Thalamic reticular control of local sleep in mouse sensory cortex
- University of Lausanne, Switzerland
Figures
Figure 1
Identification of TRN sectors through anterograde tracing.
(A) Scheme illustrating injection of AAV-ChR2_EYFP into S1 (top), AC (middle) and MD (bottom). (B) Epifluorescent micrographs of coronal sections at 2.5x magnification, anteroposterior position indicated with respect to bregma (Bg). ChR2-EYFP (green) expression in infected S1 neurons and their projection to TRN and VPM (top), in AC neurons and their projection to medial geniculate nucleus (MG) (middle) and in MD neurons (bottom), with immunostaining for PV-positive (PV+, magenta) TRN neurons. The VPM appears in light magenta due to its innervation by TRN. (C) Expanded view of the target TRN sector at 5x magnification. White arrowheads indicate sites of projection into TRN.
Figure 2 with 1 supplement
Oscillatory burst firing varies across TRN sectors and depends on CaV3.3 Ca2+ channels.
(A) Representative traces of paired EPSCs in whole-cell voltage-clamped TRN neurons at −60 mV of WT and CaV3.3 KO mice upon light activation of S1 (top), AC (middle) and MD (bottom) afferents. Afferent-specific forms of short-term plasticity are preserved across genotypes (see Results for averaged data). (B) Box-and-whisker-plots of the mean capacitance (Cm) and the resting membrane potential (VRMP) of the recorded TRN neurons (WT: n = 12 for S1, n = 13 for AC, n = 14 for MD; CaV3.3 KO: n = 12 for S1, n = 13 for AC, n = 11 for MD). It can be noted that the Cm of S1-innervated TRN neurons in CaV3.3 KO showed a reduction compared to WT, suggesting smaller cell size (Mann-Whitney test, p=0.007). (C) Representative current-clamp recordings of oscillatory bursting responses of TRN neurons across sectors, induced through hyperpolarizing current injections (−50 to −300 pA for 500 ms). Horizontal lines denote −60 mV. Note the strong repetitive burst firing in sensory sectors that is impaired in the CaV3.3 KO cells, whereas MD-innervated cells mostly discharge a single burst. (D) Graph of the number of repetitive bursts as a function of the membrane potential prior to the hyperpolarizing pulse (Cueni et al., 2008). This yields a U-shaped curve reaching a peak at −65 and −60 mV in all sectors of WT mice (D1) that was abolished in CaV3.3 KO mice (D2). Dashed horizontal lines at ordinate value one indicate the border between repetitive and not-repetitive bursting conditions. (E) Mean number of repetitive bursts of TRN neurons (between −60 and −65 mV) across sectors and genotype. Mann-Whitney tests were used for comparison between genotypes, and p-values are given above the bars. (F) Histogram of the proportion of repetitive (colored rectangles) and non-repetitive bursting (grey rectangles with color surroundings) TRN neurons in the different sectors. Chi-square test followed by pairwise proportion test with Holm’s p-value adjustment was used for statistical evaluation, with significant value given above the bars.
-
Figure 2—source data 1
Numerical data values and statistics underlying Figure 2.
- https://doi.org/10.7554/eLife.39111.006
Figure 2—figure supplement 1
Oscillatory burst firing varies across the anteroposterior extent of TRN and depends on CaV3.3 Ca2+ channels.
(A) Epifluorescent micrograph of horizontal brain sections (A1, WT; A2), CaV3.3 KO mouse) showing the TRN’s anteroposterior extent (magenta, PV-staining, PV+) and two biocytin-filled neurons (red, biocytin+) recorded through in vitro whole-cell patch-clamp in anterior and posterior locations. Squares defined by dashed-lines on the left micrograph (5x) are shown expanded on the right (10x). (B) Left: Representative traces of oscillatory burst firing of TRN neurons in response to hyperpolarizing current injections (left, WT; right CaV3.3 KO) in the posterior/sensory sector of the TRN (blue) and in the anterior/limbic sector (red). (C) Inverted U-shaped voltage-dependence of burst firing for posterior (WT, n = 4; CaV3.3 KO, n = 6) and anterior (WT, n = 4; CaV3.3 KO, n = 6) cells. Neurons in the posterior location of WT cells showed more repetitive burst discharge between −65 and −60 mV than anterior neurons (posterior: 4.4 ± 0.8, anterior: 1.7 ± 0.3, Mann-Whitney test, p=0.042). CaV3.3 KO mice showed an abolition of repetitive bursting in both locations in comparison to WT (posterior: 1.0 ± 0.0, anterior: 0.3 ± 0.2, Mann-Whitney tests, p=5.7×10−4 for posterior, p=0.019 for anterior).
Figure 3 with 1 supplement
Cortical-area specific NREMS features depend on CaV3.3 Ca2+ channels.
(A) Schematic illustrating implantation sites for LFP, EEG (fEEG and pEEG; frontal and parietal EEG) and EMG electrodes. S1, S2; primary and secondary somatosensory areas; AC, auditory cortex; PFC, medial prefrontal cortex. (B) Histological sections of representative cases confirming the location of the recording sites. Arrowheads mark site of lesion caused by electrocoagulation. Anteroposterior stereotaxic coordinates are given relative to Bregma (Bg). (C) Representative raw traces of NREMS for WT (left) and CaV3.3 KO (right) animals, showing (from top to bottom) the EMG, EEG and LFP signals for S1, S2, AC and PFC (infra-/prelimbic area). The heart beat is visible on the EMG trace. (D) Power spectra corresponding to the LFP recordings, plotted in a linear-log plot to emphasize the three frequency bands of interest: the SO (0.5–1.5 Hz), the delta (1.5–4 Hz) and the sigma band (10–15 Hz). The sigma band is colored and shown in expanded log-log plots at the bottom. Normalized mean ±S.E.M. values of power spectral density are shown for S1, S2, AC and PFC for both genotypes (WT: n = 9 for S1, n = 8 for S2, n = 6 for AC, n = 6 for PFC; CaV3.3 KO: n = 13 for S1, n = 13 for S2, n = 8 for AC, n = 7 for PFC). (E) Mean total power for the three frequency bands across S1, S2, AC and PFC, with values for individual animals shown in points (dark gray for WT, light gray for CaV3.3 KO animals), and mean values ± S.E.M. in color diamonds. Statistical significance was tested for each area separately, comparing WT and CaV3.3 KO. Mann-Whitney non-parametric test for WT vs CaV3.3 KO, for S1, p=0.003 for the SO, p=0.007 for delta, p=0.021 for sigma; for S2, p=0.008 for the SO, p=0.005 for delta, p=6.9×10−5 for sigma; for AC, p=0.02 for the SO, p>0.05 for delta and sigma; for PFC, all p-values>0.05. Bonferroni-corrected α-threshold for the three frequency bands was 0.017.
-
Figure 3—source data 1
Numerical data values and statistics underlying Figure 3.
- https://doi.org/10.7554/eLife.39111.009
Figure 3—figure supplement 1
Head-fixed animals present spectra typical for each vigilance state accompanied by distinct LFP waveforms across cortical areas in WT and CaV3.3 KO animals.
(A) Power spectra of three main vigilance states (NREMS, REMS and wakefulness) for each local site of recording (LFP), as well as for the EEG, in WT and CaV3.3 KO head-fixed animals. EEG spectra is displayed both in linear-log and log-log scale to distinguish typical differences in low- and high-frequency components of the spectra. (B) Representative S1 traces during NREMS for a WT (left) and a CaV3.3 KO (right) animals, showing (from top to bottom) raw trace, same trace filtered for the SO (0.5–1.5 Hz), delta (1.5–4 Hz), and sigma (10–15 Hz) bands.
Figure 4 with 1 supplement
Chemogenetic hyperpolarization of TRN cells mimics the NREMS phenotype of the CaV3.3 KO mice.
(A) Representative membrane voltage recording of a whole-cell patch-clamped TRN cell in vitro exposed to CNO (10 μM, bath application indicated by horizontal bar) recorded in a slice from a mouse injected with AAV8-hM4D(Gi)_IRES_mCitrine. The cell was injected every 10 s with brief negative current pulses to elicit rebound discharge. The application of CNO hyperpolarized the membrane potential, suppressed rebound bursting and decreased membrane input resistance, as evident by the smaller voltage deflection in response to the negative current step (−80 pA). Current step size was increased to −120 pA to compensate for the decreased membrane resistance. Subsequent injection of direct current (d.c.) to counteract membrane hyperpolarization then reinstated burst discharge. Numbers indicate portions of the trace shown expanded below. Horizontal dotted lines indicate mean membrane potential before and during CNO application. (B) Box-and-whisker plot of membrane hyperpolarization in vitro (ΔV, calculated as the difference before and during CNO) for fluorescent (hM4+, n = 10, ΔV = −13.9 ± 1.5 mV, paired t-test, p=6×10−6) and non-fluorescent cells (hM4-, n = 4, ΔV = 0.0 ± 0.9 mV, paired t-test, p=0.97). The CNO-effects between the two-cell groups differed significantly (unpaired t-test, p=1.1×10−4). (C) Representative traces in vivo during NREMS 30 min after the injection of NaCl (left) or CNO (right) in the same animal, showing (from top to bottom) the EMG, EEG and S1-LFP (ipsilateral and contralateral to EEG) signals. (D) Mean ±S.E.M. power spectra of the S1-LFPs for NaCl and CNO injections in vivo during the NREMS periods 20 to 65 min after injection. Expanded portion is shown below in log-log scale to emphasize the sigma band (10–15 Hz). (E) Mean total power for the three frequency bands of interest. Diamonds and error bars show the Mean ±S.E.M. across subjects. Gray lines represent individual animals. Repeated-measures ANOVA for factors ‘frequency’ and ‘treatment’, p=7.7×10−5 was followed by paired t-tests for individual frequency bands, with values given above the bars. Bonferroni-corrected, α threshold was 0.017.
-
Figure 4—source data 1
Numerical data values and statistics underlying Figure 4.
- https://doi.org/10.7554/eLife.39111.012
Figure 4—figure supplement 1
Chemogenetic inhibition of TRN cells acutely increases delta activity in a DREADD-dependent expression.
(A) Epifluorescence of coronal brain sections showing bilateral DREADD-mCherry viral expression in the somatosensory TRN sector. (B) Left: mean ±S.E.M. (shaded curve) of S1 delta power dynamics in vivo of NREMS periods after the injection of NaCl or CNO (1 mg/kg) of a representative recording (obtained from same animal as in (A). This animal was injected with NaCl on 2, and with CNO on 3 days. Right: average across animals (n = 3). Shaded vertical areas represent the 20 min after injection (gray) and the 45 min window of maximal effect chosen for analysis (yellow). (C) Epifluorescence of coronal brain sections showing the bilateral expression of a control AAV8 virus in the somatosensory TRN sector as in A. (D) As B for animals injected with a control AAV8. Representative single recording from a mouse that was injected twice for each condition. (E) Mean ±S.E.M. power spectra of the S1-LFPs for NaCl and CNO injections in mice expressing control AAV8 virus during the 20 to 65 min NREMS periods after injection. Expanded portion is shown below in log-log scale to emphasize the sigma band (10–15 Hz). (F) Mean total power for the three frequency bands of interest. Diamonds and error bars show the Mean ±S.E.M across subjects. Gray lines represent individual animals. Statistics are done analogous to Figure 4.
Figure 5 with 1 supplement
Discrete spindle events are amplified and accelerated in somatosensory areas by CaV3.3 channels.
(A) Representative traces during NREMS for WT and CaV3.3 KO for S1, S2, AC and PFC, showing examples of algorithm-detected discrete spindle events (9–16 Hz, colored and bordered by arrowheads). (B) Mean amplitude of detected spindles quantified as mean power of the band-pass filtered signal (9–16 Hz). (B1) Mean spindle power levels across animals, values for individual animals shown by dots (dark gray for WT, light gray for CaV3.3 KO animals) and mean values ± S.E.M. by colored diamonds. Statistical significance was tested for each area separately using Mann-Whitney test, comparing WT (S1, n = 7; S2, n = 6; AC, n = 6; PFC, n = 6) and CaV3.3 KO (S1, n = 13; S2, n = 13; AC, n = 8; PFC, n = 7). p-values obtained were: for S1, p=0.002; for S2, p=0.006; for AC, p>0.05; for PFC, p=0.032. (B2) Cumulative probability distributions. (C) Same for intra-spindle frequencies. (C1) p-values obtained were: for S1, p=0.029; for S2, p=0.005; for AC and PFC, p>0.05. (C2) Probability distribution of spindle events according to their intra-spindle frequency.
-
Figure 5—source data 1
Numerical data values and statistics underlying Figure 5.
- https://doi.org/10.7554/eLife.39111.015
Figure 5—figure supplement 1
Illustration of procedure for automated spindle detection Representative traces for the four areas, S1, S2, AC and PFC in a WT animal, showing for each (from top to bottom): raw trace, band-pass filtered 9–16 Hz, power of the filtered trace.
Colored rectangles show the automatically detected spindle events. Horizontal detected line shows threshold for detection.
Figure 6
Phase-locking of spindles to the active state of the SO depends on CaV3.3 channels in somatosensory cortices.
(A) Illustration of the method to determine the phase value of the SO at which a detected spindle event starts (yellow shaded rectangles). SO periods were determined based on (Mölle et al., 2009) and indicated as dashed rectangle. Spindles that did not coincide with detected SO were not included in this analysis (gray shaded rectangles). (B) Graphs presenting sleep spindle occurrence as a function of SO phase. Colored lines represent individual animals, black lines the mean ±S.E.M. (gray-shaded curve). Same animal numbers as in Figure 5. (C) Mean occurrence of spindle onsets for the active (AS, −180°, 0°) and the silent (SS, 0°, 180°) state of the SO. Statistical significance was tested for each area separately with respect to spindle occurrence as a function of AS and SS. All tests were paired t-test, except for S2 for which we used Wilcoxon signed rank-test. For WT, p-values obtained were: for S1, p=2×10−5; for S2, p=0.031; for AC, p=3×10−4; for PFC, p=0.001. For CaV3.3 KO, p-values obtained were: for S1, p=0.543; for S2, p=0.033; for AC, p=0.001; for PFC, p=0.003. Comparison between genotypes was done for the active state (AS) using unpaired t-test, except for S2 for which we used Mann-Whitney test. p-values obtained were: for S1, p=8.3×10−6; for S2, p=5×10−4; for AC, p=0.06; for PFC, p=0.59.
-
Figure 6—source data 1
Numerical data values and statistics underlying Figure 6.
- https://doi.org/10.7554/eLife.39111.017
Author response image 1
(A) Effects of chemogenetic hyperpolarization of PV-expressing TRN cells, experiment and analysis as described in Figure 4.
(B) Effects of chemogenetic hyperpolarization of SOM-expressing TRN cells, experiment and analysis as described in Figure 4.
Author response image 2
(A) Graph dots represent the amplitude of each spindle event of the WT detected during periods of slow oscillation SO (amplitude is extracted from the power of the FIR filtered signal in the 9-16Hz band) as a function of SO phase, for the 4 sites of recording, S1, S2, AC and PFC.
Superimposed in grey is the normalized spindle occurrence (right y-axis) as a function of SO phase. Horizontal plain red line (and blue in the case of PFC) corresponds to mean amplitude of spindle events in the case of the CaV3.3-KO, dashed line is the mean +/- 1 standard deviation. The dataset presented in this figure is the same as in Figure 6 of the main paper. (B) Sub-sampling of spindle events in the WT constrained by the amplitude of the CaV3.3-KO for each area recorded: all spindle events from the WT outside the mean +/- 1SD from the CaV3.3-KO are discarded. Superimposed in dark blue is the spindle occurrence (right y-axis) as a function of SO phase for the sub-sampled spindle events. (C) Mean occurrence of spindle onsets for the active (AS, -180°, 0°) and the silent (SS, 0°, 180°) state of the SO, in grey is the same WT data as in Figure 6, in dark blue is the sub-sampled WT data constrained to the CaV3.3-KO mean +/- 1SD. D. Table presenting values of mean spindle amplitude and SD across CaV3.3-KO mice and taken to apply the subsampling of WT spindle events as in B and C. Number of spindle events for each condition (across all WT mice) is indicated in the legend on the right.
Tables
Key resources table
| Reagent type (species) or resource | Designation | Source or reference | Identifiers | Additional information |
|---|---|---|---|---|
| Genetic reagent (M. musculus) | CaV3.3 KO | PMID: 21808016 | MGI:5637591 | generated by Dr. H. Prosser, then at GSK |
| Genetic reagent (M. musculus) | VGAT-Ires-Cre | PMID: 21745644 | MGI:5141270 | generated by Dr. B. Lowell, Harvard |
| Recombinant DNA reagent | AAV1-hSyn- ChR2(H134R)_ eYFP-WPRE-hGH | Penn Vector Core | 26973P | |
| Recombinant DNA reagent | AAV8-hSyn-DIO- hM4D(Gi)_mCherry | UNC Vector Core | N/A | |
| Recombinant DNA reagent | ssAAV8/2-hSyn1-dlox- HA_hM4D(Gi)_IRES_ mCitrine-dlox-WPRE- hGHp(A) | Zurich viral vector repository | v93-8 | |
| Antibody | mouse anti-PV RRID:AB_10000343 | Swant | PV 235 | Dilution 1/4000 |
| Antibody | goat anti-mouse CY5 RRID:AB_2338713 | Jackson ImmunoResearch | 115-175-146 | Dilution 1/500 |
| Peptide, recombinant protein | streptavidin coupled with Alexa Fluor 594 RRID:AB_2337250 | Jackson ImmunoResearch | 016-580-084 | Dilution 1/8000 |
| Chemical compound, drug | CNO | Tocris | 6329 | |
| Software, algorithm | Neuroexplorer | Plexon | ||
| Software, algorithm | Intan RHD2000 recording system with Matlab toolbox 1.2.2 | IntanTeck | ||
| Software, algorithm | PClamp10.2 | Molecular Devices | ||
| Software, algorithm | Igor Pro 7 | WaveMetrics | ||
| Software, algorithm | Matlab 2018 a | MathWorks | ||
| Software, algorithm | R 3.5.1 | R Core Team |
Additional files
-
Transparent reporting form
- https://doi.org/10.7554/eLife.39111.018
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Thalamic reticular control of local sleep in mouse sensory cortex
eLife 7:e39111.
https://doi.org/10.7554/eLife.39111
