Electrical and Network Neuronal Properties Are Preferentially Disrupted in Dorsal, But Not Ventral, Medial Entorhinal Cortex in a Mouse Model of Tauopathy

Dorsoventral gradients of intrinsic membrane properties in WT and rTg4510 layer II mEC-SCs. A, Traces are whole-cell current-clamp recordings from layer II mEC-SCs in dorsal (top) or ventral (bottom) locations in WT (black lines) or rTg4510 (blue lines) slices. The traces show the mean (±SEM, shaded areas) voltage response to a 3 s, −100 pA square-wave current injection from a fixed potential of −80 mV. The inset traces (right) show the initial response on an expanded time scale to illustrate the initial charging curve. Calibration: 1 s, 5 mV; inset, 10 ms. The scatter plots show R_i, C_M, τ_M, RMP, and the percentage sag from all recorded mEC-SCs at dorsal (D) and ventral (V) locations. Solid symbols are the mean ± SEM (WT mEC-SCs: D, n = 18; V, n = 28; rTg4510 mEC-SCs: D, n = 19; V, n = 19 cells from 13 WT and 11 rTg4510 mice). There was a significant difference in C_M in dorsal mEC-SCs. *p < 0.05. B, Traces are the mean (±SEM) peak-aligned waveforms of APs recorded from dorsally (top) and ventrally (bottom) located layer II mEC-SCs in WT (black lines) and rTg4510 (blue lines) slices. The analyzed APs were the first of a train elicited by a 0.5–1 nA current injection. The scatter plots show the AP peak, threshold, width at threshold, and maximum rate of rise (dV/dt).

Dorsoventral gradient in mAHP magnitude is abolished in rTg4510 layer II mEC-SCs. A, Average waveforms of AHPs elicited by a train of 5, 10, 15, 20, or 25 APs (lighter through to darker waveforms, respectively) evoked by brief, strong (2 ms, 2 nA) current injections delivered at 50 Hz. AHPs are plotted on a log–time axis to differentiate the mAHP and sAHP, respectively. The mAHP detection window was up to 0.1 s after the end of AP firing (dotted red line). The dotted black line represents the pre-AP membrane potential. Recordings were made at a fixed membrane potential of −80 mV. Note the significantly larger mAHP in dorsal mEC-SCs recorded from rTg4510 slices. B, Graphs show the mean (±SEM) mAHP amplitude in WT and rTg4510 mEC-SCs. A repeated-measures ANOVA revealed a significant main effect of AP number (F = 54.9, p < 0.001), as well as significant main effects of dorsal–ventral location and genotype, and a significant location-by-genotype interaction (Table 1). C, Representative traces showing the steady-state firing patterns in dorsal and ventral WT and rTg4510 layer II mEC-SCs. Raster plots show the timing of spikes; spikes occurring in a cluster are shown in red. Calibration: 20 mV, 2 s. D, Box plots showing the mean (filled diamonds) ±SEM (solid box lines) p_Clus and mAHP half-widths. The error bars show the interquartile range, and the dotted horizontal line is the median. Dorsal rTg4510 mEC-SCs are significantly less likely to fire spikes in a cluster and have a significantly longer mAHP half-width when compared with WT counterparts. *p < 0.05 (independent-samples Mann–Whitney U test).

Gamma (30–80 Hz) oscillations in layer II of the dorsal, but not ventral, entorhinal cortex are preferentially disrupted in rTg4510 mice. A, Traces are simultaneously recorded extracellular gamma oscillations evoked by bath application of 500 nm kainate in representative WT (Ai) and rTg4510 (Aii) slices. These recordings were made ∼15 min after the start of drug application. Gray traces are raw data; the black (dorsal) and red (ventral) traces are bandpass filtered (30–80 Hz). Calibration: 50 μV, 50 ms. B, Power spectra and autocorrelograms of the recordings shown in A. In WT slices (Bi), gamma oscillations were faster in dorsal (black) versus ventral (red) regions, both in terms of peak spectral frequency and the primary nonzero peak in the autocorrelogram. In rTg4510 slices (Bii), gamma oscillations in the dorsal mEC had significantly reduced power compared with WT slices. C, Pooled time course data showing the development of gamma oscillations following the continuous bath application of kainate (start point is indicated by arrow). Solid lines show the mean (±SEM; shaded areas) gamma power (in dB) and frequency of oscillations in 60 s bins. D, Summary data showing the maximum gamma power (Di) and frequency (Dii) in WT and rTg4510 slices. Note the selective effects of transgene expression on dorsal gamma oscillations. *p < 0.05, **p < 0.01.

Parvalbumin expression in the mEC is normal in rTg4510 mice. Photomicrographs showing representative images of a sagittal section of the mEC from WT (top) and rTg4510 (bottom) mice. The images on the left are sections stained with a PV-specific antibody, and the images on the right are sections stained by PG5, an antibody selective for phosphorylated tau. For the purposes of analysis, layers II and III of the mEC were subdivided into four regions of interest (labeled A–D, where A is most dorsal and D is most ventral). Note the strong gradient in PV staining in both the WT and rTg4510 sections. On the rTg4510 image (bottom left), a lobe of the cerebellum, with extremely dense PV staining is visible. The box plots illustrate the PV (top) and PG5 (bottom) staining in regions A–D in layers II and III in 6 μm sections from five WT and five rTg4510 mice. The mean is denoted by the filled diamonds, the SEM by the solid box lines, the median by the dashed line, and the interquartile range by the whiskers; and open symbols are data from individual mice.

Flattened dorsoventral gradients in gamma-frequency activity in awake behaving rTg4510 mice. A, Light microscopy images of parasagittal sections through the mEC of a WT and an rTg4510 mouse, which were implanted with 16-channel linear silicone probes. The position of the probe, as determined by post hoc analysis of the lesion marks (red dots), is shown on each image. Scale bar, 1 mm. Three example LFP recordings are shown to the right of each image, with the location of the recording site illustrated by the arrows. Note the prominent theta oscillation with superimposed gamma frequency oscillations occurring toward the peak of each theta cycle. Calibration: 250 ms, 500 μV. B, Pooled power spectra (mean across all sites within the mEC and then averaged across animals) from epochs of LFP recorded while mice were running at 10–15 cm/s (left). Pooled data (right) showing the significant reduction in broadband power (Total) in rTg4510 mice (n = 5) compared with WT mice (n = 3), as well as specific reductions in the theta and gamma bands. *p < 0.05, **p < 0.01. C, Averaged comodulograms showing the extent of theta–gamma cross-frequency coupling along the dorsal (top) to ventral (bottom) axis. Since broadband power is reduced in rTg4510 mice, the right column shows the mean rTg4510 comodulogram with the color scale increased by 10-fold so that theta–gamma cross-frequency coupling is discernible. D, MI plotted as a function of distance from the most dorsal electrode from example WT (left) and rTg4510 (right) mice. E, Regression analysis reveals a significantly (**p < 0.01) lower z-normalized correlation coefficient (Pearson's R; Ei) and slope (Eii) in rTg4510 mice compared with WT mice. Filled circles represent data from each individual animal; the horizontal line is the mean, and the error bars show the SD. F, Pooled MI normalized to the most ventral electrode plotted as a function of dorsoventral location.