The precise temporal pattern of prehearing spontaneous activity is necessary for tonotopic map refinement

Abstract

Patterned spontaneous activity is a hallmark of developing sensory systems. In the auditory system, rhythmic bursts of spontaneous activity are generated in cochlear hair cells and propagated along central auditory pathways. The role of these activity patterns in the development of central auditory circuits has remained speculative. Here we demonstrate that blocking efferent cholinergic neurotransmission to developing hair cells in mice that lack the α9 subunit of nicotinic acetylcholine receptors (α9 KO mice) altered the temporal fine structure of spontaneous activity without changing activity levels. KO mice showed a severe impairment in the functional and structural sharpening of an inhibitory tonotopic map, as evidenced by deficits in synaptic strengthening and silencing of connections and an absence in axonal pruning. These results provide evidence that the precise temporal pattern of spontaneous activity before hearing onset is crucial for the establishment of precise tonotopy, the major organizing principle of central auditory pathways.

Figure 1. Temporal patterns of spontaneous pre-hearing activity differ between MNTB neurons of WT and α9 KO mice

(A) Scheme of ascending neuronal pathways to the LSO. Excitatory pathways are green, inhibitory pathways are red. VCN – ventral cochlear nucleus. The tonotopic organization is reflected by shading (low frequency areas are dark, high frequency areas are white). (B–C) Example in vivo recordings of the spontaneous spiking activity of MNTB neurons in an 8-day-old WT (B) and α9 KO mouse (C). Neural traces are shown on increasingly finer time scales from left to right, as indicated by the scale bars. Blue bars indicate activity bursts. Red brackets mark the portion of the traces shown at higher resolution on the right. Individual events exhibited complex waveforms composed of pre- (Pr) and postsynaptic components (Po) typical for MNTB neurons. (D–E) Inter-spike interval (ISI) histograms plotted on a log scale for the neurons shown in (B) and (C), respectively. (F) Normalized population ISI histograms of WT (black, n = 21 cells) and α9 KO mice (red, n = 21 cells). In α9 KO mice the second peak is shifted to shorter intervals. (G–L) Quantification of spiking activity in WT (n = 21 cells) and α9 KO mice (n = 21 cells). The mean overall firing rate (G) and the frequency of bursts (H) did not differ between the two genotypes. The duration of bursts (I) was reduced in α9 KO mice (WT: 5.5 ± 0.8 s; KO: 2.9 ± 0.3 s; p = 0.0036, Student’s t-test). The mean firing rate during bursts (J) was increased in α9 KO mice (WT: 8.4 ± 0.7 Hz; KO: 15.4 ± 1.3 Hz; p = 2.6 × 10⁻⁵, Student’s t-test), as was the mean frequency of mini-bursts (K) (WT: 2.1 ± 0.2 Hz; KO: 3.6 ± 0.4 Hz; p = 0.0018, Student’s t-test) and single spikes that did not form mini-bursts (L) (WT: 3.3 ± 0.3 Hz; KO: 6.0 ± 0.4; p = 3.4 × 10⁻⁴). (M) Mean intervals between spike clusters (mini-bursts and single spikes together) plotted as a function of position within a burst (black: WT, 214 bursts; red: KO, 216 bursts). Positions 1 through 5 denote the first 5 intervals at the beginning of a burst, whereas positions −5 through −1 denote the last 5 intervals at the end of a burst. All error bars denote s.e.m. See also Figure S1

Figure 2. Topographic refinement of functional MNTB-LSO connectivity is impaired in α9 KO mice

(A) Examples of MNTB input maps in brain slices obtained from WT mice (upper row) and α9 KO mice (lower row). Example traces show postsynaptic currents (PSCs) in LSO neurons elicited by uncaging glutamate at corresponding MNTB sites. The amplitudes and transferred charge of the PSCs are quantified in Figure S3 and Table S2 The MNTB is outlined in black. Stimulation locations are depicted as squares and locations in the MNTB from which responses could be elicited are shown filled in red (see Experimental Procedures for details). (B) Mean input areas normalized to the corresponding MNTB area in WT and α9 KO mice aged P1-3 and P10-12. In WT mice, normalized MNTB input areas decreased by 63% (from 27 ± 5% at P1-3 (n = 7), to 10 ± 2% at P10-12 (n = 14); p < 0.01, Student’s t-test), while in KO mice input areas decreased by 44% (from 32 ± 3% at P1-3 (n = 6), to 18 ± 2% at P10-12 (n = 16); p < 0.01). At P10-12, input maps in KO mice were 80% larger than in WT mice (p = 0.01). (C) Developmental changes in the normalized medio-lateral width of MNTB input maps in WT and α9 KO mice. In WT mice, the width decreased by 45% (from 33 ± 5% at P1-3 (n = 7), to 18 ± 2% at P10-12 (n = 14); p < 0.01), whereas in α9 KO mice the width decreased by 34% (from 41 ± 3% at P1-3 to 27 ± 3% at P10-12; p < 0.01). At P10-12, the tonotopic width of input maps was 50% larger in α9 KO mice than in WT mice (p = 0.01). See also Figure S2 and Table S1.

Figure 3. Developmental strengthening and elimination of MNTB-LSO connections is impaired in α9 KO mice

(A–B) Examples of synaptic responses in LSO neurons elicited by minimal (min) and maximal (max) stimulation of MNTB fibers in neonatal (P1-3) WT (A) and α9 KO mice (B). Insets represent 5–30 superimposed individual current traces. (C–D) Examples of minimal and maximal synaptic responses in LSO neurons of P10-11 WT (C) and α9 KO mice (D). (E) Average amplitudes of synaptic responses elicited by minimal and maximal stimulation. In WT animals, maximal responses increased from 0.75 ± 0.12 nA at P1-3 (n = 26) to 4.23 ± 0.54 nA at P10-12 (n = 42), and minimal responses increased from 30 ± 4 pA at P1-3 (n = 22) to 470 ± 93 pA at P10-12 (n = 43). In α9 KO mice, maximal responses increased from 0.82 ± 0.16 nA at P1-3 (n = 18) to 3.93 ± 0.36 nA at P10-12 (n = 57) and minimal responses from 29 ± 9 pA at P1-3 (n = 18) to 200 ± 52 pA at P10-12 (n = 57). The amplitude of minimal responses at P10-12 was significantly smaller in α9 KO mice compared to WT mice (*p < 0.01, Student’s t-test). (F) Bootstrap analysis of MNTB-LSO convergence. In WT mice, the mean convergence decreased from 23 ± 6 MNTB fibers per LSO neuron at P1-3 to 10 ± 2 MNTB fibers per LSO neuron at P10-12. In KO mice, the mean convergence was 27 ± 7 MNTB fibers per LSO neuron at P1-3 (not significantly different from WT mice, p = 0.66, permutation test) and 21 ± 6 MNTB fibers/LSO neuron at P10-12, significantly higher than WT mice (p = 0.03, permutation test) (G) Amplitude and cumulative histograms of minimal stimulation responses in WT and α9 KO mice (*p = 0.001, Kolmogorov-Smirnov test). See also Figure S4.

Figure 4. Growth of MNTB axon terminals and formation of new boutons before hearing onset

(A–B) Example of reconstructed MNTB axon terminals in the medial LSO (outlined in blue) in WT mice (A) and α9 KO mice (B). Lower portions illustrate bouton locations of corresponding axons. Bouton clouds are fit with an ellipse that maximizes enclosed bouton density (see Experimental Procedures). D, dorsal; L, lateral. Scale bar, 100 μm. (C) Increase in the average number of boutons per MNTB axon in the LSO before hearing onset is similar in WT and α9 KO mice (P2-4: 127 ± 13 in WT vs. 112 ± 12 in KO, p = 0.39, Student’s t-test; P12-14: 301 ± 29 in WT vs. 324 ± 26 in KO, p = 0.55, Student’s t-test). (D) The average bouton area, normalized to LSO cross-sectional area, does not change before hearing onset and does not differ between WT and α9 KO mice at either age (P2-4: 19.7 ± 1.7% in WT vs. 15.5 ± 2.2% in KO, p = 0.14, Student’s t-test; P12-14: 19.6 ± 1.1% in WT vs. 21.3 ± 2.6% in KO, p = 0.54, Student’s t-test). (E) The average width of boutons along the LSO tonotopic axis, normalized to LSO length, does not change before hearing onset and does not differ between WT and α9 KO mice at either age (P2-4: 23.3 ± 2.3% in WT vs. 20.6 ± 1.5% in KO, p = 0.33, Student’s t-test; P12-14: 20.7 ± 1.1% in WT vs. 23.1 ± 2.1% in KO, p = 0.28, Student’s t-test). Errors bars, s.e.m. P2-4: n = 9 axons from 6 WT animals and 9 axons from 8 α9 KO animals. P12-14: n = 10 axons from 8 WT animals and 8 axons from 6 α9 KO animals. See Figure S6 for absolute values.

Figure 5. Changes in the spatial distribution of boutons formed by single axons before hearing onset

(A) Heat maps of the average density of boutons of individual MNTB axons in the LSO. Color bar: average number of boutons per axon per bin (0.08% LSO area). Scale bars, 20% LSO length and 40% dorso-ventral LSO height. (B) Histograms of the average distribution of boutons per MNTB axon along the LSO frequency axis in WT mice at P2-4 (grey) and P12-14 (black). The x-axis represents bouton positions with respect to the centroid of the bouton fields, and the bin size was 2% of the LSO length. Asterisks (*) mark the bins where the mean number of boutons significantly differed (p < 0.05, Student’s t-test) between the P2-4 (n = 9 axons) and P12-14 (n = 10 axons) WT groups. Error bars represent s.e.m. (C) Heat maps for α9 KO mice. (D) Histograms for α9 KO mice. Asterisks (*) mark the bins that are significantly different between the P2-4 (magenta, n = 9 axons) and P12-14 (red, n = 8 axons) KO groups. The histogram for WT P12-14 from (B) is also plotted for comparison (black). The mean bouton numbers did not significantly differ between genotypes at P12-14 (black vs red) at any bins (p > 0.1, Student’s t-test). See Figure S6 for absolute values.

Figure 6. Pruning of MNTB axon terminals and elimination of boutons after hearing onset

(A) Example of reconstructed MNTB axon terminals in the LSO (outlined in blue) in WT mice and (B) α9 KO mice. Lower portions illustrate bouton locations of corresponding axons. D, dorsal; L, lateral. Scale bar, 100 μm. (C) Average number of boutons per MNTB axon in the LSO decreases in WT mice, but not in α9 KO mice (WT: 302 ± 29 boutons/MNTB axon at P12-14 to 137 ± 21 boutons at P19-21, p < 0.001, Student’s t-test; KO: 325 ± 26 boutons/MNTB axon at P12-14 to 278 ± 28 at P19-21, p = 0.25, Student’s t-test; WT vs. KO at P19-21: p = 0.001, Student’s t-test). (D) Average bouton area normalized to LSO cross-sectional area decreases in WT mice, but not in α9 KO mice (WT: 20 ± 1% LSO cross-sectional area at P12-14 to 10 ± 2% at P19-21, p < 0.001, Student’s t-test; KO: 21 ± 3% LSO cross-sectional area at P12-14 to 22 ± 1% at P19-21, p = 0.75, Student’s t-test; WT vs. KO at P19-21: p < 0.0001). (E) Average width of bouton area along LSO tonotopic axis decreases in WT mice, but not in α9 KO mice (WT: 21 ± 1% LSO length at P12-14 to 14 ± 2% at P19-21, p = 0.002, Student’s t-test; KO: 23 ± 2% LSO length at P12-14 to 27 ± 3% at P19-21, p = 0.34, Student’s t-test; WT vs. KO at P19-21: p = 0.002). Errors bars, s.e.m. At P12-14, n = 10 axons from 8 animals in WT mice and n = 10 axons from 6 animals in α9 KO mice. At P19-21, n = 9 axons from 8 animals in WT mice and n = 9 axons from 8 animals in α9 KO mice. See Figure S6 for absolute values.

Figure 7. Changes in the spatial distribution of boutons after hearing onset

(A) Heat maps of average density of the boutons of individual MNTB axons in the LSO of WT mice. Color bar: average number of boutons per axon per bin (0.08% LSO area). Scale bars, 20% LSO length and 40% dorso-ventral LSO height. (B) Histograms of the average distribution of boutons per MNTB axon along the LSO frequency axis at P12-14 (grey circles) and P19-21 (black squares). The x-axis represents bouton positions with respect to the centroid of the bouton fields, and the bin size was 2% of the LSO length. Asterisks (*) mark the bins where the mean number of boutons significantly differed (p < 0.05, Student’s t-test) between the P12-14 (n = 10 axons) and P19-21 (n = 9 axons) WT groups. Also shown is the downshifted P12-14 histogram (thick grey line) to match the peak of the P19-21 distribution. Error bars represent s.e.m. (C) Heat maps for α9 KO mice. (D) Histograms for α9 KO mice showing the P12-14 (magenta, n = 8 axons) and P19-21 (red, n = 9 axons) groups, which did not differ significantly at any bins. The histogram for the WT P19-21 group from (B) is also plotted for comparison (black). The asterisks on the black curve mark the bins where the mean bouton numbers significantly differed between genotypes at P19-21 (black vs red; Student’s t-test, p < 0.05).

Figure 8. Rearing mice in pulsed white noise does not prevent axonal pruning after hearing onset

(A) Example of reconstructed MNTB axon terminals in the LSO in WT mice (left; same axon shown in Figure 6) and noise-reared WT (WT-NR) mice (right) at P19-21. Lower portions illustrate bouton locations of corresponding axons. D, dorsal; L, lateral. Scale bar, 100 μm. (B) Average number of boutons per MNTB axon in the LSO does not differ between WT and WT-NR mice at P19-21 (137 ± 21 boutons/MNTB axon in WT compared to 166 ± 14 boutons/MNTB axon in WT-NR, p = 0.26, Student’s t-test), but is significantly reduced compared to WT mice at P12-14 (301 ± 29 boutons/MNTB axon, p < 0.0005 for both WT and WT-NR). (C) Average bouton area normalized to LSO cross-sectional area also does not differ between WT and WT-NR mice at P19-21 (10 ± 2 % LSO area in WT compared to 11 ± 1 % LSO area in WT-NR, p = 0.34, Student’s t-test), but is significantly reduced compared to WT mice at P12-14 (20 ± 1 % LSO area, p < 0.0005 for both WT and WT-NR). (D) Average width of bouton area along LSO tonotopic axis also does not differ between WT and WT-NR mice at P19-21 (14 ± 2 % LSO length in WT compared to 14 ± 1 % LSO length in WT-NR, p = 0.71, Student’s t-test), but is significantly reduced compared to WT mice at P12-14 (21 ± 1 % LSO length, p < 0.002 for both WT and WT-NR). Errors bars, s.e.m. n = 9 axons from 8 animals in WT mice and n = 14 axons from 11 animals in WT-NR mice. (E) Heat maps of average density of the boutons of individual MNTB axons in the LSO of WT (left) and WT-NR (right) mice at P19-21. Color bar: average number of boutons per axon per bin (0.08% LSO area). Scale bars, 20% LSO length and 40% dorso-ventral LSO height. (F) Histograms of the average distribution of boutons per MNTB axon along the LSO frequency axis at P19-21 in WT (black circles) and WT-NR (green squares). The x-axis represents bouton positions with respect to the centroid of the bouton fields, and the bin size was 2% of the LSO length. Error bars represent s.e.m. The mean bouton numbers did not significantly differ between WT and WT-NR at any of the bins (Student’s t-test, p > 0.05).

Figure 9. Projections from the cochlear nucleus to the superior olivary complex appear normal in α9 KO mice

(A) Pattern of cochlear nucleus projections to the ipsilateral and contralateral superior olivary complex in a P 13 WT mouse and P12 α9 KO mouse following insertion of a biocytin crystal into the left cochlear nucleus. In both genotypes, labeled axon terminals were predominantly found in the ipsilateral LSO and contralateral MNTB. A few MNTB neurons, but never calyces, were observed in the ipsilateral MNTB in both genotypes. (B) Normal morphology of calyx of Held in α9 KO mice. Example of rhodamine dextran labeled calyces of Held in P14 WT and α9 KO mice. Each image is a maximum intensity projection of a stack of two-photon optical sections. Scale bars are 5 μm.