Tau Pathology Induces Excitatory Neuron Loss, Grid Cell Dysfunction, and Spatial Memory Deficits Reminiscent of Early Alzheimer's Disease

Abstract

The earliest stages of Alzheimer's disease (AD) are characterized by the formation of mature tangles in the entorhinal cortex and disorientation and confusion when navigating familiar places. The medial entorhinal cortex (MEC) contains specialized neurons called grid cells that form part of the spatial navigation system. Here we show in a transgenic mouse model expressing mutant human tau predominantly in the EC that the formation of mature tangles in old mice was associated with excitatory cell loss and deficits in grid cell function, including destabilized grid fields and reduced firing rates, as well as altered network activity. Overt tau pathology in the aged mice was accompanied by spatial memory deficits. Therefore, tau pathology initiated in the entorhinal cortex could lead to deficits in grid cell firing and underlie the deterioration of spatial cognition seen in human AD.

Keywords: Alzheimer’s disease; cognitive deficits; electrophysiology; entorhinal cortex; grid cells; hypoactivity; neurofibrillary tangles; neuronal loss; spatial memory; tau pathology.

Figure 1. Tau Pathology is Associated with Spatial Memory Deficits in Aged EC-Tau Mice

(A) Tau pathology was identified in the EC and the hippocampal formation as well as in extrahippocampal areas of the cortex in 30+ mo EC-Tau mice. Sections from EC-Tau mice were stained with anti-tau antibodies (MC1, CP27, AT8 and AT180) and were developed using DAB as the chromagen. Tau immunoreactivity is indicated by brown staining. High magnification images of tau staining in the MEC are shown in the lower panel. (B-E) Spatial learning and memory deficits in aged EC-Tau mice. EC-Tau mice (n = 9 at 14-mo, 7 at 30+ mo) and littermate non transgenic controls (n = 10 at 14-mo, 7 at 30+ mo) were tested in the MWM (B-D) and T-maze (E). Data are expressed as mean ± the standard error of the mean (SEM). * P < 0.05 (EC-Tau 30+ mo vs Control 30+ mo on Trial Day 2 in (B), and EC-Tau 30+ mo vs Control 30+ mo in (E), ** P < 0.01 (EC-Tau 30+ mo vs Control 30+ mo on Trial Days 3-4 in (B), 2h Probe (C) and 24 h Probe (D)). A two-way repeated measures ANOVA with Bonferroni post-tests was used to compare the escape latencies in 4 days of continuous MWM hidden platform trials. Separate unpaired t-tests were used to compare the number of platform crossings in both 2 h and 24 h probe trials at 14-mo and 30+ mo. A chi-square test was used to compare the binary data (correct vs incorrect choice) from the T-maze test. See also Figure S1.

Figure 2. Aged EC-Tau Mice Exhibit Reduced Grid Field Periodicity and Firing Rates

(A-B) MEC grid cell firing rate maps and their autocorrelations are shown for 14-mo EC-Tau mice and 30+ mo EC-Tau mice along with age-matched controls. The representative firing rate maps were chosen to illustrate that grid scores appear to get worse with age as well as with increased tau accumulation. (A) Activity maps and autocorrelations were unchanged in the 14-mo EC-Tau mice compared to controls. (B) In contrast, activity maps and autocorrelations were severely affected in 30+ mo EC-Tau mice when compared to age-matched controls. Numbers on top and to the left of the activity maps indicate peak firing rate and recording depths from the surface of the brain, respectively. Numbers on top of the autocorrelation maps indicate grid scores (GS). Scale bars = 20 μm. (C) After applying an exclusion criterion of 95% (GS = 0.21) for shuffled grid scores, we found that grid cells from 30+ mo EC-Tau mice exhibit significantly reduced grid field periodicity compared to age-matched controls (30+ mo EC-Tau mice: GS=0.32, n=26 grid cells; 30+ mo control mice: GS=0.53, n=30 grid cells, P < 0.0001, Mann-Whitney U: 143.5). The group-wise histograms of observed grid score percentages across 122 grid cells show that the distribution for 30+ mo EC-Tau cells was shifted towards lower scores compared to other groups. (D) Peak firing rates and average firing rates of grid cells were reduced in 30+ mo EC-Tau mice when compared to age-matched controls (Peak Firing: 30+ mo EC-Tau mice: 4.5Hz, n=26 grid cells; 30+ mo control mice: 9.6Hz, n=30 grid cells, P < 0.0001, Mann-Whitney U: 127.5 and Average Firing: 30+ mo EC-Tau mice: 2.2Hz, n=26; 30+ mo control mice: 3.7Hz, n=30 grid cells, P = 0.0161, Mann-Whitney U: 243.0). All data are expressed as mean ± SEM. See also Figures S2, S3 and S5.

Figure 3. Grid cell properties, Increased Interneuron Firing and Enhanced Power of Theta in the Dorsal MEC of Aged EC-Tau Mice

(A) Left. Information content in the grid cells of 30+ mo EC-Tau mice was significantly reduced compared to aged-matched controls (30+ mo EC-Tau mice: Skaggs information content, 1.254; n=26 grid cells; 30+ mo control mice: Skaggs information content, 1.706; n=30 grid cells, P < 0.0001, Mann-Whitney U: 146.5). Right. There was no difference in grid cell spatial coherence in either group. (B) Putative interneurons show no difference in peak firing rates in mice at any age or genotype. However, interneurons of 30+ mo EC-Tau mice show increased average firing rates compared to age-matched controls (30+ mo EC-Tau mice: 27.7Hz, n=22 interneurons; 30+ mo control mice: 17.66Hz, n=22 interneurons, P = 0.0193, Mann-Whitney U: 142.5), while there was no difference in average firing rates in 14-mo animals. (C-D) Local field potentials (LFPs) were recorded from the MEC of (C) 14-mo and (D) 30+ mo mice. (E) Power distributions of LFPs in both 14-mo and 30+ mo mice are shown. The power in the theta range was significantly greater in 30+ mo EC-Tau mice compared to age-matched controls and 14-mo groups **P < 0.01 (two-way ANOVA).

Figure 4. The Number of Excitatory Neurons but not Inhibitory Neurons is Reduced in Aged EC-Tau Mice

Sections from 30+ mo EC-Tau mice and age-matched controls were subjected to sequential staining with TBR1 (an excitatory neuronal marker) (A), Parvalbumin (PV, an inhibitory neuronal marker) (B) or Somatostatin (SOM, an inhibitory neuronal marker) (C) antibody, followed by incubation with the MC1 tau antibody. Scale bar = 100 μm. (D) The number of TBR1⁺ excitatory neurons, but not PV⁺ or SOM⁺ inhibitory neurons were significantly reduced in MEC layers II and III/IV of 30+ mo EC-Tau mice (n = 8) compared to age-matched controls (n = 8). All data are expressed as mean ± SEM. ** P < 0.01 vs. control (non-parametric Mann-Whitney U test). (E) Tau protein colocalized with excitatory neurons, but not inhibitory neurons in the MEC of aged EC-Tau mice. Sections were subjected to sequential staining with TBR1, Parvalbumin (PV) or Somatostatin (SOM) antibody, followed by incubation with MC1 tau antibody. MC1⁺ tau staining was exclusively colocalized with TBR1⁺ excitatory neurons, but not PV⁺ or SOM⁺ inhibitory neurons in the MEC of 30+ mo EC-Tau mice (n = 4). Scale bar = 40 μm. See also Figure S4.