Axonal dynamics of excitatory and inhibitory neurons in somatosensory cortex

Abstract

Cortical topography can be remapped as a consequence of sensory deprivation, suggesting that cortical circuits are continually modified by experience. To see the effect of altered sensory experience on specific components of cortical circuits, we imaged neurons, labeled with a genetically modified adeno-associated virus, in the intact mouse somatosensory cortex before and after whisker plucking. Following whisker plucking we observed massive and rapid reorganization of the axons of both excitatory and inhibitory neurons, accompanied by a transient increase in bouton density. For horizontally projecting axons of excitatory neurons there was a net increase in axonal projections from the non-deprived whisker barrel columns into the deprived barrel columns. The axon collaterals of inhibitory neurons located in the deprived whisker barrel columns retracted in the vicinity of their somata and sprouted long-range projections beyond their normal reach towards the non-deprived whisker barrel columns. These results suggest that alterations in the balance of excitation and inhibition in deprived and non-deprived barrel columns underlie the topographic remapping associated with sensory deprivation.

Figure 1. Methods for longitudinal imaging of cortical circuitry.

(A) Timeline of experimental procedures. The barrel cortex is mapped electrophysiologically in order to guide placement of injections of genetically modified virus. After allowing several weeks for expression of genes carried by the virus, images of fluorescent label are taken to establish baseline connectivity. The cortex is reimaged at various time points following initiation of whisker plucking. (B) Surface view of horizontally projecting axons, imaged with 2-photon microscopy. (C) Cortical representation of mouse whiskers, with barrels arranged in rows (A to E) and arcs (1 to 6). (D) Whisker plucking induced remapping of barrel cortex topography in adult animals. The movement of the representation of C row whiskers into the deprived cortex following plucking of D and E rows is plotted as a function of the number of days of plucking. The original C/D row border corresponds to zero on the y-axis. The horizontal line at 300 micrometers indicates the original boundary between the D and E rows. (E) A reconstruction of the axonal plexus of excitatory neurons labeled by AAV-eYFP injection into C3 barrel column, shown in surface view. The axonal reconstruction is superimposed on the whisker map. The virus injection site was at the center of the C3 barrel column.

Figure 2. Axons and synaptic boutons during normal experience.

(A) Axons at initial imaging session (top) and 1 wk later (bottom). (B) Axonal arbors imaged at two time points. Left, map of the whisker representation superimposed with the reconstruction of axons projecting toward rows D and E. Middle, reconstruction of the long-range horizontal axons imaged the first session. And right, 1 wk later. (C) Distribution of axonal density surrounding the injection site, averaged over all the control animals. Bins are spaced at 100 µm intervals with the injection site at (0,0,0). Axons immediately surrounding injection are not shown. Top, first imaging session. Bottom, second imaging session.

Figure 3. Bouton dynamics of selected axons in superficial cortical layers of whisker barrel columns of deprived and non-deprived animals.

Graphs depict the number of boutons on axons before and after whisker plucking. Each bar represents a single identified axon, imaged before (left set) and the same axons imaged after whisker plucking (right set). Blue: stable boutons, red: eliminated boutons, green: added boutons. (A–C) Excitatory neurons. (A) Bouton turnover in whisker barrel columns in animals that underwent normal experience. (B) Bouton turnover during whisker plucking for axons projecting into non-deprived barrel columns and (C) projecting into deprived barrel columns (D–F) Inhibitory interneurons. (D) Bouton turnover in whisker barrel columns of animals that underwent normal experience. (E) Axonal boutons of neurons located in non-deprived barrel columns before and after whisker plucking. (F) Boutons of neurons located in deprived barrel columns.

Figure 4. Axons of layer 2/3 excitatory neurons originating from C3 barrel column projecting to deprived barrel columns following whisker plucking.

(A) Left: Montage of z-projections located within a deprived barrel column before (far left) and after (second from left) 14 d of whisker plucking. Right, Montage of z-projections of a cluster of axons located within deprived cortex both before (3^rd from left) and after (far right) 1 mo of whisker plucking. * - denotes putative growth cone. (B) Axons originating from C3 injection site projecting to rows D and E following whisker plucking. Left, the map of the barrel column field. The square box depicts the region shown in the two reconstructions to the right. Middle, axons located within D and E rows before plucking. Right, same area after 1 mo of plucking. Axons that were stable over the two imaging sessions are shown in blue; axons that retracted from the first to second imaging session are in red; new axons are plotted in yellow.

Figure 5. Changes in axonal density following varying durations of whisker plucking.

(A–D) For 2 (A), 14 (B), 30 (C), and 60 (D) d. Left column, first imaging session. Right column, second imaging session. The axonal length in each bin was averaged over multiple animals and then normalized. The normalization was performed by dividing all the bins in each baseline condition by the value of the highest density bin maximum in that condition. The bins in the corresponding post-plucking time point were divided by the same number. The maximum mean axonal length for any bin in each baseline graph is therefore 1.0 (expressed in arbitrary units, A.U.), and the length in the corresponding post-plucking graph is expressed relative to that standard. The bins in the histogram are 50 µm×50 µm; the injection site is located at coordinate (0,0,0). Average location of barrel column for animals in each condition is indicated on each graph. White text depicts deprived barrel columns. Green text depicts non-deprived barrel columns.

Figure 6. Excitatory axons originating from neurons within non-deprived barrel columns.

(A) Excitatory axons located within non-deprived barrel columns. Left, axons located in rows A and B before whisker plucking. Right, same area following 1 mo of whisker plucking. (B–D) Axonal density in non-deprived cortex after 14 (B), 30 (C), and 60 (D) d of whisker plucking. Left column, first imaging session. Right column, second imaging session after the indicated period of plucking. As in Figure 5, the axonal density at each pair of time points was averaged over a set of mice and normalized by dividing each bin by the maximum region of local density in the baseline condition. Normalized mean axonal density is plotted (in arbitrary units) in 50 µm×50 µm bins with the respect to the injection site (asterisk) at coordinate (0,0,0). Distances are expressed in µm from injection site. Average location of barrel column for animals in each condition is indicated on each graph. Green labels refer to non-deprived barrel columns.

Figure 7. Cell type specificity of expression in GAD65.eGFP.AAV infected neurons.

(A) Combined eGFP expression and labeling with antibodies against calcium-binding proteins. Left, neurons labeled with GAD65.eGFP.AAV. Middle, neurons labeled with a cocktail of antibodies against the calcium-binding proteins calbindin, calretinin, and parvalbumin. Right, merged images, double-labeled neurons are yellow. (B) Number of neurons labeled by GAD65.eGFP.AAV that were co-localized with calcium binding proteins. (C) Axons of inhibitory neurons in normal cortex. Left, initial imaging session. Right, same region imaged 5 d later.

Figure 8. Axons of inhibitory interneurons located in the non-deprived rows following whisker plucking.

(A) Top, the whisker barrel map. Rectangular box depicts the region of axonal reconstructions shown on the right. Middle, axons located within non-deprived rows before whisker plucking. Right, same area after 2 d of plucking. In the reconstruction, the axons that persisted over both sessions are shown in blue, axons retracted from the first to second imaging session in red, and new axons in yellow. Scale bar  = 50 µm. (B–D) Changes in axonal length for 2 (B), 14 (C), and 30 (D) d of whisker plucking. Left, the distribution of axonal length that was lost between each baseline and post-plucking time point. Right, the distribution of axonal length that was gained between the baseline and post-plucking time points. The data for each pair of time points were obtained by averaging over several mice. Here, the magnitude of the axonal changes for each data pair is normalized with respect to the maximum length of axon that was retracted within any bin. The maximum length of retracted axon in each data pair is therefore 1.0 (in arbitrary units of length), and the length of added axon is measured with respect to that value. The dimensions of the bins are 50 µm×50 µm. The average locations of barrel columns for animals in each condition are marked on each map, with deprived barrel columns indicated in white and non-deprived barrel columns indicated in green.

Figure 9. Axons of inhibitory interneurons located within deprived rows following 2 d of whisker plucking.

(A) Left, axons located within deprived rows before plucking. Right, same area after 2 d of plucking. Axons that are stable over the two imaging sessions are color coded blue; those pruned after the first session colored red; and newly sprouted axons shown in yellow. (B–D) Quantification of changes in axonal length for 2 (B), 14 (C), and 30 (D) d of whisker plucking. As in Figure 8, the data are normalized relative to the highest density of retracted axons for each pair of time points. Retracted (left) and added (right) axons are plotted in 50 micron bins, with the injection site, shown by the asterisk, at coordinate (0,0,0). Distances are expressed in µm from injection site. The average locations of barrel columns for animals in each condition are marked on each graph. Deprived barrel column labels are in white, non-deprived barrel labels are in green.

Figure 10. Short term changes in axonal length and bouton number.

Over hours following the onset of plucking for inhibitory interneurons whose soma are located in deprived rows. (A) Location of boutons with respect to cortical topography. (B) Reconstruction of axons following whisker plucking. (C) Time course of change in axonal length (%). (D) Time course of change in bouton number (%).

Figure 11. Change in axonal density plotted as a function of days of deprivation.

Expressed as a ratio of axonal length at the indicated duration of deprivation to that measured before deprivation. The green line depicts axons of inhibitory interneurons whose somata are located in deprived barrel columns and are projecting into the non-deprived barrel columns. The yellow line depicts axons of excitatory neurons in the non-deprived barrel columns that are projecting into the deprived barrel columns.