Retrograde facilitation of efferent synapses on cochlear hair cells

Abstract

Cochlear inner hair cells (IHCs) are temporarily innervated by efferent cholinergic fibers prior to the onset of hearing. During low-frequency firing, these efferent synapses have a relatively low probability of transmitter release but facilitate strongly with repetitive stimulation. A retrograde signal from the hair cell to the efferent terminal contributes to this facilitation. When IHCs were treated with the ryanodine receptor agonist, cyclic adenosine phosphoribose (cADPR), release probability of the efferent terminal rose. This effect was quantified by computing the quantum content from a train of 100 suprathreshold stimuli to the efferent fibers. Quantum content was sevenfold higher when IHCs were treated with 100 μM cADPR (applied in the recording pipette). Since cADPR is membrane impermeant, this result implies that an extracellular messenger travels from the hair cell to the efferent terminal. cADPR is presumed to generate this messenger by increasing cytoplasmic calcium. Consistent with this presumption, voltage-gated calcium flux into the IHC also caused retrograde facilitation of efferent transmission. Retrograde facilitation was observed in IHCs of a vesicular glutamate transporter (VGlut3) null mouse and for wild-type rat hair cells subject to wide-spectrum glutamate receptor blockade, demonstrating that glutamate was unlikely to be the extracellular messenger. Rather, bath application of nitric oxide (NO) donors caused an increase in potassium-evoked efferent transmitter release while the NO scavenger carboxy-PTIO was able to prevent retrograde facilitation produced by cADPR or IHC depolarization. Thus, hair cell activity can drive retrograde facilitation of efferent input via calcium-dependent production of NO.

FIG. 1

Electrical stimulation on efferent axons. A The electrical stimulation pipette was positioned ~10–20 mm below the base of the IHC. The aim was to stimulate cholinergic medial olivocochlear efferents that contact IHCs transiently during development. B Electrical stimulation (50 μA, 500–800 μs in duration) was at 1 Hz with 100 shocks for each determination of quantum content.

FIG. 2

Effect of cADPR on ACh-evoked SK current in rat inner hair cells at V_h = −44 mV. A Representative trace of IPSCs evoked by electrical stimulation of efferent axons. Closed arrows indicate stimulation. Some shocks evoked an IPSC and some failed. Other events occurred spontaneously. B Representative trace of IPSCs with 100 μM cADPR in the recording pipette. Spontaneous and evoked responses were more frequent. C Mean evoked IPSC amplitude including failures was 23.47 ± 2.83 pA (SE) in control IHCs (371 events in three IHCs) and was 70.64 ± 5.68 pA for those treated with cADPR (523 events in three IHCs) (t(2,892) = 6.36, p < 0.0001). D Mean evoked IPSC decay time was 64.84 ± 3.1 ms in control (ct) IHCs and 93.03 ± 8.90 ms with cADPR (t(2,892) = 2.48, p = 0.013).

FIG. 3

Inward currents through the nAChR of rat inner hair cells evoked at −82 mV (E_K). A Representative traces of inward current through the hair cell nAChR evoked by electrical stimulation of efferent axons. SK currents were minimized by recording at the potassium equilibrium potential (−82 mV). B Representative traces of IPSCs with 100 μM cADPR in the recording pipette. Arrows indicate the shock artifact. Note that every stimulus evoked a response, and more spontaneous events were present in the cADPR-treated hair cell.

FIG. 4

Quantal analysis of cADPR effect in rat inner hair cells. A Amplitude distribution of electrically evoked inward currents (eIPSCs) through the hair cell nAChR from one control IHC (bin = 3 pA). Solid red line is the Poisson fit. The dashed blue lines represent the normal distributions of release events including one, two, or three vesicles. Calculated value for quantum content, m, was 0.48 (n = 252 evoked IPSCs). Inset: amplitude distribution for spontaneous events (sIPSCs) in the same IHC. Solid red line represents Gaussian fit, mean and SE were 10.33 ± 0.33 pA (SE, n = 154), corresponding to the quantal size for the evoked histogram. Quantum content was derived from the direct method (evoked/spontaneous event amplitude), by the fraction of failures (B) or by the coefficient of variation (C). By any of these measures, quantum content was consistently larger in cells treated with cADPR.

FIG. 5

Quantal size and quantum content (from fraction of failures) in control rat IHCs (gray) and IHCs treated with cADPR (black). A There was no difference in quantal size from four control cells (8.72 ± 0.57 pA, SE) compared to five cells with cADPR (9.57 ± 0.74 pA (t(2,7) = 0.87, p = 0.41). B Quantum content, m was more than fourfold larger in cells treated with calcium store release activators (m = 0.22 ± 0.02 for 17 control cells, 2.09 ± 0.33 for eight cells with 100 μM cADPR (t(2, 23) = 8.32, p < 0.0001), 1.41 ± 0.19 for five cells with 1 μM Ry, t(2, 20) = 14.01, p < 0.0001 compared to control).

FIG. 6

Depolarization of rat IHCs produced retrograde facilitation. A The control protocol included 1 Hz efferent shock trains, recording at −82 mV. In the experimental protocol, each efferent shock was preceded by a 750-ms depolarization of the IHC to −20 mV (ending 150 ms prior to the shock). B Diary plot of efferent quantum content (from fraction of failures). After 40 min of repeated 1 Hz of shock trains (every 5 min) to establish baseline quantum content, intervening depolarizing steps were included, producing a marked increase in quantum content. C For 18 control measurements and ten with interleaved depolarization in three IHCs, this manipulation produced a significant increase in quantum content (m = 0.38 ± 0.09 control, m = 1.13 ± 0.02 after depolarization, (t(2,26) = 6.78, p < 0.001).

FIG. 7

Enhancement of efferent transmission by cADPR in IHCs of the VGlut3 KO mouse. A cADPR increased efferent quantum content, m (from fraction of failures) onto IHCs of VGlut3 KO mice that fail to release vesicular glutamate (m = 0.68 ± 0.20 from four control cells and 3.29 ± 0.51 for three cells treated with cADPR, t(2,5) = 4.43, p < 0.007. B Quantal size was not significantly different with cADPR treatment in the VGlut3 KO OHCs (9.24 ± 0.33 pA and 11.80 ± 0.95 pA (t(2,5) = 1.77, p = 0.09).

FIG. 8

NO donors increased spontaneous transmitter release from efferent terminals onto young rat inner hair cells. Synaptic currents were recorded in IHCs during potassium (40 mM) depolarization and exposure to 250 μM SIN-1 (shaded region of each diary plot). A Several minutes after donor exposure began, spontaneous synaptic currents became larger. Event amplitude rose from 11.9 ± 0.49 pA to 27.6 ± 0.32 pA (t(2, 1,597) = 11.65, p < 0.0001). B Instantaneous frequency rose from 1.53 ± 0.22 Hz to 7.68 ± 0.22 Hz (t(2, 1,597) = 6.76, p < 0.0001).

FIG. 9

NO scavenger carboxy-PTIO prevents or reverses retrograde facilitation in rat IHCs treated with cADPR. A Diary plot of quantum content (from fraction of failures) determined from repeated 1-Hz shock trains to the efferent axons. Carboxy-PTIO added during the time indicated by gray-shaded regions. Upon removal of carboxy-PTIO from the bath, efferent quantum rose. Quantum content fell again with the return of carboxy-PTIO. B In another IHC, repeated presentations of carboxy-PTIO prevented or reversed activity-dependent retrograde facilitation of efferent quantum content. C In another IHC with initially high probability of efferent transmission, carboxy-PTIO reduced efferent quantum content, which recovered after removal of carboxy-PTIO. D Efferent quantum content was suppressed by carboxy-PTIO m = 0.42 ± 0.08 at 1 Hz (n = 6) and rose again after removal, m = 1.52 ± 0.22 (t(2,10) = 4.69, p = 0.0009) (n = 6). Control data (light gray bars) repeated from Figure 5B for direct comparison. Quantum content for cells treated with cADPR and carboxy-PTIO was significantly larger than that of control cells from Figure 5B (t(2,21) = 3.45, p = 0.002).

FIG. 10

NO scavenger carboxy-PTIO prevented depolarization-evoked retrograde facilitation in rat inner hair cells. A Diary plot of efferent quantum content (from fraction of failures) derived from repeated efferent shock trains (1 Hz, 100 shocks) that incorporated a 750-ms depolarization to −20 mV 150 ms prior to each shock. Quantum content rose gradually after removal of carboxy-PTIO. B, C, same protocol repeated in two other IHCs. D Carboxy-PTIO (indicated by gray-shaded regions) completely prevented retrograde facilitation normally produced by coupling efferent stimulation with hair cell depolarization (m = 0.20 ± 0.05 for six IHCs subjected to the depolarization protocol in the presence of carboxy-PTIO. With removal of carboxy-PTIO, the efferent quantum content increased significantly to 0.94 ± 0.06 (t(2,10) = 5.49, p = 0.0003). Quantum content in carboxy-PTIO was not significantly different from that of control data from Figure 5B (t(2,22) = 0.51, p = 0.62).