Cortical representations of olfactory input by trans-synaptic tracing

Abstract

In the mouse, each class of olfactory receptor neurons expressing a given odorant receptor has convergent axonal projections to two specific glomeruli in the olfactory bulb, thereby creating an odour map. However, it is unclear how this map is represented in the olfactory cortex. Here we combine rabies-virus-dependent retrograde mono-trans-synaptic labelling with genetics to control the location, number and type of 'starter' cortical neurons, from which we trace their presynaptic neurons. We find that individual cortical neurons receive input from multiple mitral cells representing broadly distributed glomeruli. Different cortical areas represent the olfactory bulb input differently. For example, the cortical amygdala preferentially receives dorsal olfactory bulb input, whereas the piriform cortex samples the whole olfactory bulb without obvious bias. These differences probably reflect different functions of these cortical areas in mediating innate odour preference or associative memory. The trans-synaptic labelling method described here should be widely applicable to mapping connections throughout the mouse nervous system.

Figure 1. Genetic control of rabies-mediated neural circuit tracing

a-b, Schematic representation of the methodology used to control the location, number and type of starter cells for RV-mediated transsynaptic labeling. tTA2 is expressed in a small subset of CreER(+) cells (grey nuclei in b). tTA2 activates an AAV-delivered transgene to express: 1) a histone-GFP marker to label the nuclei of starter cells in green, 2) EnvA receptor (TVA) to enable subsequent infection by EnvA-pseudotyped RV (rabies ΔG-mCherry+EnvA), and 3) rabies glycoprotein (G) to initiate transsynaptic labeling. c, Top left, a 60-µm coronal section that includes the injection site in the motor cortex (M1). Cells labeled with both histone-GFP (n-GFP) and mCherry (arrowheads in c₁, magnified in c₃) can be distinguished from cells labeled with mCherry alone, which are found near the injection site (c₁), in the contralateral motor cortex (c₂), in the somatosensory barrel cortex (top right; magnified in c₄), and in the motor-specific ventrolateral nucleus of the thalamus (c₅). Scale bars, 1 mm for low-magnification images on top, 100 µm for high-magnification images at the bottom.

Figure 2. The OB to AON connections display a dorsal-ventral topography

a, A 60 µm coronal section with two starter cells located in layer II of the ventrolateral AON, one of which (arrow) is magnified in the inset. RMS, rostral migratory stream. Scale, 500 µm. b, Typical examples of transsynaptically labeled mitral cells from cortical starter cells. Left, a 60 µm coronal section that captures both the cell body and the apical dendrite of a mitral cell. Right, more frequently, a mitral cell apical dendrite spans several consecutive 60 µm coronal sections. S-Glo/D-Glo, glomerulus innervated by a single or two (Ma, Mb) labeled mitral cells. A, anterior; P, posterior. Scale bar, 100 µm. c, Superimposed 3D-reconstructions of the AONs and the OBs from two injected brains. 11 red and 4 green starter cells from two AONs labeled red and green glomeruli, respectively. Light red and green, contours of two superimposed AONs. d-e, Correlations between θ_AON and θ_OB (d) and θ_AON and θ’_OB (e). Crosses represent mean θ_AON (x-axis) and mean θ_OB or θ’_OB (y-axis). Bars represent 50% of the distribution surrounding the mean θ_OB or θ’_OB. R, Pearson’s correlation coefficient; p, statistical significance tested against the null hypothesis assuming no correlation between θ_AON and θ_OB or θ’_OB. Red and blue, experiments using actin-CreER and CaMKII-CreER, respectively.

Figure 3. Representations of OB input in the amygdala and piriform cortex

a-b, Starter cells from the cortical amygdala and piriform cortex. Left, single coronal sections at the injection sites in the posterolateral cortical amygdala (a) and the posterior piriform cortex (b). Arrows point to starter cells magnified in insets. Scale bars, 100 µm in a, 200 µm in b. Right, schematic representations of 10 independent injections each into amygdala (a) or piriform cortex (b). Starter cells from each injection are labeled with a specific color. The dotted line denotes the rough border between ACo (anterior cortical amygdala) and PLCo (posterolateral cortical amygdala) based on anatomical landmarks according to a mouse brain atlas. nLOT, nucleus of lateral olfactory tract; ME, medial amygdala; PMCo, posteromedial cortical amygdala; APC, anterior piriform cortex; PPC, posterior piriform cortex; En, lateral entorhinal cortex. c-d, Superposition of three independent 3D-reconstructions of glomerular maps with starter cells from the cortical amygdala (c) or the piriform cortex (d). e, Mean θ’_OB values (crosses) from each experiment are plotted in the same column with the 95% confidence intervals for corresponding ^simθ’_OB values (gray bars). Samples with experimental mean θ’_OB outside the 95% confidence intervals are labeled with asterisks (*, p<0.05). Colors in a (scheme), c and e (AMY) are matched to represent the same samples, and so are the colors in b (scheme), d and e (PC).

Figure 4. Convergence and independence of mitral cell inputs

a, Convergence index for each cortical injection experiment is represented by a diamond, with the type and layer of starter cells specified by the color code above. b, Multiple regression analysis to estimate the convergence indices of starter cells located in different layers of the AON and PC. Estimated mean convergence indices (“x”) and the corresponding 95% confidence intervals (grey bars) are shown. Data from actin-CreER and GAD2-CreER were analyzed separately. Injections into amygdala produced only one sample that contained layer I cells and were therefore excluded. c, Schematic of dually labeled glomeruli (D) resulting from two starter cells (Dt) or a single starter cell (Ds). Comparison of the distributions of Ds derived from experimentally observed frequency of D (Data Ds, red) and from simulated D based on the null hypothesis detailed in the Methods (Random Ds, blue). For each sample, the distributions of “Data Ds” and “Random Ds” are shown by colored heat maps. *, p<0.05.