we show that cancerous inhibitor of PP2A (CIP2A), a cancerous protein with high expression level in astrocytes, is upregulated in patients with AD and 3xTg-AD transgenic mice. Overexpression of CIP2A in astrocytes through adeno-associated virus infection both in cultured cells and in mice brains results in activation of astrocytes, increased production of cytokines and Aβ, and synaptic degeneration indicated by decreased levels of synaptic proteins, spine loss, and impairment in long-term potentiation. As a result of synaptic degeneration, CIP2A overexpression in astrocytes in vivo induces significant deficits in visual episodic memory detected by novel objective recognition test and spatial memory detected by Morris water maze. We conclude that CIP2A-promoted astrogliosis induces synaptic degeneration and cognitive deficits in AD.

Keywords: CIP2A; Cognitive deficit; Reactive astrogliosis; Synaptic degeneration.

Fig.1.

CIP2A is expressed in astrocytes and upregulated in AD brains. (A) Primary neurons and astrocytes from rat hippocampus were cocultured for 2 days (DIV 2) and immunostained with antibodies against CIP2A (green), Map-2 (red) and Hoechst (blue). Scale bar = 20 μm. (B) Rat primary astrocytes (DIV 3, 7, 14 and 21) were immunostained with antibodies against CIP2A (green), GFAP (red) and Hoechst (blue). Scale bar = 40 μm. (C) Left: CIP2A and GFAP expression levels in the temporal cortex from AD patients were detected by Western blotting. GAPDH was used as a loading control. Right: Quantitative analysis of the CIP2A and GFAP levels. *p < 0.05 versus Control, n = 5 control group, n = 4 AD group. (D) Left: CIP2A and GFAP expression levels in the hippocampus from 3xTg-AD and control mice (male, 9 months old) were detected by Western blotting. GAPDH was used as a loading control. Right: Quantitative analysis of the CIP2A and GFAP levels. **p < 0.01, ***p < 0.001 versus WT mice, n = 3. Data are mean ± SEM.

Fig.2.

Overexpression of CIP2A induces reactive astrogliosis. Primary rat astrocytes (DIV15) were infected with AAV-CIP2A for 5 days. (A) Immunofluorescence staining of GFAP (red) and Hoechst (blue). Scale bar = 40 μm. (B) The numbers of astrocytes were counted in control and CIP2A overexpression group, n = 9 well per group. (C) Left: CIP2A and GFAP expression levels were detected by Western blotting. DM1A was used as a loading control. Right: Quantitative analysis of the CIP2A and GFAP levels, n = 3 per group. (D) IL-1α, IL-6 and TNF-α in culture media were detected by ELISA, n = 9 well per group. (E) Aβ₄₀ and Aβ₄₂ levels in culture media were detected by ELISA, n = 9 well per group. *p < 0.05, **p < 0.01, ***p < 0.001 versus Vector. Data are mean ± SEM.

Fig.3.

CIP2A-induced astrogliosis promotes synaptic degeneration in primary neurons. Rat primary cortical neurons (DIV 10) were cultured in the conditional media from CIP2A-overexpressed asrocytes (DIV 10, infected with AAV-CIP2A for 5 days) for 4 days. (A) Cell viability was detected by LDH cytotoxicity assay kit. n = 9 well per group. (B) Left: PSD95, synapsin I and synaptophysin expression levels were detected by Western blotting using the antibodies indicated. DM1A was used as a loading control. Right: Quantitative analysis of the PSD95, synapsin I and synaptophysin levels. *** p < 0.001 versus Vector, n = 3 per group. Data are mean ± SEM.

Fig.4.

Overexpression of CIP2A in astrocytes induces visual episodic memory deficit in mice. (A) Timeline of the experiment. (B) The structure diagram of AAV-CIP2A. (C) AAV-CIP2A was injected into the hippocampal CA3 region of the C57/BL6 mice, and CIP2A expression was observed (green) 4 weeks after injection. Cell nuclei were stained with Hoechst (Blue) to show the hippocampus. Scale bar = 500 μm. (D) Brain slices in C were immunostained with anti-GFAP to show astrocytes; virus infection efficiency was obtained through calculating the ratio of co-stained (red and green) cells/ GFAP-positive (red) cells. (E) The experimental design of novel object recognition test (NOR). One day before the memory acquisition, the mice were habituated to the arenas for 5 min, as indicated in the first box. The second box showed the acquisition trial. The third and forth box showed the test trial conducted 1 h and 24 h after the acquisition trial. (F) The recognition index of object A and B in the acquisition trial. (G) The recognition index of object A and C in the test trial 1 h post the acquisition trial. *** p < 0.001 unfamiliar object C versus familiar object A. (H) The recognition index of object A and D in the test trial 24 h post the acquisition trial. *** p < 0.001 unfamiliar object D versus familiar object A. (I) The discrimination index of 1 h test trial and 24 h test trial. *** p < 0.001 versus Vector group, n = 6 mice per group. Data are mean ± SEM.

Fig.5.

Overexpression of CIP2A in astrocytes induces spatial memory deficit in mice. (A) The experimental design of Morris water maze test (MWM). (B) The representative searching trace on day 6 of the training. (C) The latency to reach the hidden platform. (D) Left: Swimming speed after removing the platform (1 h) during the probe trial. Middle: The number of times of crossing the position of the platform. Right: The percentage of time spent in target quadrant. Quadrant 3 is the target quadrant. (E) Left: Swimming speed after removing the platform (48 h) during the probe trial. Middle: The number of times to cross the position of the platform. Right: The percentage of time spent in target quadrant. *** p < 0.001 versus Vector, n = 6 mice per group. Data are mean ± SEM.

Fig.6.

Overexpression of CIP2A in astrocytes induces neuronal loss, synaptic degeneration and impairment in hippocampal LTP in mouse brains. (A) Left: Nissl staining of the brain slices from Con, AAV-Vector and AAV-CIP2A infected mouse brains. Right: Quantitative analysis of the neuron numbers in CA1 and CA3 regions of hippocampus, n = 3 brain slices per mouse, three mice per group. (B) Top: Representative dendritic spines of neurons in hippocampal CA3 region from Golgi impregnated hippocampus. Bottom: Quantitative analysis of averaged spine density (mean spine number per 10 μm dendrite segment), n = 10, from three mice per group. (C) Top: CIP2A, PSD95, synapsin I and synaptophysin expression levels were detected by Western blotting in mouse cortex. DM1A was used as a loading control. Bottom: Quantitative analysis of the protein levels, n = 3 per group. (D) Top: CIP2A, PSD95, synapsin I and synaptophysin expression levels were detected by Western blotting in mouse hippocampus. DM1A was used as a loading control. Bottom: Quantitative analysis of the protein levels, n = 3 per group. (E) Top: Normalized CA3-CA1 fEPSP mean slope recorded from the CA1 dendritic region in acute hippocampal slices. Bottom: Quantitative analysis of the normalized fEPSP slope, n = 15, 3 mice per group, 5 brain slices per mouse were recorded. * p < 0.05, ** p < 0.01, ***p < 0.001 versus Vector. Data are mean ± SEM.

Fig.7.

Overexpression of CIP2A in astrocytes induces increased production of inflammatory cytokines and Aβ in mouse brains. (A) IL-1α, IL-6 and TNF-α in brain homogenates from the cortex and hippocampus of the mice were detected by ELISA, n = 6 per group. (B) Aβ₄₀ and Aβ₄₂ levels in brain homogenates were detected by ELISA, n = 6 per group. *** p < 0.001 versus Vector. Data are mean ± SEM. (C) Cultured primary astrocytes were treated with Aβ oligomers (2 μM) for 24 h, CIP2A levels were detected by Western blotting. Left, representative blots. Right, Quantitative analysis of the protein levels, n = 3 per group. *p < 0.05 versus Con. Data are mean ± SEM. (D) Cell culture media in (C) were collected, IL-1α, IL-6 and TNF-α in culture media were detected by ELISA, n = 9 well per group. * p < 0.05, *** p < 0.001, **** p < 0.0001 versus Con. Data are mean ± SEM. (E) A working model: In AD brains, CIP2A is up-regulated in astrocytes and induces reactive astrogliosis, the latter, results in overproduction of toxic cytokines and Aβ (identified in the present study), and probably reduced release of neurotrophins, thus promotes neuronal degeneration. Aβ further stimulates CIP2A overexpression and thus forms a vicious cycle.