Mutation of Npr2 leads to blurred tonotopic organization of central auditory circuits in mice

Abstract

Tonotopy is a fundamental organizational feature of the auditory system. Sounds are encoded by the spatial and temporal patterns of electrical activity in spiral ganglion neurons (SGNs) and are transmitted via tonotopically ordered processes from the cochlea through the eighth nerve to the cochlear nuclei. Upon reaching the brainstem, SGN axons bifurcate in a stereotyped pattern, innervating target neurons in the anteroventral cochlear nucleus (aVCN) with one branch and in the posteroventral and dorsal cochlear nuclei (pVCN and DCN) with the other. Each branch is tonotopically organized, thereby distributing acoustic information systematically along multiple parallel pathways for processing in the brainstem. In mice with a mutation in the receptor guanylyl cyclase Npr2, this spatial organization is disrupted. Peripheral SGN processes appear normal, but central SGN processes fail to bifurcate and are disorganized as they exit the auditory nerve. Within the cochlear nuclei, the tonotopic organization of the SGN terminal arbors is blurred and the aVCN is underinnervated with a reduced convergence of SGN inputs onto target neurons. The tonotopy of circuitry within the cochlear nuclei is also degraded, as revealed by changes in the topographic mapping of tuberculoventral cell projections from DCN to VCN. Nonetheless, Npr2 mutant SGN axons are able to transmit acoustic information with normal sensitivity and timing, as revealed by auditory brainstem responses and electrophysiological recordings from VCN neurons. Although most features of signal transmission are normal, intermittent failures were observed in responses to trains of shocks, likely due to a failure in action potential conduction at branch points in Npr2 mutant afferent fibers. Our results show that Npr2 is necessary for the precise spatial organization typical of central auditory circuits, but that signals are still transmitted with normal timing, and that mutant mice can hear even with these deficits.

Figure 1. Peripheral SGN connectivity is normal in Npr2 mutant mice.

(A,B) Cochlear innervation was visualized by neurofilament immunostaining of P18 whole cochleae. Visual inspection revealed no obvious difference in the peripheral pattern of connections between wild-type control (Ctl, n = 2) (A) and Npr2 mutant (Mut, n = 2) (B) animals. (C,D) SGN projections in the P14 cochlea were labeled by crossing Neurog1-CreER^T2 to the AI14:tdTomato reporter strain, which enables visualization of a subset of SGNs along the length of the cochlea. Individual SGNs of control animals (Ctl, n = 3 heterozygotes) (C) show neatly organized projections within the cochlea. Individual SGN processes of Npr2 mutant (Mut, n = 5) cochleae (D) showed no qualitative differences compared to controls. Note that the degree of labeling can vary slightly independent of genotype, due to fluctuations in Cre activity. (E) Electron micrograph of a transverse section of myelinated SGN axons in the eighth nerve in a control P21 animal. (F) Similar electron micrograph of the eighth nerve of an Npr2 mutant at P21 shows normal axonal diameters and normal myelination. (G) The g-ratio of Npr2 mutants did not differ from controls (P = 0.87, Student's t-test) and was near optimal. Scale bars in A, 50 µm; C, 2 µm.

Figure 2. Npr2 mutant mice show SGN central axon guidance and bifurcation defects.

(A) Schematic diagram of E16.5 embryo head showing SGN axons projecting from the cochlea into the hindbrain. The boxed area indicates the hindbrain region shown in B and C. (B,C) Dye labeling of SGN central axons in the hindbrain at E16.5. (B) SGN axons normally exhibited a stereotyped bifurcation pattern in the developing brainstem at E16.5, as shown in a control heterozygous embryo (Ctl). (C) In Npr2 mutants, axons appeared disorganized in the region where they would normally bifurcate (arrow), and the nerve root lacked the distinctive Y-shape. Dorsal (D) is up and rostral (R) is to the left. (D–I) Genetic labeling of SGN central axons at postnatal stages (P14–18) using Ngn1-cre^ERT2 and AI14-tdTomato, which allows random, relatively sparse labeling of SGNs. (D–E) Tiled confocal stack projections showing the entire cochlear nucleus of control (Ctl) (D) and Npr2 mutant (Mut) (E) animals at P14, an approach that permits the overall pattern of SGN axon organization to be qualitatively assessed. Control SGN axons projected in a highly organized fashion to the aVCN, pVCN, and DCN (D). In an Npr2 mutant, SGN axons still projected to aVCN, pVCN, and DCN, but in a disorganized pattern (E). Yellow arrowheads indicate regions in the aVCN that appear under-innervated. (F–G) Confocal stack projections of vibratome-sectioned control and Npr2 mutant cochlear nuclei at P18. Visual inspection of SGN axons confirmed the presence of stereotypical Y-shaped branch points and orderly projections to aVCN and pVCN in controls (F). In contrast, SGN projections to aVCN and pVCN were disarrayed in Npr2 mutants (G). (H–H′) Control SGN axons exhibited characteristic bifurcations (yellow arrowheads, H) and formed organized bundles of axonal branches in aVCN (H′) at P14. (I–I′) Npr2 mutant SGN axons generally turned instead of bifurcating (yellow arrowheads, I) and followed aberrant trajectories in aVCN (I′). Scale bar in A, 50 µm. Scale bars in D and F, 100 µm. Scale bar in H, 10 µm.

Figure 3. Npr2 mutant SGN axons do not bifurcate properly but can form interstitial branches and morphologically normal synapses.

(A) Injection of biocytin into the aVCN (left) in a parasagittal slice of the cochlear nuclei labeled fibers not only in the nerve root but also in the pVCN (right) in control (Ctl) animals at P18. (A′) A close up view of boxed region in A showing labeled descending branches of SGNs. (B) In an Npr2 mutant (Mut) at P19, biocytin injection into the aVCN also labeled fibers in the nerve root and in the pVCN, showing that at least some fibers branched at the nerve root. (B′) A close up view of the boxed region in B shows labeled SGN fibers and terminals in the pVCN. (C,D) Magnified views of bifurcation zone in control (C) and mutant (D) animals. (C) Examples of stereotyped bifurcations in control animals. (D) In mutants, rare fibers that do branch in the appropriate region do so at irregular angles and with one thinner and one thicker branch (arrows). (E–H) SGN axons can still form interstitial branches and elaborate morphologically normal synaptic endings in Npr2 mutants. Interstitial SGN axon branches, which were distinguished from bifurcations according to morphological criteria, are present in both control (E, arrowheads) and mutant (G, arrowheads) animals. Similarly, endbulbs of Held, which are one terminal whereby SGNs contact bushy cells, show the same types of branching patterns in the control (F) and Npr2 mutants (H). Scale bars in B, 50 µm. Scale bars in D and H, 10 µm. Dorsal is up and rostral is to the left in all panels.

Figure 4. Tonotopy of SGN axons in the cochlear nuclei is blurred in Npr2 mutants.

(A) Schematic diagram illustrating tonotopic dye labeling at E16. Labeling of SGNs in the cochlea with red (apex) and green (base) lipophilic dyes allowed their relative positions to be traced to the hindbrain. (B) In a wild-type (WT) E16 embryo, fibers from the base and apex bifurcated in the nerve root in separate bundles. (C–C′) Similar labeling in two different Npr2 mutants (Mut) shows that fibers from the base and apex were incompletely segregated in the hindbrain. The gross tonotopy was preserved, but some overlap in dye labeling was seen between axons from basal and apical SGNs (arrowheads). Scale bar, 50 µm. (D) The intensity of labeling of SGN axons revealed a caudal bias in Npr2 mutants, with more axons projecting towards the developing pVCN than the aVCN (P<0.05, Student's t-test). (E) Schematic diagram illustrating tonotopic dye labeling at P14. The planes of sections illustrated in F–H are indicated by dotted lines. (F–F″) In a control animal, axons arising from the middle and from the apex of the cochlea were segregated in the auditory nerve and in the aVCN and pVCN, as assessed qualitatively using confocal imaging. (G–G″, H–H″) Two examples of Npr2 mutants. Axons arising from the middle and apical turns of the cochlea were properly segregated in the nerve, indicating that the dyes labeled physically distinct populations of neurons as in controls (G,H). However, axons from these neurons were intermingled in the aVCN (G′, H′) and pVCN (G″, H″) of the same animals.

Figure 5. Tonotopy of TV neuron projections from the DCN to the aVCN is also blurred.

(A) Photomicrograph of a section of a parasagittal slice from a P18 heterozygote control animal with a biocytin injection into the aVCN (arrow). Biocytin labeled processes and cell bodies of neurons that passed through the injection site. Labeled processes included auditory nerve fibers (asterisk), the axons and terminals of TV cells whose cell bodies lie in the DCN (bracket), and the dendrites and local axonal collaterals of T stellate and D stellate cells whose cell bodies formed a halo around the injection site. Labeled TV cells clustered in a band (bracket) among terminals of auditory nerve fibers that were labeled by the same injection, showing that the TV cells lie in the same isofrequency lamina as their VCN targets. A few TV cells whose axons crossed the injection site on their way to more ventral regions were labeled ventral to (arrowhead) the labeled band of TV cells. No cell bodies were labeled in the DCN dorsal to the labeled band, indicative of the sharp tonotopic organization of the projection. (B) A similar injection of biocytin into the aVCN of a P19 Npr2 mutant animal labeled a more diffuse bundle of auditory nerve fibers, a halo of neurons in the aVCN, and TV cells that were more scattered than in the heterozygote (bracket). (C) To compare the distribution of labeled neurons between control and mutant animals, peaks of distributions were lined up and normalized. In WT (n = 6 slices, 258 cells) and Hets (n = 20 slices, 914 cells), labeled cells were distributed in a sharp band, with no labeled cells more than 150 µm from the peak on the dorsal side. Since no clear bands were observed in Npr2 mutants (Mut) (n = 11 slices, 858 cells), histograms were aligned along their medians, at which half of the labeled cells lay more dorsal and half more ventral. The bands were sharp in both WT and Het animals, but were significantly broader in Npr2 mutant mice (P<0.001, ANOVA). No differences were detected over the age range examined between P14 and P26. (D) Examples of reconstructed slices, with labeled cells marked with dots, gray regions denoting areas containing granule cells in the largest section, and lines indicating the location of some of the labeled fibers. Numbers of labeled TV cell bodies were plotted as a function of distance along the tonotopic axis of the DCN, as illustrated. Scale bar, 500 µm.

Figure 6. Npr2 mutants show normal auditory responses.

(A) Average ABR waveforms at 16 kHz for wild-type (WT) (purple, n = 6) and Npr2 mutant (Mut) (green, n = 15) animals. The average is shown by the dark lines, and the shaded areas show the standard error of the mean. (B) Wave 1 reflects the synchronous firing of auditory nerve fibers. Its amplitude decreased with sound level similarly in WT (purple) and Mut (green) (P>0.3 at all frequencies, Student's t-test). (C) Average ABR thresholds for WT (purple) and Npr2 mutant (green) animals across frequencies. No significant difference was observed between WT and Npr2 mutants (P>0.3, Student's t-test).

Figure 7. Convergence of SGNs onto bushy and T stellate cells tends to be lower in Npr2 mutants.

(A) Voltage-clamp recordings (−65 mV) from individual bushy cells showed that in wild-type and heterozygote controls, gradual increase in the strength of shocks applied to auditory nerve fiber bundles evoked first one or two small jumps in current, presumably from bringing to threshold one or two fibers that contacted the recorded bushy cells, and then a large jump likely from bringing to threshold a fiber that contacted the bushy cell with an endbulb of Held. In the mutant, the response was all-or-none, increasing in a single step. In all genotypes, at least one of the steps was >1 nA. There was no obvious difference in the amplitudes of steps between mutant, heterozygote and wild type mice (P17–19). Five out of seven mutant bushy cells likely received one input through an endbulb. (B) Recordings from individual control T stellate cells showed that the synaptic current grew in 8 or 9 small steps, but that in a T stellate cell recorded anteriorly in an Npr2 mutant, the response grew in only 4, larger steps. Small numbers of steps were recorded in 4 of 10 mutant T stellate cells, all of which lay anteriorly; the remaining 6 cells were located posteriorly. This difference is statistically significant (P<0.001, Student's t-test). (C) Recordings from individual octopus cells grew in more steps that were too numerous to count both in the wild type and in mutants. There was no discernible difference between them.

Figure 8. Npr2 mutants exhibit normal synaptic depression but abnormal intermittent synaptic failures.

(A) Trains of shocks delivered at 100 Hz evoked synaptic responses that tended to diminish in amplitude, confirming synaptic depression in both control and Npr2 mutant animals. However, some mutant bushy cells failed to respond to some of the shocks in the train. (A′) The small response (boxed) in the middle panel is shown at higher magnification to show that this cell had strong synaptic depression but that every shock evoked a synaptic current. 7 of 12 bushy cells showed synaptic failures such as those shown in the right panel. (A″) At higher magnification, no response was detected in the boxed region, indicating that this was a synaptic failure. (B) In T stellate cells, trains evoked responses with less synaptic depression than in bushy cells in the WT as well as in mutants. Failures in transmission were observed in mutant but not in WT animals. (C) Quantification of EPSC failures in control (WT and Het) and Npr2 mutant (Mut) bushy and T stellate cells in response to the first 10 shocks in a train. No failures were observed in WT or Het cells, whereas in Npr2 mutants, failures sometimes occurred, even in responses to the first shock. (D, E) Synaptic failures in Npr2 mutants were reversibly eliminated by 4-aminopyridine (4-AP). In a bushy cell (D) and in a T stellate cell (E), failure to evoke EPSCs was reversibly abolished by 0.1 mM 4-AP. Blocking K⁺ channels heightens and lengthens action potentials, making them less susceptible to conduction block.