Cortical sensory maps are remodeled during early life to adapt to the surrounding environment. Both sensory and contextual signals are important for induction of this plasticity, but how these signals converge to sculpt developing thalamocortical circuits remains largely unknown. Here we show that layer 1 (L1) of primary auditory cortex (A1) is a key hub where neuromodulatory and topographically organized thalamic inputs meet to tune the cortical layers below. Inhibitory interneurons in L1 send narrowly descending projections to differentially modulate thalamic drive to pyramidal and parvalbumin-expressing (PV) cells in L4, creating brief windows of intracolumnar activation. Silencing of L1 (but not VIP-expressing) cells abolishes map plasticity during the tonotopic critical period. Developmental transitions in nicotinic acetylcholine receptor (nAChR) sensitivity in these cells caused by Lynx1 protein can be overridden to extend critical-period closure. Notably, thalamocortical maps in L1 are themselves stable, and serve as a scaffold for cortical plasticity throughout life.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank J. Miwa (Lehigh University, Bethlehem, PA, USA) for providing the original Lynx1-knockout breeder mice and Lynx1-knockout × α7-nAChR-knockout mice (care of A. Beaudet, Baylor University, Waco, TX, USA); H. Monyer (Heidelberg University, Heidelberg, Germany) for the original PV–GFP mice; M. Fagiolini for comments and discussions; A. Covello and M. Snyder for assistance with Brainbow tracing; A. Castros and E. He for assistance with immunocytochemistry quantification and in situ hybridization; A. Galvin for ABR measurements to test DHβE effects; T. Barkat, E. Diel, H. Lee, R. Reh, S. Mierau, A. Patrizi, D. Cai, G. Corfas, J. Holt, G. Géléoc, A. Thompson, and C. Chen for technical assistance and discussions; and N. De Souza, M. Nakamura, H. Bond, E. Centofante, and M. Marcotrigiano for animal care. This work was supported by the Nancy Lurie Marks Family Foundation and the Canadian Institute for Advanced Research (CIFAR; to A.E.T.), the Ellison Medical Foundation, and the NIMH Silvio O Conte Center (P50MH094271 to T.K.H.).
Integrated supplementary information
Projections from the medial geniculate body (MGB; red) onto 5-HT3AR interneurons within layer 1 of primary auditory cortex (A1) colocalize with vesicular glutamate transporter 2 (vGluT2; green). Representative images from one of 3 mice. Scale bar = 5 µm.
(a) Layer (L) 1 interneurons and L4 pyramidal cells in primary auditory cortex (A1) show similar responses to medial geniculate body (MGB) stimulation. Left; Schematic illustrating recording configuration. L1 interneurons and L4 pyramidal cells were recorded during electrical stimulation of the MGB fibers. Right; Individual (gray) and mean (bolded) MGB-evoked EPSPs recorded in representative L1 late-spiking (LS) or non-LS cell subtypes and a L4 pyramidal cell. (b) Left; Mean (± SEM) EPSP amplitudes evoked by maximal MGB stimulation (EPSP maximum amplitude (mV): L1 LS = 7.34 ± 1.41, n = 7 cells/4 mice; L1 non-LS = 6.16 ± 2.14, n = 4 cells/4 mice; L4 = 7.33 ± 1.96, n = 11 cells/4 mice; Kruskal-Wallis test, χ2(2) = 0.47, P = 0.79). Right; Mean (± SEM) EPSP onset time, EPSP 10–90% rise time, and width at half maximum EPSP amplitude recorded in L1 interneurons and L4 pyramidal cells (onset time (ms): L1 = 11.36 ± 0.80, n = 11 cells/4 mice; L4 = 8.11 ± 0.84, n = 11 cells/4 mice; unpaired t test, two-tailed, t(20) = −2.80, P = 0.011, CL = −5.66, −0.83; rise time (ms): L1 = 8.18 ± 1.39, n = 11 cells/4 mice; L4 = 7.45 ± 1.35, n = 11 cells/4 mice; Mann-Whitney U test, two-tailed, z = −0.53, P = 0.599, CL = −3.13, 1.97; half-width (ms): L1 = 50.95 ± 4.96, n = 11 cells/4 mice; L4 = 57.64 ± 8.09, n = 11 cells/4 mice; Mann-Whitney U test, two-tailed, z = 0.33, P = 0.743, CL = −15.55, 32.66). (c) L1 interneurons and L4 pyramidal cells exhibit similar short-term depression of MGB-evoked EPSPs. Left; Schematic illustrating recording configuration. Trains of electrical stimuli were delivered to the MBG fibers during recordings in L1 interneurons or L4 pyramidal cells. Right; Mean (± SEM, shaded region) EPSPs recorded in all L1 interneurons (green, n = 8 cells/3 mice) and all L4 pyramidal cells (gray, n = 10 cells/3 mice) during repetitive electrical stimulation of the MGB fibers (10Hz). (d) Left; Mean (± SEM) EPSP amplitudes normalized to the first EPSP (EPSPN/EPSP1) evoked by each of 10 stimuli applied to the MGB fibers at 10 or 20Hz (L1, n = 8 cells/3 mice; L4, n = 10 cells/3 mice). Box indicates EPSP10/EPSP1. Right; Mean (± SEM) EPSP10/EPSP1 shows no significant differences between L1 interneurons and L4 pyramidal cells (10 Hz, L1 = 0.36 ± 0.06, L4 = 0.66 ± 0.18, z = −1.64, P = 0.10, CL = −0.06, 0.43; 20 Hz, L1 = 0.20 ± 0.04, L4 = 0.23 ± 0.04, z = −0.49, P = 0.63, CL = −0.11, 0.15, Mann-Whitney U tests, two-tailed, L1, n = 8 cells/3 mice; L4, n = 10 cells/3 mice). n.s. (not significant), P > 0.05; *P < 0.05.
Supplementary Figure 3 5-HT3AR+ interneurons differentially affect MGB-evoked EPSPs in PV and pyramidal cells.
(a) Schematic of recording configuration. (b) Example of ventral medial geniculate body (MGBv)-evoked subthreshold EPSPs recorded in L4 PV cells alone (red) or during optogenetic activation of 5HT3AR interneurons (blue). (c) Mean (± SEM) effects of 5-HT3AR interneuron activation on MGBv-evoked EPSPs in PV interneurons (red) and pyramidal cells (black) within L4 at varying intervals between 5-HT3AR interneuron activation and MGB stimulation (PV cell, n = 4 cells/4 mice; PYR cell, n = 12 cells/9 mice, Mann-Whitney U tests, two-tailed, *P < 0.05).
(a) Averaged responses (normalized ΔF/F) in primary auditory cortex (A1) to stimulation of ventral medial geniculate body (MGBv) using voltage-sensitive dye imaging (VSDI) within layers (L) 2/3, 4 and 5 of P12–15 mice (n = 10 mice) at baseline (black) or during 5-HT3AR cell optogenetic silencing (blue). (b) 5-HT3AR cell silencing decreased the early (0–40ms) peak response within L2/3/4 of activated columns (normalized ΔF/F: baseline = 0.68 ± 0.11; 5-HT3AR cell silencing = 0.49 ± 0.06, n = 10 mice; paired t test, two tailed, t(9) = −2.40, P = 0.0396, CL = −0.37, −0.01), but increased the late (200–450ms) response in neighboring columns (normalized ΔF/F: baseline = 0.51 ± 0.11; 5-HT3AR cell silencing = 0.68 ± 0.09, n = 10 mice; paired t test, two tailed, t(9) = 4.00, P = 0.0031, CL = 0.07, 0.27). Mean ± SEM, *P < 0.05, **P < 0.01.
Supplementary Figure 5 Silencing of 5-HT3AR+ cells abolishes a critical period for auditory thalamocortical plasticity.
Normalized (norm.) ∆F/F (mean ± SEM) across primary auditory cortex (A1) locations in response to ventral medial geniculate body (MGBv) stimuli imaged in slices from naïve C57 mice (P20–25) raised in a normal acoustic environment (a, n = 15 mice), 7kHz-exposed controls (b, n = 14 mice), and 7kHz-exposed mice with silenced 5-HT3AR cells using hM4D (c, n = 28 mice). Colors indicate MGBv positions. Insets show the MGBv positions that elicited the maximal average responses across A1.
Supplementary Figure 6 Developmental changes in genes encoding 5-HT3AR and nAChRs within 5-HT3AR+ interneurons.
Normalized expression (mean ± SEM) of Htr3a, Chrna4, Chrna7, and Chrnb2 genes encoding 5-HT3AR and nAChR subunits (α4, α7, and β2) within 5-HT3AR-expressing interneurons from primary auditory cortex (A1) in mice aged postnatal day (P) 11 and P20. (Normalized quantity Htr3a: P11 = 0.74 ± 0.12 (95% CI, 0.37, 1.11), P20 = 1.87 ± 0.58 (95% CI, 0.01, 3.73), z = 1.01, P = 0.31; Normalized quantity Chrna4: P11 = 1.52 ± 0.03 (95% CI, 1.43, 1.62), P20 = 1.27 ± 0.07 (95% CI, 1.05, 1.48), z = −2.17, P = 0.03; Normalized quantity Chrna7: P11 = 1.08 ± 0.07 (95% CI, 0.85, 1.32), P20 = 1.22 ± 0.05 (95% CI, 1.06, 1.37), z = 1.30, P = 0.19; Normalized quantity Chrnb2: P11 = 1.02 ± 0.09 (95% CI, 0.73, 1.31), P20 = 1.00 ± 0.05 (95% CI, 0.83, 1.18), z = 0, P = 1.00; Mann-Whitney U tests, two-tailed, n = 4 mice each). *P < 0.05.
Expression of Lynx1 (mean ± SEM) measured within inhibitory interneuron subtypes from primary auditory cortex (A1) in adult mice. (Normalized quantity Lynx1: 5-HT3AR = 0.90 ± 0.13 (95% CI, 0.55, 1.26), n = 5 mice; non-5-HT3AR = 1.38 ± 0.13 (95% CI, 1.01, 1.74), n = 5 mice; PV = 2.10 ± 0.15 (95% CI, 1.68, 2.52), n = 5 mice; SOM = 1.22 ± 0.02 (95% CI, 1.16, 1.29), n = 4 mice; Kruskal-Wallis, χ2(3) = 14.25, P = 0.003, compared to 5-HT3AR using Steel test, non-5-HT3AR, P = 0.15; PV, P = 0.03; SOM, P = 0.10). n.s. (not significant), P > 0.05; *P < 0.05.
Supplementary Figure 8 Nicotine induces a barrage of IPSCs in PV and pyramidal cells from Lynx1-knockout mice.
(a) Schematic of recordings from PV and pyramidal (PYR) cells in primary auditory cortex (A1) from P25–45 wild-type (WT) and Lynx1 KO mice. These cell types are targeted by 5HT3AR inhibitory interneurons that express α4β2- and α7-containing nAChRs. (b, c) Examples of spontaneous inhibitory postsynaptic currents (sIPSCs) before (b) and after (c) bath application of nicotine (10 µM). Insets show expanded sIPSCs from gray region. (d) Mean (± SEM, shaded region) sIPSC frequency following nicotine application in pyramidal and PV cells from Lynx1 KO mice. Nicotine-induced increase in sIPSC frequency persisted in α7-nAChR/Lynx1 KO mice or during application of the α7-nAChR selective antagonist methyllycaconitine (MLA; bath-applied, 10 nM), but was abolished by pharmacological blockade of α4-containing nAChRs with DHβE (bath-applied, 500 nM). (PV cells: WT, n = 11 cells/10 mice; Lynx1 KO, n = 9 cells/6 mice; PYR cells: WT, n = 10 cells/7 mice; Lynx1 KO, n = 9 cells/9 mice; α7-nAChR/Lynx1 KO, n = 6 cells/6 mice; MLA, n = 10 cells/7 mice; DHβE, n = 8 cells/7 mice). (e) Mean (± SEM) nicotine effect (peak sIPSC frequency during nicotine application divided by baseline) in pyramidal and PV cells from WT and Lynx1 KO mice (nicotine effect on PV cells: WT = 1.55 ± 0.29, n = 11 cells/10 mice; Lynx1 KO = 4.98 ± 1.33, n = 9 cells/6 mice, Mann-Whitney U test, two-tailed, z = 2.74, P = 0.006, CL = 0.61, 6.74; nicotine effect on PYR cells: WT = 1.50 ± 0.18, n = 10 cells/7 mice; Lynx1 KO = 3.29 ± 0.47, n = 9 cells/9 mice; α7-nAChR/Lynx1 KO = 6.15 ± 2.80, n = 6 cells/6 mice; MLA = 4.82 ± 1.17, n = 10 cells/7 mice; DHβE = 1.35 ± 0.36, n = 8 cells/7 mice; Kruskal-Wallis, χ2(4) = 12.31, P = 0.015, compared to Lynx1 KO using Steel test, WT, P = 0.028, CL = 0.67, 3.50; α7-nAChR/Lynx1 KO, P = 1.0, CL = −3.66, 12.39; MLA, P = 0.989, CL = −2.83, 5.94; DHβE, P = 0.049, CL = −3.84, 0.09). n.s. (not significant), P > 0.05; *P < 0.05; **P < 0.01.
(a, b) Representative auditory brainstem responses (ABR) recorded from C57 and Lynx1 KO adult mice during 11.3 kHz tone pips of increasing sound pressure level (decibels; dB). Circles represent location of the five positive (P) wave peaks (1–5). (c) C57 and Lynx1 KO mice showed similar ABR thresholds (mean ± SEM) across tone frequencies (MANOVA, ABR threshold changes by genotype, P = 0.51, frequency, P = 0.0018, genotype x frequency, P = 0.77). (d) C57 and Lynx1 KO mice showed similar Distortion Product Otoacoustic Emissions (DPOAE) thresholds (mean ± SEM) across tone frequencies (MANOVA, threshold changes by genotype, P = 0.10, frequency, P < 0.0001, genotype x frequency, P = 0.50, C57, n = 4 mice; Lynx1 KO, n = 6 mice). (e, f) Representative ABR waves recorded from Lynx1 KO adult mice before and 30 minutes after administration of DHβE (i.p. 1mg/kg) during 11.3 kHz tone pips of increasing dB. (g, h) DHβE treatment did not impact ABR or DPOAE thresholds (mean ± SEM) across tone frequencies (MANOVA, ABR threshold by treatment condition, P = 0.54, frequency, P = 0.0116, treatment x frequency, P = 0.98; DPOAE threshold by treatment condition, P = 0.26, frequency, P = 0.13, treatment × frequency, P = 0.70, n = 4 mice).
Imaging of 5-HT3AR interneurons expressing the genetically-encoded calcium indicator (GCaMP6f) using 5-HT3AR-Cre x Ai95 mice. (a) Schematic of an auditory thalamocortical slice illustrates the lateral (1) and medial (6) positions stimulated within the ventral medial geniculate body (MGBv). (b) Average normalized traces (ΔF/F) across all mice (n = 3 mice) within layer (L) 1 locations of primary auditory cortex (A1) to stimulation of MGBv positions 1 and 6. (c) Mean peak responses (norm. ∆F/F) within L1 of A1 to stimulation of MGBv positions 1 and 6 across all mice (n = 3 mice). Black arrows indicate the rostro-caudal location of peak responses within L1.
5-HT3AR cells within layer (L) 1 of primary auditory cortex (A1) are driven by direct, topographically-organized inputs from the ventral medial geniculate body (MGBv). These cells are also co-activated by neuromodulators such as acetylcholine (ACh) via nAChRs. The nAChR recruitment of these cells is developmentally regulated by Lynx1, a protein that reduces nAChR function31, 32. The axons of non-late-spiking (non-LS) 5-HT3AR cells descend to L4, preferentially targeting intracolumnar PV-cell somata. This suppresses MGBv-evoked spiking in PV cells, creating temporal windows of enhanced thalamic drive to L4 pyramidal cells that may gate plasticity mechanisms.