Developmental loss of ErbB4 in PV interneurons disrupts state-dependent cortical circuit dynamics

Abstract

GABAergic inhibition plays an important role in the establishment and maintenance of cortical circuits during development. Neuregulin 1 (Nrg1) and its interneuron-specific receptor ErbB4 are key elements of a signaling pathway critical for the maturation and proper synaptic connectivity of interneurons. Using conditional deletions of the ERBB4 gene in mice, we tested the role of this signaling pathway at two developmental timepoints in parvalbumin-expressing (PV) interneurons, the largest subpopulation of cortical GABAergic cells. Loss of ErbB4 in PV interneurons during embryonic, but not late postnatal development leads to alterations in the activity of excitatory and inhibitory cortical neurons, along with severe disruption of cortical temporal organization. These impairments emerge by the end of the second postnatal week, prior to the complete maturation of the PV interneurons themselves. Early loss of ErbB4 in PV interneurons also results in profound dysregulation of excitatory pyramidal neuron dendritic architecture and a redistribution of spine density at the apical dendritic tuft. In association with these deficits, excitatory cortical neurons exhibit normal tuning for sensory inputs, but a loss of state-dependent modulation of the gain of sensory responses. Together these data support a key role for early developmental Nrg1/ErbB4 signaling in PV interneurons as a powerful mechanism underlying the maturation of both the inhibitory and excitatory components of cortical circuits.

Fig. 1. Early, but not late, ErbB4 deletion from PV interneurons alters cortical firing and state transitions.

A Distribution of firing rates across the population of regular spiking (RS, upper panel) and fast-spiking (FS, lower panel) primary visual cortex (V1) neurons in controls (black) and Lhx6 mutants (cyan). B Average firing rate of RS cells during quiescence in controls and Lhx6 (cyan), PV (green), and SST (blue) mutants. Controls: 119 cells, 6 mice. Lhx6 mutants: 109 cells, 6 mice. PV mutants: 23 cells, 4 mice. SST mutants: 26 cells, 6 mice. C LFP power spectra during locomotion for controls and Lhx6, PV, and SST mutants. D Average firing rate for RS cells around locomotion onset (L-on) in controls (upper panel) and Lhx6 mutants (lower panel). E Firing rate modulation index (L–Q/L + Q) in early locomotion period (L; −0.5 to 0.5 s around L-on) as compared to quiescence period (Q) for RS cells in each group. Controls: 85 cells, 5 mice. Lhx6 mutants: 96 cells, 6 mice. PV mutants: 16 cells, 4 mice. SST mutants: 22 cells, 6 mice. F Average firing rate for RS cells during quiescence for P15 controls and Lhx6 mutants. Controls: 18 cells, 4 mice. Lhx6 mutants: 19 cells, 6 mice. G LFP power spectra during quiescence (left) and locomotion (right) for P15 controls and Lhx6 mutants. Shaded areas denote s.e.m.

Fig. 2. Early loss of ErbB4 in PV cells disrupts the temporal organization of neocortical activity.

A Spike-triggered LFP average in 1–6 Hz and 40–60 Hz bands during locomotion for controls (black) and Lhx6 mutants (cyan). B Average spike-LFP phase-locking during locomotion, measured as pairwise phase consistency (PPC). C Preferred LFP gamma-phase of firing during locomotion for RS (upper) and FS (lower) cells. D Consistency of preferred LFP gamma-phases for RS (upper) and FS (lower) cells. Controls: 130 RS cells, 16 FS cells, 6 mice. Lhx6 mutants: 117 RS cells, 31 FS cells, 6 mice. E Average normalized cross-correlograms during quiescence for RS-RS (upper) and FS-FS (lower) pairs. Controls: 192 RS-RS pairs, 15 FS-FS pairs, 6 mice. Lhx6 mutants: 166 RS-RS pairs, 19 FS-FS pairs, 6 mice. Shaded areas denote s.e.m.

Fig. 3. Early PV ErbB4 deletion impairs excitatory pyramidal neuron morphology.

A Example layer 5 pyramidal neuron reconstructions from an ErbB4^F/F control (left) and an ErbB4^F/F,Lhx6^Cre mutant (right). B Average total dendritic length for layer 5 pyramidal neurons in controls (black) and Lhx6 mutants (cyan). Controls: 15 cells. Lhx6 mutants: 13 cells. C Average mean dendritic intersections for control and Lhx6 mutant layer 5 cells. D Average maximum dendritic intersections for control and Lhx6 mutant layer 5 cells. E Average critical radius for control and Lhx6 mutant layer 5 cells. F Example spine density at the apical dendritic tuft of layer 5 pyramidal neurons in controls and Lhx6 mutants. G. Spine density on the main apical dendrite, basal dendrites, oblique dendrites, and apical dendritic tuft for layer 5 pyramidal neurons in controls and Lhx6 mutants. Controls: 10 cells. Lhx6 mutants: 10 cells.

Fig. 4. ErbB4 mutants exhibit reduced visual response modulation but intact selectivity.

A Upper panel: visual responses of individual RS cells in controls (left) and Lhx6 mutants (right) to the preferred stimulus of each cell. Lower panel: average visual response in each population, shown as signal:noise ratio (SNR) over the stimulus presentation period. B Average signal-to-noise ratio of visual responses for RS cells in controls (black) and Lhx6 mutants (cyan). Controls: 92 cells, 6 mice. Lhx6 mutants: 140 cells, 8 mice. C Average orientation selectivity index (OSI) of RS cells in controls and Lhx6 mutants. Controls: 83 cells, 6 mice. Lhx6 mutants: 108 cells, 6 mice. D Radial plots of preferred orientations of all RS cells in each group. E. Increase in stimulus rate modulation of RS cells during locomotion as compared to quiescence. Controls: 86 cells, 6 mice. Lhx6 mutants: 100 cells, 6 mice.