Inhibiting cholesterol degradation induces neuronal sclerosis and epileptic activity in mouse hippocampus

Abstract

Elevations in neuronal cholesterol have been associated with several degenerative diseases. An enhanced excitability and synchronous firing in surviving neurons are among the sequels of neuronal death in these diseases and also in some epileptic syndromes. Here, we attempted to increase neuronal cholesterol levels, using a short hairpin RNA to suppress expression of the enzyme cytochrome P450 family 46, subfamily A, polypeptide 1 gene (CYP46A1). This protein hydroxylates cholesterol and so facilitates transmembrane extrusion. A short hairpin RNA CYP46A1construction coupled to the adeno-associated virus type 5 was injected focally and unilaterally into mouse hippocampus. It was selectively expressed first in neurons of the cornu ammonis (hippocampus) (CA)3a region. Cytoplasmic and membrane cholesterol increased, and the neuronal soma volume increased and then decreased before pyramidal cells died. As CA3a pyramidal cells died, interictal electroencephalographic (EEG) events occurred during exploration and non-rapid eye movement sleep. With time, neuronal death spread to involve pyramidal cells and interneurons of the CA1 region. CA1 neuronal death was correlated with a delayed local expression of phosphorylated tau. Astrocytes were activated throughout the hippocampus and microglial activation was specific to regions of neuronal death. CA1 neuronal death was correlated with distinct aberrant EEG activity. During exploratory behaviour and rapid eye movement sleep, EEG oscillations at 7-10 Hz (theta) could accelerate to 14-21 Hz (beta) waves. They were accompanied by low-amplitude, high-frequency oscillations of peak power at ~300 Hz and a range of 250-350 Hz. Although episodes of EEG acceleration were not correlated with changes in exploratory behaviour, they were followed in some animals by structured seizure-like discharges. These data strengthen links between increased cholesterol, neuronal sclerosis and epileptic behaviour.

Keywords: cholesterol; cytochrome P450 family 46 subfamily A; epilepsy; neurodegeneration.

Fig. 1

Regional uptake of the sh-CYP46A1 vector and effects on neuronal cholesterol. (A) Uptake of the AAV5-shRNA vector at 2 weeks after intra-hippocampal injection. The GFP (green) signal was most intense in regions occupied by the somata and apical dendrites of neurons at the border between the CA3 and CA1 hippocampal regions. (B-D) Neurons expressing the AAV5-shRNA vector (GFP, green) and related changes in cholesterol detected by filipin (red). The lower images are merged. (B) First filipin signals detected as small spots in CA3a neurones at 3 weeks after vector injection. (C) At 4 weeks, a membrane and diffuse cytoplasmic filipin staining was detected at membrane and cytoplasmic sites of GFP positive neurons. (D) At 6 weeks, strong filipin staining was evident throughout GFP containing neurons. Signs of neuronal damage included a diminished somatic volume. (E) Mean somatic area of GFP-stained CA3a neurons at 3, 4 and 6 weeks.

Fig. 2

Time course of changes in AAV5-shRNA vector expression and neuronal survival. (A) Regional distribution of GFP expression indicating the presence of the AAV5-shRNA vector at 2, 4, 8 and 12 weeks after injection. At 2 weeks the vector was expressed in a restricted zone at the border between the CA3 and CA1 regions. It spread to most of the CA1 region at 8 weeks. (B) NeuN immuno-staining at 2, 4, 8 and 12 weeks after injection. An absence of staining corresponds to a regions of neuronal loss. Neuronal death began in the CA3a region and spread to most of the CA1 region at 12 weeks. (C) Counts of neuronal density in the CA3c, CA3a and CA1 regions between 2 and 12 weeks after injection of the AAV5-shRNA vector. CA3a pyramidal cell numbers were reduced first, CA1b pyramidal cell death was delayed. No death of CA3c pyramidal cells, or dentate granule cells, was detected. (D) NeuN staining in the ipsilateral and contralateral CA1 region at 8 weeks after vector injections reveals not only a reduction of intensity in st. pyramidale, but also fewer stained somata (arrows) in the st. oriens (below st. pyramidale) and st. radiatum (above pyramidale).

Fig. 3

Glial activation and propagating neuronal death. (A) GFAP immunostaining (red) reveals astrocytes distributed throughout the hippocampus, at 4 weeks after injection of the AAV5-shRNA vector. (B) A similar distribution of GFAP staining at 10 weeks after injection. (C) In contrast, Iba1 immuno-staining (red) was restricted to regions of neuronal death, in CA3a at 4 weeks. (D) Iba1staining throughout the CA1 region at 8 weeks after vector injection. (E) Activated astrocytes stained with GFAP (red) at 4 weeks after vector injection. (F) Iba1-positive microglial cells (red) and GFP-immunopositive neurons (green) in the CA3a region at 4 weeks. (G) Larger microglial cells (red) with weakly GFP positive neurons (green) in the CA1 region at 12 weeks. Phosphorylated tau staining in the CA1 region reveals an increase at 8 weeks (H) compared to levels at 4 weeks after vector injection (I).

Fig. 4

Interictal-like EEG activity emerges after neuronal death in the CA3 region. (A) EEG activity recorded during exploration from the injected hippocampus (H-ipsi) and from ipsilateral (C-Ipsi) and contralateral cortex (C-cont) at 2 weeks after vector injection. (B) Spikes emerged during exploration in the hippocampal and cortical EEG at 6 weeks after injection correlated with the initial death of CA3a pyramidal cells. (C) EEG activity recorded during non-REM sleep at 2 weeks after vector injection. (D) More complex sequences of interictal-like events in cortical and hippocampal EEG signals recorded during non-REM sleep at 6 weeks after vector injection. (E) The frequency of simple or complex EEG spikes measured during non-REM sleep in records made at 2-10 weeks after vector injection.

Fig. 5

Acceleration of EEG oscillations during exploration. (A) An acceleration from theta to beta frequency EEG oscillations during exploration. Records from the injected hippocampus and both ipsi- and contra-lateral cortex at 7 weeks after vector injection. (B) Time-frequency analysis of this EEG episode showing the frequency acceleration. (C) Hippocampal EEG during the acceleration on an expanded time scale. (D) Video records of mouse movements show that active exploration continues throughout this episode of EEG acceleration. Frames 1-8 correspond to time points 1-8 in A. (E) Frequency of episodes of beta oscillations during exploration with time after vector injection (n=7).

Fig. 6

The emergence of high frequency, low amplitude EEG oscillations. (A) Low voltage, high frequency activity detected around the peak of beta oscillations in an hippocampal EEG record. Upper trace band-pass filtered at 0.5-500 Hz, lower trace at 200-500 Hz. (B) High frequency oscillations showed a peak frequency near 300 Hz and a rather narrow frequency distribution from 250-350 Hz. (C) They emerged with time after vector injection. Average power plotted against frequency for hippocampal EEG signals from the same animal recorded during sessions of duration 2 hours at 4, 5 and 7 weeks after vector injection.

Fig. 7

Ictal-like EEG activity during exploration. (A) An episode of accelerated beta oscillations during exploration develops into an ictal-like event. Records from the injected hippocampus and both ipsi- and contra-lateral cortex at 7 weeks after vector injection. (B) A time-frequency plot of the event shows the EEG frequency increases into the gamma band after the theta to beta transition. (C) Hippocampal EEG towards the end of the seizure on an expanded time scale. The arrows in A and C indicate large, slow EEG motifs. (D) Video records of mouse movements before and during the seizure show that active exploration is not interrupted. Frames 1-8 correspond to time points 1-8 in A. (E) Frequency of ictal-like events plotted against time after vector injection (n=4 mice). (F) Differences in frequency content of EEG frequency accelerations during exploration and seizure-like events on a relative power graph.