EHMT1 regulates Parvalbumin-positive interneuron development and GABAergic input in sensory cortical areas

Abstract

Mutations in the Euchromatic Histone Methyltransferase 1 (EHMT1) gene cause Kleefstra syndrome, a rare form of intellectual disability (ID) with strong autistic traits and sensory processing deficits. Proper development of inhibitory interneurons is crucial for sensory function. Here we report a timeline of Parvalbumin-positive (PV⁺) interneuron development in the three most important sensory cortical areas in the Ehmt1^+/- mouse. We find a hitherto unreported delay of PV⁺ neuron maturation early in sensory development, with layer- and region-specific variability later in development. The delayed PV⁺ maturation is also reflected in a delayed maturation of GABAergic transmission in Ehmt1^+/- auditory cortex, where we find a reduced GABA release probability specifically in putative PV⁺ synapses. Together with earlier reports of excitatory impairments in Ehmt1^+/- neurons, we propose a shift in excitatory-inhibitory balance towards overexcitability in Ehmt1^+/- sensory cortices as a consequence of early deficits in inhibitory maturation.

Keywords: Autism spectrum disorder; Cortical development; Critical period; EHMT1; Gabaergic synapse; Kleefstra syndrome; Parvalbumin-positive interneurons.

Fig. 1

Ehmt1 haploinsufficiency delays PV and PNN expression in the auditory cortex during the auditory critical period. a Primary auditory cortex, layers 2/3 and 4 are marked in red. b Overview of PV⁺ (green) and PNN⁺ (red) development in the auditory cortex over time. Layer numbers are marked on the left. c Neurons stained for Parvalbumin (green) and PNNs (red) in the primary auditory cortex, layer 2/3. d Quantification of the density of PV⁺, PNN⁺, and PV/PNN double-labelled cells. N P14 = 63/38, P28 = 40/39, P56 = 22/24 images, from 1 to 4 slices per animal, P14 = 6/4, P28 = 3/3, P56 = 3/3 Ehmt1^+/+/ Ehmt1^+/− animals. e Neurons stained for Parvalbumin (green) and PNNs (red) in the primary auditory cortex, layer 4. f Quantification of the density of PV⁺, PNN⁺, and PV/PNN double-labelled cells. N P14 = 66/38, P28 = 40/37, P56 = 13/13 images, from 1 to 4 slices per animal, P14 = 6/4, P28 = 3/3, P56 = 3/3 Ehmt1^+/+/ Ehmt1^+/− animals. Source for a Allen Adult Mouse Brain Atlas, https://atlas.brain-map.org/atlas?atlas=1. Image Credit: Allen Institute. Scale bars in b 500 µm, c, e 20 µm. Note that the intensities in c, e are matched within a timepoint, but not across timepoints. Scale in d, f cells/mm², Dot plot with mean ± SEM, color-coded per mouse. *p < 0.05, **p < 0.01; Nested ANOVA with mice as subgroups. See Supplementary Table 1 for exact values

Fig. 2

A persistent reduction in PV⁺ neuron markers in Ehmt1^+/− somatosensory cortex layer 4, but not in layer 2/3. a Primary somatosensory cortex, Barrel field, layers 2/3 and 4 are marked in red. b Overview of PV⁺ (green) and PNN⁺ (red) development in the somatosensory cortex over time. Layer numbers are marked on the left. c Neurons stained for Parvalbumin (green) and PNNs (red) in the primary somatosensory cortex, layer 2/3. d Quantification of the density of PV⁺, PNN⁺, and PV/PNN double-labelled cells. N P14 = 30/28, P28 = 54/53, P56 = 63/63 images, four slices each from P14 = 3/3, P28 = 3/3, P56 = 6/6 Ehmt1^+/+/ Ehmt1^+/− animals. Color-code per mouse. e: Neurons stained for Parvalbumin (green) and PNNs (red) in the primary somatosensory cortex, layer 4. f Quantification of the density of PV⁺, PNN⁺, and PV/PNN double-labelled cells. N P14 = 29/29, P28 = 53/52, P56 = 68/63 images, four slices each from P14 = 3/3, P28 = 3/3, P56 = 6/6 Ehmt1^+/+/ Ehmt1^+/− animals. Color-code per mouse. Source for a Allen Adult Mouse Brain Atlas, https://atlas.brain-map.org/atlas?atlas=1. Image Credit: Allen Institute. Scale bars in b 500 µm, c, e 20 µm. Note that the intensities in c, e are matched within a timepoint, but not across timepoints. Scale in d, f cells/mm², Dot plot with mean ± SEM, color-coded per mouse. *p < 0.05, **p < 0.01; nested ANOVA with mice as subgroups. See Supplementary Table 1 for exact values

Fig. 3

An early delay in PV + maturation in Ehmt1^+/− visual cortex layer 4, but not layer 2/3. a Primary visual cortex, layers 2/3 and 4 are marked in red. b Overview of PV⁺ (green) and PNN⁺ (red) development in the visual cortex over time. Layer numbers are marked on the left. c Neurons stained for Parvalbumin (green) and PNNs (red) in the primary Visual cortex, layer 2/3. d Quantification of the density of PV⁺, PNN⁺, and PV/PNN double-labelled cells. N P14 = 30/28, P28 = 32/35, P56 = 30/32 images, 12/12 slices from 3/3 Ehmt1^+/+/ Ehmt1^+/− animals per age. e Neurons stained for Parvalbumin (green) and PNNs (red) in the primary Visual cortex, layer 4. e, f Quantification of the density of PV⁺, PNN⁺, and PV/PNN double-labelled cells. N P14 = 30/29, P28 = 39/35, P56 = 31/34 images, 12/12 slices from 3/3 Ehmt1^+/+/ Ehmt1^+/− animals per age. Source for a Allen Adult Mouse Brain Atlas, https://atlas.brain-map.org/atlas?atlas=1. Image Credit: Allen Institute. Scale bars in b 500 µm, c, e 20 µm. Note that the intensities in c, e are matched within a timepoint, but not across timepoints. Scale in d, f cells/mm², Dot plot with mean ± SEM, color-coded per mouse. *p < 0.05, **p < 0.01; Nested ANOVA with mice as subgroups. Please see Supplementary Table 1 for exact values

Fig. 4

Inhibitory GABAergic transmission is impaired in Ehmt1^+/− auditory cortex at P14-16. a Representative traces of inhibitory miniature postsynaptic currents (mIPSCs) recorded from pyramidal neurons in Ehmt1^+/+ (black) and Ehmt1^+/− (red). Scale bar = 20 pA, 100 ms. b Cumulative probability plots of mIPSC Frequency (left), and mIPSC amplitude (right). N = 9/8 cells. c Left, Putative inhibitory presynapses (arrowheads) stained with Gad67 (green) next to neuronal somata (NeuN, red). Right, quantification of perisomatic Gad67⁺ puncta. N = 319/237 cells in 12/12 slices from 3/3 Ehmt1^+/+/ Ehmt1^+/− mice. Student’s t test. d Composite traces illustrating the paired-pulse paradigm. GABAergic inputs to A1 L2/3 pyramidal cells are stimulated at intervals of respectively 50, 100, 150, 200, or 500 ms. Scale bar = 100 ms. e Paired-pulse ratio for Ehmt1^+/+ (black) and Ehmt1^+/− (red). N = 14/14 cells from 4/4 Ehmt1^+/+/ Ehmt1^+/− animals. f Paired-pulse ratio in the presence of ω-Agatoxin-IVA, a specific blocker of P/Q type Ca²⁺ channels. g Comparison of Ehmt1^+/+ PPR at physiological (2 mM, black) and elevated (4 mM, grey) extracellular Ca²⁺. h Comparison of Ehmt1^+/− PPR at physiological (2 mM, red) and elevated (4 mM, salmon) extracellular Ca²⁺. i Run-down of vesicle pools by 10 Hz stimulation at 4 mM extracellular Ca²⁺. Normalized amplitudes plotted against time; 100 stimuli with 100 ms ISI, followed by control pulses at 500 ms and 30 s (sweep end). Multiple t tests with Benjamini–Hochberg correction. j Readily releasable pool determined from the 10 Hz stimulus train. RRP displayed normalized to the initial stimulus amplitude. k Release Probability determined from the 10 Hz stimulus train. *p < 0.05, **p < 0.01. Please see Supplementary Table 2 for exact values