Skip to main content
. 2020 Sep 17;9:e55585. doi: 10.7554/eLife.55585

Figure 1. Tracing strategy and localizations for input and output viral tracings from distinct IC subregions.

Schematic representation of Cre-dependent (A) monosynaptic retrograde Rabies virus tracing (RV) and (B) anterograde axonal AAV tracings (AAV) used to determine respective input and output connectivity to the IC. Tracings were performed in both excitatory (Camk2a-Cre) and inhibitory (Gad2-Cre) mouse lines for RV, and only in the Camk2a-Cre mouse line for AAV. For RV tracings, AAV-FLEX helper viruses expressing mCherry-tagged TVA (1) and rabies-virus-specific G protein (2) were co-injected into the IC region of interest. Three weeks later EnvA-coated, eGFP-expressing modified RV lacking G protein was injected at the same location (3). For anterograde tracings, a one-off injection of eYFP-expressing AAV-FLEX virus was administered into the chosen location. Three distinct IC subregions were chosen for each tracing technique: anterior (aIC, red), medial (mIC, green) or posterior (pIC, blue). (C) Schematic illustration of the lateral view of the IC including distances from Bregma (top panel) and heatmap showing average starter cell distribution for each tracing strategy at each specific IC target (bottom panels). The three IC target subregions were mostly non-overlapping, and only a minimal percentage of cells were detected in the Motor and Sensory Cortex (M/S), or Piriform Cortex (Pir) neighboring the IC. n = 3 mice per injection site/tracing strategy. Heatmap intensity scale is the same for all three IC target subregions. Regions absent at specific Bregma levels indicated by dark gray squares.

Figure 1.

Figure 1—figure supplement 1. Starter cell identification.

Figure 1—figure supplement 1.

(A) Starter cell identification pipeline for RV helper system. (1) High-resolution image of a representative section at the injection site in the IC. Starter cells are double-labeled with TVA-mCherry and RV-GFP, and appear yellow. Scale bars 200 µm (main image), 50 µm (inset). Number of starter cells were identified in an automated fashion using Cell Profiler. First RV+ cells were identified ([2], yellow cell outlines) from the GFP image, then RV+ cells that also contain mCherry-TVA were identified from the mCherry signal ([3], red rings within yellow RV+ cell outlines). Double-labeled cells were counted as starter cells. (B) Starter cell identification pipeline for AAV tracing system. (1) Representative epifluorescent image of YFP-labeled AAV starter cells. Scale bars 200 µm (main image), 50 µm (inset). (2) YFP-positive cells were identified in an automated manner using Cell Profiler. Data given as cells per brain subregion, which was manually defined before cell identification. (C) Raw data for each individual animal used. Starter cell range values given as distance from Bregma in mm. Ratio shown as total cells/starter cells. Hemisphere indicates injection site.
Figure 1—figure supplement 2. Starter cells.

Figure 1—figure supplement 2.

(A) Starter cell exclusion threshold. High-resolution image of a representative section at the injection site in the IC. Starter cells are double-labeled with TVA-mCherry and RV-GFP, and appear yellow. The center of the population is marked with a circle. Scale bars 200 µm. (1) Example of a brain without spillover into adjacent regions (2) Example of a brain with an acceptable amount of spillover into Primary Sensory Cortex (S1) (3) Image of an excluded brain with high spillover into S1 and Piriform Area (Pir). (B) Percent of total starter cell population sorted by layer occupation. A two-way ANOVA followed by Tuckey’s multiple comparison test was performed to compare starter cell distribution between and within layers. Significant differences were labeled as ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05. (C) Percent of total starter cell population sorted by region occupation; showing spillover. Group comparisons per region were made using one-way ANOVAs followed by Tuckey’s multiple comparison tests. Significant differences were labeled as **p<0.01, *p<0.05. (D) Comparison of percent of total input between individual brains with varying amount of spillover to either the Motor- and Somatosensory cortices (M/S) or Piriform cortex (Pir). The darkest color indicates the sample with the highest spillover and lightest color the lowest/no spillover. For detailed statistics see Supplementary file 3. n = 3 mice per condition, data shown as average ± SEM.
Figure 1—figure supplement 3. Cell counting.

Figure 1—figure supplement 3.

(A–C) Example pictures depicting (1) The raw GFP channel signal of labeled input cells, (2) Human counts laid over the raw GFP signal and (3) The automated counts laid over the segmented image of the GFP channel. Exemplary counts are shown for (A) VP, (B) Po and (C) parts of the Amygdala (CeA, LA and BLA). (D–E) Cell count comparisons, including number of sections on which the brain region was present and cells were counted as well as the average difference in cell counts per section (human vs. automated) and Relative Percent Difference (RPD). (D) Comparison within a single brain showing differences between several human counters (left part of the table) and the difference between averaged human counts and the automated counts (right part of the table). (E) Comparisons between human and automated counts for three different brains.
Figure 1—figure supplement 4. Workflow of tracing quantification and data analysis.

Figure 1—figure supplement 4.

All data available at https://github.com/GogollaLab/tracing_quantification_and_analysis (Gehrlach, 2020; copy archived at https://github.com/elifesciences-publications/tracing_quantification_and_analysis). Pipeline to detect (A) RV+ input neurons to the IC and (B) AAV+ output neurons from the IC. Brains were fixed and coronally sectioned (thickness: 70 µm). Every second section was stained for DAPI and imaged either using slide scanner epifluorescent microscopy (RV) or scanning confocal microscopy (AAV). For RV, positive cells were identified using supervised machine learning (https://github.com/GogollaLab/tracing_quantification_and_analysis/blob/master/autonomous_neuron_detection.ijm) and allocated to manually adjusted ROIs corresponding to the Paxinos and Franklin mouse brain atlas (https://github.com/GogollaLab/tracing_quantification_and_analysis/blob/master/counting_RV.ijm). For AAV tracings, YFP-positive pixels were segmented with hessian ridge detection (https://github.com/GogollaLab/tracing_quantification_and_analysis/blob/master/autonomous_pixel_detection.ijm) and allocated to manually adjusted ROIs from the mouse brain atlas (https://github.com/GogollaLab/tracing_quantification_and_analysis/blob/master/counting_AAV.ijm). (C) Pipeline of data analysis in python. After cleaning of the raw data, it was clustered and calcualtions done before a final clustering. The output was pivot tables. https://github.com/GogollaLab/tracing_quantification_and_analysis/blob/master/analysis_AAV.py and https://github.com/GogollaLab/tracing_quantification_and_analysis/blob/master/analysis_RV.py. (D) Assessment of the specificity and spread of experimental and control conditions. In contrast to experimental conditions (red: Camk2a-Cre, N = 9 mice; blue: Gad2-Cre, N = 9 mice), we could not detect long-range RV+ neurons in the wildtype- (black, N = 2 mice) and TVA control conditions (grey, N = 2 mice). This confirms, that in the experimental conditions, the brain-wide signals are indeed from a transsynaptic retrograde transfer. Further, combining the results from the WT controls (black) and TVA controls (grey) revealed that there is some leakage of the AAV-FLEX system and that SADdG-eGFP(EnvA) can still infect a minor fraction of TVA-negative neurons. Guided by these control experiments, we omitted quantification of RV+ neurons for ±1 mm from the injection center within the IC and claustrum.