Connectivity of mouse somatosensory and prefrontal cortex examined with trans-synaptic tracing

Abstract

Information processing in neocortical circuits requires integrating inputs over a wide range of spatial scales, from local microcircuits to long-range cortical and subcortical connections. We used rabies virus-based trans-synaptic tracing to analyze the laminar distribution of local and long-range inputs to pyramidal neurons in the mouse barrel cortex and medial prefrontal cortex (mPFC). In barrel cortex, we found substantial inputs from layer 3 (L3) to L6, prevalent translaminar inhibitory inputs, and long-range inputs to L2/3 or L5/6 preferentially from L2/3 or L5/6 of input cortical areas, respectively. These layer-specific input patterns were largely independent of NMDA receptor function in the recipient neurons. mPFC L5 received proportionally more long-range inputs and more local inhibitory inputs than barrel cortex L5. Our results provide new insight into the organization and development of neocortical networks and identify important differences in the circuit organization in sensory and association cortices.

Figure 1

Layer-specific input tracing in mouse barrel cortex. (a) Experimental design and timeline of layer-specific RV tracing. (b–d) Layer-specific tracing in barrel cortex. Left, representative coronal sections showing local tracing (insets: confocal images of starter cells). Starter cells (SCs, Middle) and local input (Right) distributions are quantified according to cortical layers for SCs in L2/3 (b: SCs, 76±27; total Inputs, 1,148±340; n=6 mice), L5 (c: SCs, 84±27; total Inputs, 1,677±359; n=9 mice), and L6 (d: SCs, 117±42; total Inputs, 1,751±542; n=9 mice). Scale bars represent 200 μm or 20 μm (insets). See Supplementary Table 1 for numerical values, and Supplementary Fig. 1 for controls for RV tracing. (e–h) L3→L6 CRACM. (e) Example coronal sections from an Ntsr1^Cre;Rosa^Ai14 mouse expressing ChR2-mVenus in L2/3. Right: recorded L5 pyramidal (pyr) and tdTomato+ L6 cell (Ntsr1). Scale bars represent 100 μm (left panels) and 10 μm (zoomed panels, right). (f) Traces from a L5 and L6 cell following laser stimulation of ChR2-expressing L3 axons in the presence of TTX, 4-AP, and PTX. (g) Evoked EPSC amplitudes in L5 and L6 (L5: 205.2±57.48 pA, n=6 cells; L6: 23.05±5.50 pA, n=6 cells from 3 mice; p=0.0052, Student's t-test). (h) Representative trace (top) and quantification (bottom) of blockade of light-evoked EPSC following bath DNQX application (Baseline: 35.89±2.64 pA, DNQX: 2.64±1.61 pA, n=6 cells from 3 mice, p=0.01, paired t-test). All summary statistics are presented as mean±SEM. See Supplementary Table 6 for test results and p-values.

Figure 2

Spatial analyses of synaptic inputs from barrel cortex to starter cells in L2/3, L5, and L6. (a) Schematic of fractional analysis of inputs along the A–P axis. (b) Schematic of fractional analysis of inputs along M–L axis in the central sections. (c–e) Left, heat maps of distribution of starter cells (SCs, lower panels) and inputs (upper panels) along the A–P axis for inputs to L2/3 (n=6 mice), to L5 (n=9 mice), and to L6 (n=9 mice). Colors represent fraction of total barrel cortex inputs according to the index at right. Bin widths are 120 μm. Right, fraction of inputs in central versus peripheral sections. (f–h) Left, heat maps of distribution of starter cells (SCs, lower panels) and inputs (upper panels) along the M–L axis for inputs to L2/3 (n=5 mice), to L5 (n=6 mice), and to L6 (n=9 mice). Colors represent fraction of barrel cortex inputs within the sections analyzed. Bin widths are 120 μm. Right, quantification of fraction of inputs in middle versus side regions. * denotes significant p-values from multiple t-tests with Holm-Sidak correction for multiple comparisons (α=0.05). Summary statistics presented as mean±SEM. See Supplementary Table 6 for test results and p-values.

Figure 3

Analysis of local inhibitory inputs to starter cells in L2/3, L5, and L6. (a) Representative coronal sections of RV tracing to L2/3 in a SepW1^Cre mouse combined with Gad1/2 in situ hybridization. Inhibitory inputs are double-positive for GFP (green) and Gad1/2 (red) and are labeled with arrows. Starter cells are double-positive for GFP (green) and mCherry (white) and a subset are indicated with arrowheads. Excitatory inputs are green-only. Scale bar represents 200 μm (1^st panel) or 100 μm (2^nd–4^th panels, magnified from the dashed box in the 1^st panel). (b–d) Starter cells are mostly restricted to the defined layer (L2/3 SCs: 48±19, n=3 mice; L5 SCs: 16±2, n=3 mice; L6 SCs: 95±6, n=3 mice). (e–g) Fraction of input (Gad⁺ and Gad⁻) within sections on which Gad1/2 in situ hybridization was performed (L2/3 total inputs: 342±154, n=3 mice; L5 total inputs: 983±173, n=3 mice; L6 total inputs: 605±184, n=3 mice). (h–j) Inhibitory inputs are distributed throughout all layers in a pattern that was specific to the layer origin of starter cells (L2/3 Gad⁺ inputs: 91±16, n=3 mice; L5 Gad⁺ inputs: 34±7, n=3 mice; L6 Gad⁺ inputs: 52±16, n=3 mice). (k–m) Inhibitory inputs occupy specific fractions of inputs within each layer. Summary statistics presented as mean±SEM. See Supplementary Table 3 for numerical values).

Figure 4

Laminar analyses of long-range inputs to starter cells in L2/3, L5, and L6. (a–c) Atlas locations and example images showing long-range inputs to barrel cortex L5 from (a) secondary motor cortex (M2), (b) primary motor cortex (M1), primary somatosensory cortex (S1_body), (c) somatosensory thalamus (VPM/POm) and secondary somatosensory cortex (S2). Scale bars represent 500 μm (hemisection images), or 100 μm (zoomed images). (d) Quantification of fraction of total inputs contributed by each area. (e–r) Regional analysis of thalamic inputs (e–g) and laminar analysis of long-range inputs to L2/3 (n=6 mice), L5 (n=9 mice), and L6 (n=9 mice) from S1_body (h–j), S2 (k–m), M1 (n–p), and M2 (q–r). Summary statistics presented as mean±SEM. Schematics are modified from Paxinos and Franklin, 2001.

Figure 5

Layer-specific tracing from starter cells lacking GluN1. (a) Experimental design. RV tracing was performed as described in Fig. 1, except on mice in which starter cells (as well as other Cre-expressing cells in the starter cell layer, blue circles) lacked GluN1. (b–d) Electrophysiological characterization of GluN1 mutant cells. (b) Synaptic currents measured in Ntsr1^Cre;GluN1^fl/+ (heterozygous control, left: −60 mV: 34.6±8.4 pA; +40 mV: 47.7±13.3 pA, n=6 cells from 2 mice) and Ntsr1^Cre;GluN1^fl/Δ (mutant, right: −60 mV: 26.2±3.1 pA; +40 mV: 3.1±1.0 pA, n=8 cells from 2 mice). (c) Quantification of NMDA synaptic current in heterozygous control (black) and mutant (red) (fl/+: 47.69±13.33 pA, n=6 cells from 2 mice; fl/Δ: 3.07±0.97 pA, n=8 cells from 2 mice, p=0.002, Student's t test). (d) Example recordings from Ntsr1^Cre;GluN1^fl/+ (black) and Ntsr1^Cre;GluN1^fl/Δ (red) cells. (e, h, k) Representative coronal sections of layer-specific RV tracing in which starter cells lack GluN1. Scale bars represent 200 μm. (f, i, l) Starter cells are mostly restricted to the defined layer. (g, j, m) Comparisons of the patterns of input to control (same as data in Fig. 1) versus GluN1-lacking starter cells (L2/3 SCs: 18±10, inputs: 564±117, n=4 mice; L5 SCs: 152±64, inputs: 4,447±1559, n=6 mice L6→L5, p=0.006; L6 SCs: 146±26, inputs: 1,583±348, n=7 mice, L6→L6 p=0.002). * denotes significant p-values from Student's t-test (c) or multiple t-tests with Holm-Sidak correction for multiple comparisons (α=0.05) (g, j, m). Summary statistics presented as mean±SEM. See Supplementary Table 4 for numerical values. See Supplementary Table 6 for test results and p-values.

Figure 6

Local input to mPFC L5. (a) Low (left) and high (right) magnification example images of tracing from L5 mPFC neurons. Scale bar represents 200 μm. (b) Fraction of starter cells by layer. (c) Fraction of local inputs by layer (colored, SCs: 114±33, total inputs: 2,641±664, n=4 mice) in comparison with barrel cortex (BC, grey, n=8 mice). (d, e) Left, heat maps of distribution of starter cells (SCs, lower panel) and inputs (upper panel) in each layer along the (d) A–P or(e) D–V axis. Colors represent fraction of total mPFC cells according to the index at right. Bin widths are 120μm. Right, fraction of mPFC inputs in central versus peripheral sections (d; n=4 mice) or in middle versus side regions (e; n=4 mice). (f) Tracing from L5 starter cells with Gad1/2 in situ hybridization. Arrows indicate Gad⁺ inputs. Scale bar represents 100 μm. (g) Fraction of starter cells by layer (n=3 mice). (h) Fraction of total Gad⁺ cells by layer (n=3 mice). (i) Gad⁺ fraction of total inputs in each layer (n=3 mice). (j) Comparison of Gad⁺ fraction of local inputs to L5 mPFC versus L5 BC (mPFC: 0.24±0.04, n=3 mice; BC: 0.09±0.008, n=3 mice; p=0.03, Student's t-test). (k) Comparison of the local fraction of total inputs to mPFC and BC L5 (mPFC: 0.21±0.02, n=4 mice; BC: 0.79±0.6, n=8 mice; p=5.14×10⁻⁵, Student's t-test). (l) Spread of local inputs along the A–P axis within mPFC (n=4 mice) or BC (n=8 mice), presented as fraction of total inputs. (m) Spread of local inputs along the D–V axis within mPFC (n=4 mice) or BC (n=8 mice), presented as fraction of total inputs. (n) Quantification of cell density based on NeuN staining (mPFC: 97,633±9010 cells/mm³, n=5 sections from 2 mice; BC: 102,760±4876 cells/mm³, n=8 sections from 3 mice; p=0.59, Student's t-test). * denotes significant p-values from Student's t-tests (j, k, n) or multiple t-tests with Holm-Sidak correction for multiple comparisons (α=0.05) (c, l, m). Summary statistics presented as mean±SEM. See Supplementary Table 5 for numerical values. See Supplementary Table 6 for test results and p-values.

Figure 7

Long-range input to mPFC L5. (a–g) Example images of long-range inputs to L5 mPFC. (h) Quantification of major inputs to L5 mPFC (n=4 mice). (i–q) Detailed quantification of subregional inputs to mPFC L5 (n=4 mice). Abbreviations not mentioned in text: AVPV, anteroventral periventricular nucleus; Ect, ectorhinal cortex; Ent, entorhinal cortex; LHA, lateral hypothalamic area; LPO, lateral preoptic area; MPO, medial preoptic area; MS, medial septum; NDB, nucleus of the diagonal band; Pir, piriform cortex; PRh, perirhinal cortex; PTLp, posterior parietal association area; PVH, paraventricular hypothalamic nucleus; RSP, retrosplenial cortex; SUMl, supramammillary nucleus, lateral part; ZI, zona inserta. Scale bars represent 500 μm (hemisection images) or 100 μm (zoomed images).

Figure 8

Summary of major findings. (a) Summary of major connections in barrel cortex. (b) Summary of major inhibitory connections in barrel cortex. (c) Summary of laminar analysis of long-range inputs to barrel cortex L2/3, L5 and L6. (d) Summary of major connections in mPFC. (e) Summary of major inhibitory connections in mPFC. Arrow thickness in a, b, d, e, and circle diameter in c correspond to strength of fractional inputs. Only inputs that consist of more than 10% of total inputs to a given layer are represented by arrows. Dotted lines between L2 and L3, as well as L5a and L5b, indicate that while our experiments distinguish the input from these layers (arrows leaving from the middle of these layers), we treat them collectively as recipient of inputs (arrows pointing to the dotted lines).