Activity-induced remodeling of olfactory bulb microcircuits revealed by monosynaptic tracing

Abstract

The continued addition of new neurons to mature olfactory circuits represents a remarkable mode of cellular and structural brain plasticity. However, the anatomical configuration of newly established circuits, the types and numbers of neurons that form new synaptic connections, and the effect of sensory experience on synaptic connectivity in the olfactory bulb remain poorly understood. Using in vivo electroporation and monosynaptic tracing, we show that postnatal-born granule cells form synaptic connections with centrifugal inputs and mitral/tufted cells in the mouse olfactory bulb. In addition, newly born granule cells receive extensive input from local inhibitory short axon cells, a poorly understood cell population. The connectivity of short axon cells shows clustered organization, and their synaptic input onto newborn granule cells dramatically and selectively expands with odor stimulation. Our findings suggest that sensory experience promotes the synaptic integration of new neurons into cell type-specific olfactory circuits.

Figure 1. Engineered Rabies Virus Allows Monosynaptic Circuit Tracing in Olfactory Bulb.

(a) Illustration of the ROSA26-stop^flox-tdTomato conditional knock-in allele. (b) Illustration of the tri-cistronic G-IRES-TVA-IRES-Cre expression construct used for electroporation into the subventricular zone (SVZ) of ROSA26-stop^flox-tdTomato mice. (c) Illustration of the in vivo electroporation procedure. Left, plasmid DNA encoding G-IRES-TVA-IRES-Cre was injected into the lateral ventricle of newborn mice. Middle, a square-pulse voltage was applied across the head to introduce the expression construct into SVZ progenitors. Right, 30 d following electroporation, pseudotyped SADΔG-EGFP RV was injected into the olfactory bulb for targeted infection of electroporated granule cells. (d) Diagram showing how pre- and postsynaptic neurons can be identified by two-color monosynaptic viral tracing in the conditional ROSA26-stop^flox-tdTomato background. Red cells represent conditional tdTomato reporter expression following G-IRES-TVA-IRES-Cre electroporation. Yellow cells represent electroporated cells that are also infected by SADΔG-EGFP RV. Green cells represent presynaptic targets of the infected granule cells. GL, glomerular layer; EPL, external plexiform layer; ML, mitral cell layer; IPL, internal plexiform layer; GCL, granule cell layer. (e) Schematic illustrating the estimated domain of SADΔG-EGFP RV infection with respect to the entire bulb. On average, infected granule cells were detected within 302 µm of the injection site (green circle)±214 µm (yellow outer circle). Scale bar, 300 µm. (f) Whole mount view of ROSA26-stop^flox-tdTomato mouse brain 30 d after unilateral G-IRES-TVA-IRES-Cre electroporation. Dashed line represents the coronal section imaged in (g)–(i). OB, olfactory bulb. Scale bar, 1 mm. (g) Coronal section through an electroporated and infected olfactory bulb showing tdTomato and SADΔG-EGFP expression. Scale bar, 300 µm. (h) Coronal section through the olfactory bulb of electroporated mice showing tdTomato-expressing granule cells at higher magnification. (i) Section shown in (h) imaged for SADΔG-EGFP following RV infection. (j) Merged view of (h) and (i). MC, mitral cell; GC, granule cell origin; dashed arrow, granule cell dendrite. Scale bar, 25 µm. (k)–(m) Dual tdTomato plus SADΔG-EGFP reporter expression, identifying a local granule cell microcircuit. (k) tdTomato expression in a single granule cell. (l) SADΔG-EGFP expression in the same ‘source’ granule cell shown in (k) and local presynaptic partners. (m) Merged reporter expression delineating source cell (yellow) and the local presynaptic partners (green). Short arrows in (k–m) point to the source granule cell. Scale bar, 10 µm.

Figure 2. SADΔG-EGFP RV is Retrogradely Transported from Granule Cells to Presynaptic Targets in Olfactory Bulb.

(a) A coronal section of olfactory bulb following monosynaptic viral tracing showing a clustered pattern of presynaptic labeling. The dashed line shows the midline through a coronal section of MOB. Scale bar, 100 µm. (b) SADΔG-EGFP expression in presynaptic mitral cells. Scale bar, 25 µm. (c)–(d) SADΔG-EGFP expression in short axon cells in the external plexiform and granule cell layers (hatched yellow boxes). Scale bars, 20 µm. (e)–(g) Partial colabeling of calretinin in SADΔG-EGFP expressing cells. Scale bar, 20 µm. (h)–(j) Partial colabeling of parvalbumin in SADΔG-EGFP expressing cells. Scale bar, 20 µm. (k)–(m) Overlapping expression of GABA_A R α1 in SADΔG-EGFP labeled short axon cells. Scale bar, 15 µm. (n)–(p) SADΔG-EGFP expressing cells do not express tyrosine hydroxylase, and thus are not periglomerular. Dashed arrows point to a tyrosine hydroxylase positive cell not labeled by SADΔG-EGFP RV. Scale bar, 20 µm. (q)–(s) Expression of GFAP in occasional SADΔG-EGFP labeled glial cells. Scale bar, 10 µm. Arrows point to overlapping marker expression in SADΔG-EGFP labeled cells. For all panels, GL, glomerular layer; EPL, external plexiform layer; ML, mitral cell layer; GCL, granule cell layer.

Figure 3. Newborn Granule Cells Receive Extensive Input from Short Axon Cells.

(a)–(c) An SADΔG-EGFP labeled olfactory bulb microcircuit in which pre- and postsynaptic cell types can be identified by differential reporter expression. Electroporated cells appear red due to Cre activation of tdTomato expression. SADΔG-EGFP infected source cells appear yellow due to co-expression of tdTomato and EGFP. Presynaptic partners become trans-synaptically infected with SADΔG-EGFP but lack Cre and thus appear green. The dashed arrow points to a mitral cell, MC. Arrows point to short axon cells, SACs. Arrowheads indicate a source granule cell, GC. For (a)–(d): EPL, external plexiform layer; ML, mitral cell layer; GCL, granule cell layer. (a) tdTomato expression (red) in recombined Cre-expressing granule cells. (b) SADΔG-EGFP expression in a granule source cell (arrowhead) and its presynaptic targets (arrows). (c) Merge of (a) and (b). Scale bar, 20 µm. (d) Examples of volume-rendered reconstructions showing local short axon cell microcircuits with synaptic contacts onto newborn granule cells. The postsynaptic granule cells are shown in red, and the presynaptic short axon cells are shown in green. Scale bar, 15 µm. (e)–(f) Examples of action potential responses to depolarizing current injection and images of the short axon cell types observed to make synaptic contacts onto newborn granule cells. Shown are firing responses (left) and cellular morphologies (right) of a representative deep short axon cell (e) and superficial short axon cell (f) with contacts onto a newborn granule cell. Scale bars, 15 and 10 µm, respectively.

Figure 4. Odor Enrichment Increases SAC Connectivity onto Granule Cells.

(a) Schematic of experimental paradigm to track changes in granule cell connectivity following odor enrichment. (b) Dual labeled region of a control olfactory bulb showing SADΔG-EGFP expression in presynaptic targets. Scale bar, 10 µm. (c) Dual labeled olfactory bulb from a mouse subjected to odor enrichment showing increased presynaptic labeling (green). Arrows points to source granule cells (GC) which appear yellow due to co-expression of tdTomato and EGFP. Presynaptic partners appear green. Scale bar, 10 µm. EPL, external plexiform layer; ML, mitral cell layer; GCL, granule cell layer. (d) Connectivity ratio between postsynaptic granule cells and presynaptic short axon cells (SAC∶GC) under control conditions or following odor enrichment. ^*p<0.01, Student's t-test.

Figure 5. Odor Stimulation Increases Synaptic Inputs onto Newborn Granule Cells.

(a) Dendrites from granule cells in control or odor exposed mice showing the increased number of spines following odor enrichment. Yellow arrows point to individual spines. Scale bar, 3 µm. (b) Data represent means ± SEM of spine number per 25 µm dendrite on granule cells in mice exposed to cycled odorants (odor) compared to non-odor exposed controls. ^*p<0.001, Student's t-test. (c) Gephyrin labeling to reveal inhibitory GABAergic synapses on control granule cell dendrites. Insets in (c) and (d) show tdTomato expression in doubly labeled dendrites. Scale bars, 2 µm. (d) Increased number of gephyrin labeled inhibitory synapses contacting dendrites of granule cells from mice subjected to odor enrichment.

Figure 6. Odor Enrichment Increases Inhibitory Drive Onto Newborn Granule Cells.

(a)–(b) Quantitative analysis of (a) average frequency and (b) amplitude of mIPSCs recorded from labeled granule cells in control and odor-enriched mice. Odor enrichment increased the frequency, but not amplitude, of granule cell mIPSCs (control, n = 9; odor, n = 10, ^*p<0.01, unpaired t-test). (c) Representative voltage-clamp recordings of mIPSCs from granule cells in acute MOB slices from control and odor-enriched mice.