Auditory hair cells and spiral ganglion neurons regenerate synapses with refined release properties in vitro

Abstract

Ribbon synapses between inner hair cells (IHCs) and type I spiral ganglion neurons (SGNs) in the inner ear are damaged by noise trauma and with aging, causing "synaptopathy" and hearing loss. Cocultures of neonatal denervated organs of Corti and newly introduced SGNs have been developed to find strategies for improving IHC synapse regeneration, but evidence of the physiological normality of regenerated synapses is missing. This study utilizes IHC optogenetic stimulation and SGN recordings, showing that, when P3-5 denervated organs of Corti are cocultured with SGNs, newly formed IHC/SGN synapses are indeed functional, exhibiting glutamatergic excitatory postsynaptic currents. When using older organs of Corti at P10-11, synaptic activity probed by deconvolution showed more mature release properties, closer to the specialized mode of IHC synaptic transmission crucial for coding the sound signal. This functional assessment of newly formed IHC synapses developed here, provides a powerful tool for testing approaches to improve synapse regeneration.

Keywords: hair cell ribbon synapses; optogenetics; postsynaptic currents; regeneration; type-I spiral ganglion neurons.

Fig. 1.

Cocultures of denervated organs of Corti and isolated SGNs for testing regenerated hair cell synaptic function. (A) Organs of Corti were dissected from Gfi1^Cre/+; Ai32^fl/+mice expressing Channelrhodopsin-2 (ChR2) in both IHCs and outer hair cells (OHCs) for light stimulation and from Gfi1^+/+; Ai32^fl/+mice, where hair cell stimulation was performed by the local perfusion of 40 mM K⁺. Organs of Corti were separated from the lateral wall (black dotted line) and denervated by cutting through SGN endings close to the IHCs (red dotted line). Denervated organs of Corti were plated at P3-5, P7-8, or P10-11. Fluorescent SGNs were isolated at P0-2 from one of three mouse lines used indiscriminately (Mapt-GFP^+/+; Bhlhb5^Cre/+; GCaMP6f^fl/+or Bhlhb5^Cre/+; Ai9^fl/+), and cocultured with denervated organs of Corti. At 10 to 12 DIV, recordings were performed from SGN somata while stimulating hair cells. Functional newly formed synapses were identified by EPSCs in SGNs in response to IHC stimulation. (B) Confocal image of a denervated P_3-5 Gfi1^Cre/+; Ai32^fl/+ organ of Corti. Inset: ChR2-eYFP is observed in IHC and OHC membranes. (C) In current clamp (I_holding = 0 pA), the IHC membrane potential is recorded in response to two 1 s blue light pulses separated by a 4 s interval (blue line). Resting membrane potential indicated at the trace. (D) Same protocol as in C. IHC recording in a native P_4-5 Gfi1^Cre/+; Ai32^fl/+ acutely isolated organ of Corti. Light stimulation induces IHC depolarization and superimposed Ca²⁺ APs which are not found in cultured IHCs like in C. EPSPs (asterisks) and Ca²⁺ APs (black arrow) also occur spontaneously. (E) Confocal image of a cultured BhlhB5^Cre/+; GCaMP6f^fl/+ positive SGN with soma and projecting fiber. (F) SGNs with diverse electrical response properties in culture; “Slowly Adapting”: APs slightly decrease in size throughout the pulse; “Intermediately Adapting”: multiple APs only at the beginning of the pulse; “Rapidly Adapting”: single AP at the pulse onset. Current clamp recordings of SGNs; 100 ms-long current step protocols from resting membrane potential as indicated at trace, with initial 5 mV step and subsequent 10 mV increasing steps.

Fig. 2.

New contacts between hair cells and SGNs appear in coculture and express AMPA receptors. (A–C) Representative examples of live cocultures at DIV10-11 showing Bhlhb5^Cre/+; Ai9^fl/+, td-Tomato expressing SGN endings (red) close to ChR2⁺ hair cells (green). In A and B, several IHCs are contacted by multiple endings from a single SGN. Arrowheads in A and C point to a SGN soma. In C, SGN fiber travels along the row of IHCs. The arrow points to a SGN projection reaching the OHC region. (D) Confocal maximum-intensity projection image from an immuno-labeled coculture at DIV 12. The Left Insets show individual labels and overlay at a single synapse, with presynaptic ribbon (Anti-CtBP2; red) juxtaposed with postsynaptic AMPA receptors (Anti-PAN GluA1-4; blue). Anti-Myosin VI (orange) labels hair cells and Anti-NF200 (white) labels nerve fibers.

Fig. 3.

Hair cell stimulation activates glutamatergic synaptic currents with various waveforms in newly formed synapses. (A) Superimposed difference interference contrast (DIC) and ChR2-YFP confocal image of a native acutely excised P_4-5 Gfi1^Cre/+; Ai32^fl/+organ of Corti. ChR2⁺ hair cells in green. SG region with SGN somata is located between white dashed lines; recording patch pipette is highlighted by red dashed lines. (A1) Trace of the SGN recording shown in A from a native P_4-5 synapse, in response to two 1 s long light pulses (blue lines). Holding potential −79 mV. A flurry of fast EPSCs is observed in response to each light pulse (Top trace, Control) that are blocked by the combined perfusion of NBQX and (Rs)-CPP (20 µM each), an AMPA/kainate and NMDA receptor blocker, respectively (Bottom trace). (B) A superimposed DIC and ChR2-YFP confocal image of a live coculture at DIV 10. Row of ChR2⁺ IHCs in yellow. Here, SGNs express the Ca²⁺ indicator GCaMP6_f (green) under the control of the BHLHB5 Cre promoter. The Inset shows a drawing of the experimental setting. (C–E) Trace examples of SGN recordings of newly formed synapses in coculture using P_3-5 denervated organs of Corti, in response to two 1 s long light pulses (blue lines). Holding potential −79 mV. (C1–E1) show extended traces of (C–E); their extent indicated in red in (C–E). Blue arrowheads in (C1–E2) indicate the beginning of the light pulse. Two individual EPSCs are shown at further extended time scale in (C2). Three response types with different waveforms were found: (C), fast; (D), slow and (E), mixed, the last having both fast and slow responses. Inward currents of all response types were blocked by glutamate receptor blockers; here in the examples (C1 and D1) completely by NBQX (20 µM), in E1 mostly by NBQX, and in (E2) completely by NBQX and (Rs)-CPP (20 µM).

Fig. 4.

Age of the plated organs of Corti affects the maturation state of newly formed synapses. (A–E) Properties of SGN recordings are reported for newly formed synapses in cocultures of SGNs and denervated organs of Corti, plated at ages P3-5, P7-8, or P10-11; and for native acutely excised P_4-5 organs of Corti (P_4-5 Native, black). A, Left, EPSC waveform of a P_4-5 Native synapse (black) is slower that EPSC waveform from a P_3-5 newly formed synapse (pink; P_3-5 Newly formed). Individual waveforms are shown in gray and averaged in black or pink. Holding potential −79 mV. Right, EPSC analysis (10 to 90% rise time, 90 to 10% decay time, amplitude, and area) is described for a monophasic (Top) and a multiphasic event (Bottom). Note that for multiphasic events, only the biggest peak was considered to calculate rise time and amplitude. Decay time is always calculated from the last peak. (B–E) Comparison of EPSC waveform parameters for newly formed and native synapses. Each data point represents the median value from an individual recording. Numbers above each box indicated the number of SGN recordings used for analysis. Boxes represent the median (horizontal line) and 10th and 90th percentiles. Whiskers represent maximum and minimum values of the distribution. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant. (B) The 10 to 90% rise time of native EPSCs (EPSCs^{(P4-5 Native)}; 1.21 ms; n = 7) was significantly slower compared to EPSCs from newly formed synapses at all age conditions (one-way ANOVA with Tukey’s post hoc test): EPSCs^{(P3-5 New)} (0.42 ms; n = 17; P < 0.001); EPSCs^{(P7-8 New)} (0.33 ms; n = 8; P < 0.001) and EPSCs^{(P10-11 New)} (0.29 ms, n = 6; P < 0.0001). However, rise time was not statistically different between newly formed EPSCs at all age conditions. (C) 90 to 10% median decay time of native EPSCs (EPSCs^{(P4-5 Native)}; 1.81 ms; n = 7) were slower compared to EPSCs from newly formed synapses of all age conditions (Kruskal–Wallis with Dunn’s post hoc test): EPSCs^{(P3-5 New)} (0.66 ms; n = 17; P < 0.01); EPSCs^{(P7-8 New)} (0.60 ms; n = 8; P < 0.01) and EPSCs^{(P10-11 New)} (0.58 ms, n = 6; P < 0.01). However, decay time was not statistically different between regenerated EPSCs of all age conditions. (D) The median amplitude of native EPSCs (EPSCs^{(P4-5 Native)}: 15.59 pA; n = 7) was similar compared to EPSCs^{(P3-5 New)} (18.66 pA; n = 17). However, EPSCs^{(P4-5 Native)} were significantly smaller when compared to EPSCs^{(P7-8 New)} (21.89 pA; n = 8; P < 0.01) and to EPSCs^{(P10-11 New)} (25.77 pA; n = 6; P < 0.05) (Kruskal–Wallis with Dunn’s post hoc test). (E) The median area of native EPSCs (EPSCs^{(P4-5 Native)}; 46.00 fC; n = 7) was similar compared to regenerated EPSCs at all age conditions (one-way ANOVA with Tukey’s post hoc test): EPSCs^{(P3-5 New)} (19.23 fC; n = 17, P = 0.12), EPSCs^{(P7-8 New)} (30.25 fC; n = 8, P = 0.99) and EPSCs^{(P10-11 New)} (24.00 fC; n = 6, P = 0.41). EPSCs from newly formed synapses also displayed similar area values at all age conditions.

Fig. 5.

EPSC waveforms at newly formed synapses with older IHCs reveal properties closer to mature IHC ribbon synapses. (A1–A4) Modeling of EPSC waveforms using deconvolution. Four EPSCs (blue traces) from a single SGN recording of a newly formed synapse are displayed. Holding potential: –79 mV. P_10-11 denervated organ of Corti was plated for this coculture with P_0-2 SGNs. (A1) Shows a monophasic EPSC. Monophasic EPSC like this example, were averaged to create a kernel (standardized release event) for individual recording. Kernels (amplitude and time-of-peak) are depicted in green dashed line above each EPSC. The scales for the green schemata are the same as the scales for the EPSC currents. The fits calculated from the event sequences are shown in orange. EPSC amplitude (EPSC Amp.) and EPSC area (gray filled area) are defined in (A2). Horizontal black lines represent EPSC extent and numbers indicate the smallest number of events (i.e., kernels) that best fit this EPSC. (B–D) Mean values of newly formed EPSC amplitude are plotted against the number of events per EPSC, for the three coculture conditions, P_3-5 (B, n = 14 synapses, pink), P_7-8 (C, n = 8 synapses, red), and P_10-11 (D, n = 6 synapses, purple) denervated organ of Corti. Gray thin lines represent individual recordings and bold colored lines represent the averages. Black dotted lines represent the fit of the data, including only data with 1 to 8 events per EPSC. (E) EPSC amplitude versus events/EPSC plots were normalized to their minimum value and superimposed for different conditions. These include average traces for newly formed synapses from (B–D), (P_3-5 New, P_7-8 New, P_10-11 New; pink, red, and purple), for immature P_4-5 Native synapses (black dashed lines) and mature ribbon synapses (blue dotted lines, data from ref. 50). For EPSCs^{(P4-5 Native)}, EPSCs from the seven recordings were pooled. (F–I) Same as (B–E), but with EPSC area plotted against number of events per EPSC. (J) Slopes of the EPSC amplitude versus the number of events per EPSC calculated from (B–E), are shown for each condition. EPSCs^{(P4-5 Native)}: 1.79pA/no. event (n = 7 pooled cells); EPSCs^{(P3-5 New)}: −0.32pA/no. events (n = 16); EPSCs^{(P7-8 New)}: −1.44 pA/no. event (n = 8); EPSCs^{(P10-11 New)}: −3.26 pA/no. event (n = 6) and EPSCs^{(mature native)}: −35.76 pA/no. event (n = 8; data from ref. 50). One-way ANOVA with Tukey’s post hoc test. (K) Slopes of the EPSC area versus the number of events per EPSC calculated from (F–I), are shown for each condition. EPSCs^{(Native P4-5)}: 8.00fC/no. event (n = 7 pooled cells); EPSCs^{(P3-5 New)}: 3.82 fC/no. event (n = 16); EPSCs^{(P7-8 New)}: 4.01 fC/no. event (n = 8); EPSCs^{(P10-11 New)}: 1.7 fC/no. event (n = 6) and EPSCs^{(mature native)}: −4.32 fC/no. event (n = 8; data from ref. 50). One-way ANOVA with Tukey’s post hoc test. (L) Percentage of monophasic EPSCs per recording is shown for each condition. EPSCs^{(Native P4-5)}: 10% (n = 7); EPSCs^{(P3-5 New)}: 36.70% (n = 16); EPSCs^{(P7-8 New)}: 34.04% (n = 8); EPSCs^{(P10-11 New)}: 35.64% (n = 6) and EPSCs^{(mature native)}: 60.50% (n = 8; data from ref. 50). Kruskal–Wallis with Dunn’s post hoc test. (J–L) Each data point represents an individual recording. Number of SGN recordings used for analysis is indicated. Boxes represent the median (horizontal line) and 10th and 90th percentiles. Whiskers represent maximum and minimum values of the distribution. *P < 0.05, **P <0.01, ***P < 0.001, “n.s.”: not significant.