Immunostaining using the neuronal marker MAP2 and the oligodendroglial marker, Olig1, in brainstem sections from wild-type (WT) mice showed that BDNF is highly expressed in MNTB principal neurons (Figure 1A). It is of note that oligodendrocytes located close to the calyx–MNTB neuron synapse also expressed BDNF. BDNF expression was notably decreased in all cell types in the MNTB in Bdnf^+/−mice at P21 (Figure 1A). To examine the effect of endogenous BDNF on fast glutamatergic transmission in the auditory brainstem, we recorded miniature excitatory post-synaptic currents (mEPSCs) from MNTB principal neurons in brainstem slices from P16–20 WT and Bdnf^+/− mice (Figure 1B). There was no significant difference in the amplitude or kinetics, including rise and decay times, of mEPSCs in WT and Bdnf^+/– mice (amplitude: 39.9 ± 2.51 pA, n = 13 vs 33.4 ± 2.92 pA, n = 12, respectively, p=0.10; rise time: 0.3 ± 0.01 ms, n = 13 vs 0.3 ± 0.02 ms, n = 12, respectively, p=0.37; decay time: 0.6 ± 0.03 ms, n = 13 vs 0.7 ± 0.07 ms, n = 12, respectively, p=0.58, unpaired t-test; Figure 1C–E). In addition, the frequency of mEPSCs was not statistically different (2.7 ± 0.69 Hz, n = 8 vs 2.4 ± 0.41 Hz, n = 11 in WT and Bdnf^+/− mice, respectively; p=0.76, unpaired t-test; Figure 1F). However, the amplitude of evoked EPSCs (eEPSCs) triggered by afferent fiber stimulation was significantly smaller in Bdnf^+/− mice (3.1 ± 0.31 nA, n = 17 in Bdnf^+/− mice vs 5.9 ± 0.35 nA, n = 9 in WT; p<0.0001, unpaired t-test; Figure 1G,H). To examine the changes in postsynaptic receptor kinetics or in asynchronous release, we analyzed the decay of eEPSCs. The line corresponding to eEPSC decay was well fit as a single exponential with a time constant tau (τ)=1.0 ± 0.07 ms (n = 9) in WT and 0.9 ± 0.05 ms (n = 17) in Bdnf^+/− mice, which were not significantly different (p=0.31, unpaired t-test; Figure 1G,I). These results suggest that a reduction in endogenous BDNF results in impaired glutamatergic transmission, which is caused by alterations in presynaptic properties rather than postsynaptic components.

Figure 1

Download asset Open asset

To test the effect of reduced BDNF on presynaptic properties, we examined the paired pulse ratio (PPR), which was similar in both groups (0.8 ± 0.02, n = 9 in WT and 0.8 ± 0.03, n = 17 in Bdnf^+/− mice; p=0.93, unpaired t-test; Figure 2A). This indicates that a reduction in endogenous BDNF does not alter the Ca²⁺-dependent release probability at presynaptic terminals. Next, we examined the short-term depression and the RRP size of available glutamate vesicles at presynaptic terminals in WT and Bdnf^+/− mice. During a train of stimuli at 100 Hz (50 pulses), the amplitude of eEPSCs displayed strong depression, falling to ~20% of the initial amplitude near the end of the train in both WT and Bdnf^+/− mice (Figure 2B). There was no notable difference in short-term depression between WT and Bdnf^+/− mice (n = 7 vs 12). To predict the RRP size, the release probability (P_r), and the synaptic vesicle replenishment rate, we used two variants of the cumulative analysis of EPSC trains (Figure 2C,D). In the Elmqvist and Quastel (EQ) method (Elmqvist and Quastel, 1965), the RRP size was estimated by fitting a line to the linear portion of these data (corresponding to the second through the fourth EPSC) and extrapolating to the x axis, we measured the total equivalent EPSC at the beginning of the train, which indirectly indicates the available pool of vesicles for release (Taschenberger et al., 2002; Kushmerick et al., 2006). We plotted the eEPSC amplitudes during a train versus their cumulative amplitudes at the end of a train (Figure 2C), which were 91.5 ± 9.46 nA (n = 6) in WT and 46.3 ± 6.89 nA (n = 13) in Bdnf^+/− mice (p=0.0016, unpaired t-test; Figure 2D). The forward extrapolation linear fits revealed that calyces in Bdnf^+/− mice had a much smaller RRP of glutamate vesicles as compared with WT mice (17.6 ± 2.73 nA in Bdnf^+/− mice, n = 11 vs 33.4 ± 3.31 nA in WT, n = 5; p=0.004, unpaired t-test; Figure 2C). The RRP divided by the mEPSC amplitude (Figure 1C) approximately estimates the number of vesicles, which was reduced in in Bdnf^+/− mice (~837 vesicles in WT and ~527 vesicles in Bdnf^+/− mice). There was no significant difference in Pr, determined as the slop of the linear fit (0.25 ± 0.035 in Bdnf^+/− mice, n = 10 vs 0.26 ± 0.037 in WT, n = 6; p=0.8582, unpaired t-test; Figure 2E).

Figure 2

Download asset Open asset

In the Schneggenburger-Meyer-Neher (SMN) method (Schneggenburger et al., 1999), EPSC amplitudes from trains are plotted cumulatively against the stimulus number (Figure 2F). A line fit to the steady-state points (the last 10 of 50 points) is back-extrapolated to the y-axis, and the y-intercept divided by the mEPSC amplitude estimates the RRP size. This analysis also revealed that calyces in Bdnf^+/− mice had a much smaller RRP of glutamate vesicles as compared with WT mice (9.8 ± 1.29 nA in Bdnf^+/− mice, n = 11 vs 19.4 ± 1.71 nA in WT, n = 5; p=0.0008, unpaired t-test; Figure 2F). Conversely, the release probability (P_r), which is calculated by dividing the amplitude of the first eEPSC by the RRP size, was not different in WT and Bdnf^+/– mice (0.35 ± 0.02, n = 5 vs 0.38 ± 0.05, n = 11, respectively, p=0.73, unpaired t-test; Figure 2G). Another interesting finding was the reduced replenishment rate of vesicles in Bdnf^+/− mice, which was estimated by the slope of the linear fit (0.61 ± 0.10 in Bdnf^+/− mice, n = 11 vs 1.39 ± 0.19 in WT, n = 5; p=0.0018, unpaired t-test; Figure 2F), indicating that BDNF signaling plays a role in the replenishment of vesicles at the calyx terminal. The values of RRP obtained by the SMN method were smaller than those obtained by the EQ method, because this RRP was measured as the pool decrement during stimulation, whereas the RRP estimated using the EQ Method indicates the size of a pre-existing pool of vesicles (Neher, 2015). Taken together, these results show that a reduction in endogenous BDNF decreased the pool of glutamate vesicles available for release at the beginning of a train in Bdnf^+/− mice, suggesting that BDNF is important for determining the RRP at presynaptic terminals.

An increase in the RRP of vesicles can contribute to short-term plasticity such as post-tetanic potentiation (PTP; Habets and Borst, 2005; Regehr, 2012). Thus, we tested whether a decrease in the RRP caused by reduction of BDNF alters PTP at the calyx synapse (Figure 2H,I). Tetanic stimulation (100 Hz, 3 s) increased the amplitude of eEPSCs from 5.6 ± 1.01 pA to 8.0 ± 0.96 pA in WT (n = 5; p=0.02, paired t-test), indicating a PTP induction, whereas there was no significant increase in the eEPSC amplitude after tetanus in Bdnf^+/− mice (2.8 ± 0.37 pA to 2.7 ± 0.39 pA, n = 9; p=0.05, paired t-test). Taken together, BDNF controls synaptic plasticity as well as neurotransmitter release by regulating the RRP at the calyx terminal.

We next evaluated the regulatory effect of endogenous BDNF on the exocytosis of vesicular neurotransmitter and Ca²⁺ influx at presynaptic terminals by measuring the membrane capacitance jump (ΔC_m) and voltage-activated Ca²⁺ channel current (I_Ca) in P9–13 calyces from WT and Bdnf^+/− mice (Figure 3A). Depolarization induced Ca²⁺ currents and, consequently, ΔC_m in a pulse duration–dependent manner (2 to 20 ms). ΔC_m in Bdnf^+/− mice was much smaller than that in WT (ΔC_m after a 20 ms depolarization: 93.4 ± 8.82 fF, n = 11 in Bdnf^+/− vs 148 ± 11.32 fF, n = 11 in WT mice; p=0.0013, unpaired t-test; Figure 3A,B). However, the membrane capacitances of calyces from WT and Bdnf^+/− mice were similar (20 ± 1.5 pF, n = 10 vs 21.5 ± 1.35 pF, n = 9, respectively, p=0.43, unpaired t-test; Figure 3C), indicating that the sizes of the calyx terminals are similar, and thus the difference in ΔC_m was not due to the terminal size. In addition, plotting ΔC_m (for the 20 ms depolarization) as a function of the C_m showed the distribution of C_m as estimated by linear regression analysis (R² = 0.1 in WT, n = 9; R² = 0.06 in Bdnf^+/– mice, n = 9; Figure 3C). Despite their similar resting values for C_m, ΔC_m was much smaller in calyces from Bdnf^+/− mice as compared with comparably sized calyces from WT. However, a positive correlation between the resting C_m and ΔC_m suggest that larger calyces release more vesicles in both genotypes (Figure 3C). Next, we examined the effects of BDNF reduction on presynaptic Ca²⁺ channel and I_Ca related to changes in exocytosis of glutamate vesicles. There was no significant difference in the presynaptic I_Ca charge (QI_Ca) in response to depolarizing pulses. The smaller ΔC_m in calyces from Bdnf^+/− mice was not attributed to alterations in presynaptic Ca²⁺ currents. In WT calyces, a 20 ms depolarization induced a I_Ca of 6.5 ± 0.53 pC/pF (n = 10), whereas in Bdnf^+/− calyces, I_Ca was 5.9 ± 0.96 pC/pF (n = 9; p=0.50, unpaired t-test; Figure 3D). In addition, the current–voltage relationship (I–V) curve for these voltage-activated Ca²⁺ channels at the calyx terminal in WT and Bdnf^+/− mice (at P10-12) exhibited a similar pattern with the peak current of −625 ± 48.3 pA vs −614 ± 16.5 pA at −10 mV in WT and Bdnf^+/− mice, respectively (n = 15 vs n = 15; Figure 3E,F). Furthermore, we examined presynaptic action potential evoked by afferent fiber stimulation, and there was no significant difference in amplitude and half-with of presynaptic action potential, which is associated with Ca²⁺ channel activation and release probability (amplitude, 122.5 ± 7.8 mV in WT, n = 3 vs 124.6 ± 4.4 mV in Bdnf^+/−, n = 4, p=0.8082 and half-width, 318 ± 41.2 μs in WT, n = 3 vs 283 ± 45.5 μs in Bdnf^+/−, n = 4, p=0.6123, unpaired t-test, data not shown). These data suggest that a reduction in endogenous BDNF decreases exocytosis of vesicular glutamate, but this reduction is not associated with changes in Ca²⁺ influx via Ca²⁺ channels at the presynaptic terminals.

Figure 3 with 1 supplement see all

Download asset Open asset

To confirm whether BDNF regulates the RRP and exocytosis of glutamate vesicles at presynaptic terminals, we assessed the vesicle pool size by gradually depleting the RRP using 20 depolarizing pulses (10 ms, 10 Hz). The accumulated capacitance jump (ΣΔC_m), which is a measure of the sum of available glutamate vesicles released by stimulation, was significantly smaller in Bdnf^+/− mice (260 ± 54.1 fF, n = 9) relative to WT (508 ± 86.2 fF, n = 11; p=0.04, t-test; Figure 3G–I). This difference, however, was not associated with the endocytosis rate, which was similar in both groups (19.5 ± 2.75 s, n = 10 in WT vs 17.3 ± 1.91 s, n = 9 in Bdnf^+/− mice; p=0.51, unpaired t-test; Figure 3—figure supplement 1). Therefore, the decreased ΣΔC_m resulted mainly from the reduction in the RRP size in Bdnf^+/– mice. Taken together, these results suggest that the endogenous BDNF level is directly involved in glutamatergic transmission based on its ability to determine the RRP size at the presynaptic terminal.

We examined whether BDNF function in determining the RRP is mediated by endogenous BDNF at the terminal or by BDNF derived from neighboring cells, which activate TrkB signaling at presynaptic terminals. TrkB was expressed at the calyx terminal and axon, which was immunolabelled with VGluT1 and detected by Alexa 568 dye filling during presynaptic whole-cell recording (Figure 4A). We directly activated TrkB using its agonist, 7,8-dihydroxyflavone (7,8-DHF), which binds to the TrkB extracellular domain and activates TrkB-mediated downstream signaling (Jang et al., 2010; Marongiu et al., 2013), to determine whether this activation rescues the impaired RRP size and glutamate release at the calyx synapses in Bdnf^+/– mice. In WT mice, the pre-application of 20 μM 7,8-DHF to brainstem slices (30 min) had no effect on the amplitude (6.6 ± 0.53 nA for control vs 6.3 ± 0.70 nA for 7,8-DHF, n = 5, p=0.70, unpaired t-test) or PPR (0.75 ± 0.03 for control vs 0.74 ± 0.01 for 7,8-DHF, n = 5, p=0.86, unpaired t-test) of eEPSCs (Figure 4B,C). In Bdnf^+/− mice, TrkB activation using 7,8-DHF significantly increased the amplitude of eEPSCs (from 2.8 ± 0.54 nA, n = 8 to 4.5 ± 0.40 nA, n = 17, p=0.02, unpaired t-test) without changing the PPR (0.7 ± 0.05 for control, n = 8 vs 0.8 ± 0.03 for 7,8-DHF, n = 17, p=0.58, unpaired t-test; Figure 4B,C). Acute application of 7,8-DHF also increased the amplitude of eEPSCs during the 10 min after its application in Bdnf^+/− mice in a time-dependent manner (2.8 ± 0.97 nA for control, n = 5 vs 3.9 ± 1.14 nA for 7,8-DHF at 10 min vs, n = 5, p=0.023, paired t-test; Figure 4D,E). In addition, the pre-application of 7,8-DHF significantly increased the cumulative eEPSC size (from 45.9 ± 7.9 nA, n = 11 to 75.2 ± 9.7 nA, n = 6; p=0.03, unpaired t-test) and partially restored the RRP in Bdnf^+/– mice (from 17.6 ± 2.73 nA, n = 11 to 27.8 ± 3.64 nA, n = 7; p=0.03; unpaired t-test, Figure 4F,G). These findings suggest that the down-regulation of BDNF–TrkB signaling at the presynaptic terminal impairs the RRP size and glutamatergic transmission in the MNTB.

Figure 4

Download asset Open asset

Our results suggested that local BDNF signaling from neighboring cells around presynaptic terminals is critical for determining the RRP at presynaptic terminals during postnatal development. We investigated whether glial cells, specifically oligodendrocytes, are the source of this BDNF signaling. A number of oligodendrocytes are apposed to the calyx synapse in the MNTB during postnatal development (Figures 1A and 5A; Berret et al., 2017). To study the specific role of oligodendrocytes in presynaptic functions as BDNF providers, we generated Bdnf cKO mice, in which BDNF was specifically deleted in CNPase-expressing oligodendrocytes using the Cre/loxP system (Figure 5B). To confirm the specificity of the Cnp^cre line, Cnp^cre mice were crossed to a GCaMP6f-GFP mouse (or tdTomato reporter) as a reporter line. GFP+ cells were positive for Olig2 and were present next to the calyx synapse, but did not express MAP2, NeuN, and GFAP expression in the MNTB (Figure 5—figure supplement 1). This confirms that Cnp^cre is specific to oligodendrocytes and is not expressed in neurons or astrocytes in the MNTB of the auditory brainstem. In addition, CNP+ cells were positive for CC1, but negative for PDGFRa, indicating that most CNP+ cells in the MNTB are pre-myelinating oligodendrocytes beyond the precursor stage (Figure 5—figure supplement 1). Bdnf cKO mice (Cnp^cre:Bdnf^fl/fl; Figure 5B) were generated by crossing Cnp^cre mice with mice containing a floxed allele of BDNF (Bdnf^fl/fl). To further confirm the oligodendrocyte-specific depletion of BDNF, oligodendrocytes were isolated via the fluorescent activated cell sorting (FACS) using an O1 antibody, which is specific to oligodendrocytes, or Cnp- driven GCaMP6f-GFP (GFP). Utilizing quantitative PCR, we confirmed that the sorted O1+ or GFP+ fraction expressed a substantial level of Bdnf in control mice (Cnp^cre:Bdnf^fl/+), whereas the O1+ or GFP+ fraction from Bdnf cKO mice showed significantly reduced level of Bdnf (Figure 5—figure supplement 2). Using presynaptic terminal recordings, we compared Bdnf cKO mice with control mice to examine how oligodendroglial BDNF affects presynaptic properties (Figure 5C).

Figure 5 with 2 supplements see all

Download asset Open asset

The deletion of BDNF in oligodendrocytes significantly decreased exocytosis of glutamate vesicles at the calyx terminal in brainstem slices from Bdnf cKO mice. In P9–12 Bdnf cKO mice, ΔC_m in response to a 2-, 3-, 5-, 10-, 20-, or 40 ms depolarization was much smaller than in control mice (for 20 ms, 123 ± 20 fF for Bdnf cKO, n = 19 vs 266 ± 35.4 fF for the control, n = 20; p=0.0016, unpaired t-test, Figure 5D,E). Longer depolarization induced a larger ΔCm and 40 ms- pulse exhibited saturation of ΔC_m in both control and cKO calyces. ΔC_m resulting from 2 ms depolarization was difficult to resolve in ~50% of Bdnf cKO calyces. In both genotypes, the C_m (15 ± 1.9 pF, n = 14 for control vs 15 ± 3.5 pF, n = 17 for Bdnf cKO; Figure 5F) and the Ca²⁺ influx during depolarizing pulses were similar (the QI_Ca for 20 ms was 5.1 ± 0.45 pC/pF, n = 13 for the control vs 5.5 ± 0.39 pC/pF, n = 11 for the Bdnf cKO; p=0.45, unpaired t-test; Figure 5G), indicating that the loss of oligodendroglial BDNF impairs vesicular exocytosis without a change in Ca²⁺ channel activation, similar to what was observed in Bdnf^+/– mice. In addition, the current–voltage relationship (I–V) curve for these voltage-activated Ca²⁺ channels at the calyx terminal in control and Bdnf cKO mice (at P10-12) exhibited a similar pattern with the peak current of −633 ± 35.4 pA vs −540 ± 107.1 pA at −10 mV in control and Bdnf cKO mice, respectively (n = 4 vs 5; Figure 5H). Furthermore, we assessed the ΣΔC_m during 20 pulses of a 10 ms depolarization at 10 Hz, a protocol that gradually depletes the RRP and thus reflects the RRP size. The ΣΔC_m evoked by 20 depolarizing pulses was reduced in the Bdnf cKO mice (107 ± 19.14 fF, n = 10) as compared with that in the control (525 ± 134 fF, n = 8; p=0.006, unpaired t-test; Figure 5I). There was no difference in the endocytosis rate (19.7 ± 8.86 s, n = 7 for the control vs 20 ± 5.28 s, n = 9 for Bdnf cKO; p=0.97, unpaired t-test; data not shown). These findings suggest that oligodendrocytes are critically involved in determining the presynaptic RRP and vesicular glutamate release through BDNF signaling during postnatal development.

We examined the role of oligodendroglial BDNF in glutmatergic transmission in the immature (P10-P12, before hearing onset, Figure 6A) and mature calyx synapses (P16-P20, after hearing onset, Figure 6B) during postnatal development. The amplitude of eEPSCs was significantly smaller in both immature and mature Bdnf cKO mice (2.4 ± 0.53 nA, n = 6 in Bdnf cKO vs 4.9 ± 0.85 nA, n = 5 in control at P10-12; p=0.032, unpaired t-test; Figure 6A,C, and 1.6 ± 0.37 nA, n = 11 in Bdnf cKO vs 6.1 ± 0.51 nA, n = 9 in control at P16-20; p<0.0001, unpaired t-test; Figure 6B,E). In both immature and mature synapses, there was no difference in PPR (Figure 6D,F). Next, we examined the RRP size of available glutamate vesicles and its release probability at presynaptic terminals in control and Bdnf cKO mice at different ages (P10-12 vs P16-20). Using the EQ method, calyces in Bdnf cKO mice had a much smaller RRP of glutamate vesicles as compared with control mice (9.8 ± 0.58 nA in Bdnf cKO mice, n = 6 vs 20.1 ± 2.57 nA in control at P10-12, n = 3; p=0.0238, Mann-Whitney test; Figure 6G and 9.6 ± 3.95 nA in Bdnf cKO mice, n = 6 vs 34.6 ± 3.39 nA in control at P16-20, n = 9; p=0.0004, unpaired t-test; Figure 6I). Conversely, the P_r was not different in both immature and mature control and Bdnf cKO mice (0.33 ± 0.026 in Bdnf cKO mice, n = 6 vs 0.32 ± 0.012 in control at P10-12, n = 3; p=0.7619, Mann-Whitney test; Figure 6H and 0.41 ± 0.038 in Bdnf cKO mice, n = 4 vs 0.38 ± 0.026 in control at P16-20, n = 9; p=0.4459, unpaired t-test; Figure 6J). In addition, the SMN method analysis showed a reduction in the RRP and the replenishment rate of RRP in Bdnf cKO mice, without significant difference in P_r (Figure 6—figure supplement 1). A deletion of BDNF from oligodendrocytes around the calyx synapses significantly impaired the RRP and glutamate release at immature and mature calyx synapses in Bdnf cKO mice, suggesting that oligodendroglial BDNF is important for regulating glutamatergic transmission in the auditory brainstem before and after hearing onset.

Figure 6 with 1 supplement see all

Download asset Open asset

To visualize changes in the presynaptic RRP and to quantify the number of glutamate vesicles at the active zone at the calyx terminal, we performed ultrastructural analysis of the calyx–MNTB neuron synapse (at P10-12 and P20) using electron microscopy (EM). Within individual active zones of the calyx terminals, an average of 2–3 docked vesicles was observed in control mice, whereas there were fewer docked vesicles or an absence of docked vesicles at the active zones in the Bdnf cKO mice (Figure 7A). In immature calyx synapses at P10-12, the number of docked vesicles located within 10 nm from the presynaptic active zone membrane was 2.1 ± 0.24 vesicles (62 active zones of three individual cells), whereas in the Bdnf cKO mice the number of docked vesicles was significantly reduced to1.5 ± 0.18 vesicles (67 active zones of five individual cells, p=0.0285, unpaired t-test, Figure 7B, Figure 7—figure supplement 1). To test whether oligodendroglial BDNF influences the development of calyces, we assessed the size of calyces using 3D reconstruction of confocal images of the calyx terminals from control and Bdnf cKO mice (at P10-12) after presynaptic recordings. The volume of the calyx terminal was not significantly different in Bdnf cKO mice (1378 ± 143.7 μm³, n = 7 for control and 1199 ± 146 μm³, n = 6 for Bdnf cKO; p=0.3308, Mann-Whitney test; Figure 7—figure supplement 2). This result was consistent with the membrane capacitance (C_m) measurement; there was no difference (15 ± 1.9 pF, n = 14 for control vs 15 ± 3.5 pF, n = 17 for Bdnf cKO, Figure 5F). In mature calyx synapses at P20, 2.7 ± 0.14 vesicles were located within 10 nm and 20.2 ± 0.73 vesicles were within 200 nm of the active zone of the calyx terminal in the control (counted in 166 active zones from four cells; Figure 7A,C). In Bdnf cKO mice, 1.0 ± 0.12 and 19.1 ± 0.75 vesicles were located within 10 nm and 200 nm of the active zone, respectively, (196 active zones from five individual cells; <10 nm, p<0.0001; <200 nm, p=0.26, unpaired t-test; Figure 7C). Thus, the number of docked vesicles was significantly decreased in both immature and mature calyces in the Bdnf cKO mice. These anatomical changes in presynaptic terminals strongly indicate that oligodendroglial BDNF signaling is important for determining the RRP and specifically for mobilizing glutamate vesicles at the presynaptic terminal during postnatal development.

Figure 7 with 2 supplements see all

Download asset Open asset

We next tested whether activation of presynaptic TrkB using an agonist can recover the reduced number of docked vesicles at the active zone in Bdnf cKO mice. Auditory brainstem slices from control and Bdnf cKO mice (at P20) were prepared for EM imaging after 30 min pre-treatment with 7,8-DHF (20 μM) as described in Figure 4. Application of 7,8-DHF recovered the reduced number of docked vesicles to 2.2 ± 0.22 within 10 nm of the active zone (63 active zones from three individual cells). There was no change in the number of docked vesicles within 200 nm of the active zone (Figure 7A,C). This result indicates that activation of BDNF-TrkB signaling rescues the docking defect or impaired mobilization of vesicles at the active zone, resulting in recovery of the reduced RRP in Bdnf cKO mice.

We next tested whether extracellular application of BDNF or 7,8-DHF can recover the impaired glutamate vesicle release at presynaptic terminals in Bdnf cKO mice. The pre-application of BDNF (100 ng/ml) to brainstem slices for 30 min increased ΔC_m in response to depolarizing pulses at the calyx terminal from Bdnf cKO mice. After 20 ms depolarizing pulses, ΔC_m was much larger at calyces after BDNF application (200 ± 12.72 fF, n = 5) relative to untreated terminals from Bdnf cKO mice (93.8 ± 23.51 fF, n = 13; p=0.04, unpaired t-test; Figure 8A,B). There were no corresponding changes in QI_Ca in treated and untreated terminals (for 20 ms pulses, 5.1 ± 0.47 pC/pF, n = 13 vs 5.8 ± 0.65 pC/pF, n = 5, respectively; p=0.42, unpaired t-test; Figure 8B). Interestingly, the application of BDNF had no effect on ΔC_m and QI_Ca in control calyces with a normal RRP (Figure 8—figure supplement 1). In addition, the direct activation of TrkB also rescued the impaired RRP and glutamate release at the calyx terminal in Bdnf cKO mice. After 20 ms depolarization pulses, ΔC_m was much larger at calyces in the presence of 7,8-DHF as compared with those from Bdnf cKO without the 7,8-DHF application (177.6 ± 12.72 fF, n = 5 vs 93.8 ± 23.51 fF, n = 13, respectively; p=0.04, unpaired t-test; Figure 8A,B). There was no change in the QI_Ca in the presence of 7,8-DHF (for 20 ms pulses, 5.1 ± 0.47 pC/pF, n = 13; p=0.42, unpaired t-test; Figure 8B). Furthermore, the ΣΔC_m induced by 20 pulses of 10 ms depolarization at 10 Hz was significantly increased by ~100% in the presence of 7,8-DHF in the Bdnf cKO (107 ± 19.14 fF, n = 10 in the absence of 7,8-DHF vs 328 ± 60 fF, n = 7 in the presence of 7,8-DHF; p=0.001, unpaired t-test; Figure 8C–E). The extracellular BDNF application also partially restored the ΣΔC_m to 399 ± 120 fF in the Bdnf cKO (n = 6; p=0.008, unpaired t-test; Figure 8E). Thus, the activation of BDNF–TrkB signaling by the application of BDNF or 7,8-DHF partially recovered the impaired RRP and exocytosis at the presynaptic terminal in the Bdnf cKO. These findings suggest that oligodendrocyte-derived BDNF activates presynaptic TrkB signaling, which modulates the RRP and enhances glutamatergic transmission in the MNTB during postnatal development.

Figure 8 with 1 supplement see all

Download asset Open asset

To assess how loss of oligodendroglial BDNF and subsequent synaptic dysfunction influence auditory functions along the central auditory system, we measured auditory brainstem responses (ABRs), which represent the summed synchronized activity of neurons in the auditory pathway (Kim et al., 2013), in control and Bdnf cKO mice (P20–25). In both, the ABR waveform consisted of five distinct peaks (herein referred to as waves I–V) during the 6 ms following a click stimulus and each wave corresponds to electrical responses from the auditory nerve (wave I) and the ascending auditory pathway (e.g. cochlea nucleus, the superior olivary complex, lateral lemniscus, and inferior colliculus; wave II-V). There was no difference in the threshold of ABRs in response to click stimulation in control and Bdnf cKO mice (42.8 ± 2.39 dB vs 42.8 ± 3.04 dB, n = 19 vs 14, respectively; Figure 9A,B). In addition, the latency of wave I, and the time difference between wave I and wave IV, indicating central conduction, did not show significant difference in Bdnf cKO mice. We did not observe a significant difference in the amplitude of wave I, whereas the amplitudes of ABR waves II–IV were significantly reduced in Bdnf cKO mice (Figure 9A,B). In particular, the amplitude of wave III, which reflects the summed neuronal activities of the superior olivary complex, was significantly reduced in the range of click intensities from 55 dB to 85 dB in Bdnf cKO mice (at 75 dB, 2.6 ± 0.19 μV, n = 21 for control and 1.7 ± 0.16 μV, n = 16 for the Bdnf cKO; p=0.002, unpaired t-test; Figure 9B). There was no significant difference in the latency of wave I, indicating peripheral conduction, and in central conduction, which was estimated by the time difference between wave IV and wave II (Figure 9B). These ABRs indicate that neuronal activity and synaptic synchrony in central auditory nuclei are impaired in Bdnf cKO mice. Taken together, the ABRs suggest that endogenous oligodendroglial BDNF regulates the synchrony of synaptic activities and critically influences auditory transmission during postnatal development.

Figure 9

Download asset Open asset

In this study, we found that a reduction in endogenous BDNF in Bdnf^+/− and Bdnf cKO mice impaired glutamatergic transmission without altering presynaptic Ca²⁺ channel activation. The results are comparable to a previous study in the inner ear showing the deletion of endogenous BDNF significantly reduced exocytosis of glutamate vesicles but did not affect Ca²⁺ currents in cochlear hair cells of mice (Zuccotti et al., 2012). Furthermore, we found that the application of BDNF or a TrkB agonist (7,8-DHF, 20 μM) led to partial recovery of the reduction in the RRP and in exocytosis of vesicular glutamate in Bdnf^+/− and Bdnf cKO, but there was no significant effect in WT or control mice (Figures 7 and 8, Figure 8—figure supplement 1). However, a previous study in the MNTB of the rat brainstem showed that exogenous BDNF application reduces glutamate release by slowing down presynaptic Ca²⁺ channel activation and inhibiting exocytosis and endocytosis (Baydyuk et al., 2015). These conflicting findings may result from the differences between species, ages, or BDNF application method. In particular, the timing of the BDNF signal induction may differentially modulate synapse function as either acute or chronic applications of BDNF can differentially modulate synaptic plasticity (Sherwood and Lo, 1999; Schildt et al., 2013; Guo et al., 2018). Exogenous administration of BDNF to brain slices has limitations. Depending on administration time, exogenous application could result in the non-specific binding effect of BDNF to presynaptic Ca²⁺ channels, resulting in inhibition of Ca²⁺ channel activation rather than through BDNF-TrkB signaling. Further studies are required to determine the effect of exogenous BDNF on Ca²⁺ channel subtypes expressed in the presynaptic terminal and what aspects of BDNF signaling generate differential responses.

The mechanisms underlying the presynaptic effects of BDNF-TrkB signaling remain elusive. Activation of TrkB leads to the induction of a combination of downstream signaling pathways, including the mitogen-activated protein kinase (MAPK), the PLC pathway, and the phosphatidylinositol 3-kinase (PI3K) pathway, that could modulate synaptic vesicles at the presynaptic terminal (Yoshii and Constantine-Paton, 2010; Reichardt, 2006). The acute and local effects of oligodendrocyte-derived BDNF on the RRP could be mediated by the increases of intracellular Ca²⁺ levels, which may depend on the activation of the PLC pathway (Matsumoto et al., 2001; Reichardt, 2006). Recent studies demonstrated that BDNF-induced rise in intracellular Ca²⁺ concentration at the presynaptic terminal was mediated by Ca²⁺ influx through TRPC3 channels, resulting in a transient increase in spontaneous glutamate release (Cheng et al., 2017), and/or release of Ca²⁺ from intracellular stores (Amaral and Pozzo-Miller, 2012).

It is important to identify the source of BDNF release at the synapse to understand how BDNF functions and acquires target specificity. BDNF increases by ~10-fold in the mouse CNS in the first 3 postnatal weeks (Kolbeck et al., 1999; Tao et al., 1998). Although the major source of BDNF in the adult brain appears to be neurons (Hofer et al., 1990; Rauskolb et al., 2010), BDNF is frequently detected in oligodendrocytes, astrocytes, and microglia in the developing brain (Dougherty et al., 2000). BDNF expression is observed in auditory brainstem nuclei in the mouse from P6, and its expression follows the protracted period of development in the auditory pathway, with expression beginning in the ventral cochlear nucleus and continuing to the MNTB and then to the medial superior olive and the lateral superior olive (Wiechers et al., 1999; Hafidi, 1999). Glial cells may participate in modulating synaptic structure and function during the development of the auditory circuitry by providing a permissive environment through the secretion of BDNF. Oligodendrocytes populate the MNTB prior to astrocytes, indicating oligodendrocytes have a primary role in the maturation of synapses during MNTB development. During the early postnatal weeks, oligodendrocytes are present throughout the auditory brainstem including the MNTB nuclei (as they were at birth), whereas GFAP-positive astrocytes appear in the MNTB during the second postnatal week (Dinh et al., 2014). Our immunostaining results, which are consistent with this previous study, showed that there is a greater oligodendrocyte population in the MNTB as compared with GFAP-positive astrocytes by P8 and that most oligodendrocytes are located in proximity to the calyx synapse (data not shown).

This study demonstrated the presence of BDNF in oligodendrocytes in the MNTB of mouse brainstem during early postnatal development. Oligodendrocytes expressed BDNF in the MNTB of the auditory brainstem (Figure 1A), and isolated O1+ or Cnp^cre-driven GCaMP6f-GFP cells expressed a substantial amount of Bdnf mRNA (Figure 5—figure supplement 2). Previous study demonstrated that oligodendrocytes release BDNF in response to glutamate application (Bagayogo and Dreyfus, 2009). Oligodendrocyte processes contact the calyx terminal, which releases glutamate, before forming a myelin sheath during early development (Figure 5A). This suggests that glutamate-mediated signaling between oligodendrocytes and the calyx synapses induces BDNF release from oligodendrocytes to increase synaptic strength. Due to the biochemical nature of BDNF, it is thought to act locally at the synapse with limited diffusion within the micrometer range (Horch and Katz, 2002; Sasi et al., 2017). Oligodendrocytes likely exert a direct impact through the localized secretion of BDNF to the calyx synapses. The results demonstrate oligodendrocytes actively participate in bidirectional neuron–glia communication at the calyx synapse through BDNF-dependent signaling during early postnatal development.

This study utilizes Cnp^cre to generate a Bdnf deletion specifically in oligodendrocytes. A recent study reported that Cnp^cre-driven YFP reporter signal was detected in 5.5% of NeuN+ neurons, suggesting a potential limitation on the specificity of recombination of Cnp^cre mouse line (Tognatta et al., 2017). To address the specificity of Cnp^cre in the MNTB, we have analyzed reporter expression and verified a specific reduction of Bdnf in isolated oligodendrocytes. Using two reporter lines, Rosa-GCaMP6f-GFP and Rosa-tdTomato, we identified that <5% of neurons in the MNTB expressed Cnp^cre-driven reporter in early postnatal ages (P10- P20). In addition, using FACS, isolated GCaMP6f-GFP+ cell population contains high levels of Olig2 mRNA with very low levels of Kcc2 mRNA, a neuronal marker (Figure 5—figure supplement 2). The majority (>95%) of Cnp^cre expressing cells in the MNTB are oligodendrocytes as shown through immunohistochemistry. We demonstrated that GCaMP6f-GFP+ cells have detectable Bdnf mRNA, which was significantly reduced in Bdnf cKO mice. There was no significant difference in the level of Bdnf mRNA in the GFP− or an O1− fraction, which considered as a non-oligodendroglial population, although there was a trend toward lower Bdnf mRNA in the O1− or GFP− fraction. There is the possibility that the small percentage of neurons, affected by Cnp^cre, is sufficient to reduce global levels of Bdnf and impact on the synaptic phenotype in the cKO. In cultured neurons, the effect of BDNF within a synapse has been observed to occur within a distance of 4.5 μm (Horch and Katz, 2002). Thus, BDNF reduction in a small portion of neurons (<5%) is unlikely to have widespread effects or global impact on the synaptic phenotype observed in the cKO. We interpret that functional alterations of the calyx synapse were caused by the loss of BDNF in oligodendrocytes, which constitute the majority of CNP-expressing cells.

In cultured oligodendrocytes, the activation of glutamate receptors and the phospholipase C pathway enhances the release of dense-core vesicles containing BDNF (Bagayogo and Dreyfus, 2009), suggesting that release of BDNF from oligodendrocytes depends on neuronal activity and is mediated by neuron–oligodendrocyte interactions. Our recent study demonstrated that a sub-population of oligodendrocytes interacts with neurons via synapses and displays action potentials in response to intensive neuronal activities in the auditory brainstem (Berret et al., 2017). It is intriguing to speculate that bidirectional signaling occurs between oligodendrocytes and nerve terminals at synapses, in which glutamatergic inputs from neurons trigger oligodendrocytes to release BDNF, and then oligodendrocyte-derived BDNF binds to presynaptic TrkB, and finally modulates the glutamate vesicle pool at the nerve terminal. These findings indicate that oligodendrocytes may modulate synaptic plasticity in an activity-dependent manner. An increase in the RRP of vesicles could also contribute to short-term plasticity such as PTP (Habets and Borst, 2005; Regehr, 2012). We show that the reduction of global BDNF significantly impairs the induction of PTP at the calyx synapse in Bdnf^+/– mice (Figure 2). Oligodendrocytes can regulate synaptic strength and plasticity at the calyx synapse by modulating the RRP size through BDNF signaling. Oligodendrocytes that are closely apposed to synapses thus monitor and sense synaptic activity and modulate synaptic plasticity at the presynaptic terminals. In the case of the calyx synapse, this occurs through BDNF–TrkB signaling, which may represent an efficient way for oligodendrocytes to find active nerve terminals and to assist in maintaining synaptic activities. During this critical window of development, when activity-dependent synaptic refinement can occur along the auditory nervous system, oligodendrocytes actively participate in synaptic transmission and plasticity through BDNF signaling in the developing brain.

All animal procedures were performed in accordance with the guidelines approved by the University of Texas Health Science Center, San Antonio (UTHSCSA) Institutional Animal Care and Use Committee protocols. BDNF heterozygous (Bdnf^+/−) mice were generated by crossing Bdnf^+/− mice (B6.129S4-Bdnf^tm1Jae/J; The Jackson Laboratory) with WT mice (C57B[L]6/J). The offspring were genotyped with a standard PCR assay. Primer sequences were as follows: forward, 5’-ATGCGTACCTGACTTTCTCCTTCT-3’; reverse, 5’-ACTGGGTGCTCAGGTACTGGTTGT-3’, which amplify a 280 bp and 350 bp fragment for Bdnf^+/– mice and a 280 bp fragment for WT.

To create the cKO mice (Cnp^cre: Bdnf^fl/fl), mice carrying the floxed allele of Bdnf (Bdnf^fl/fl; The Jackson Laboratory; Rios et al., 2001) were crossed to Cnp^cre heterozygous mice (Lappe-Siefke et al., 2003). The constitutive KO allele is obtained after Cre-mediated recombination by crossing Cnp^cre mice with Bdnf^fl/fl mice to obtain the deletion of Bdnf only in CNPase-expressing cells (Lappe-Siefke et al., 2003; Rios et al., 2001). Genotypes of all mice were determined by PCR analysis of tail genomic DNA using the appropriate primers: for Bdnf^fl/fl, forward, 5’-TGTGATTGTGTTTCTGGTGAC-3’ and reverse, 5’-GCCTTCATGCAACCGAAGTATG-3’, which amplifies a 487 bp (floxed Bdnf allele) and a 437 bp (Bdnf WT allele) fragment; for Cnp^cre, forward, 5’-GCCACACATTCCTGCCCAAGCTC-3’ and reverse 1, 5’-GTCGCCACGCTGTCTTGGGCTCC-3’, and reverse 2, 5’-CTCCCACCGTCAGTACGTGAGAT-3’, which amplifies a 400 bp (Cnp WT allele) and 550 bp (Cnp^cre allele) fragment. Control mice (Cnp^cre: Bdnf^fl/+) were identified by PCR amplification of a 400 bp, 437 bp, and 487 bp fragment, whereas Bdnf cKO mice (Cnp^cre: Bdnf^fl/fl) were identified by PCR amplification of a 400 bp, 550 bp, and 487 bp fragment. Recombination efficiency in oligodendrocytes in the Cnp^cre mice was determined by transgenic crosses to the GCaMP6f reporter mouse (provided by Dr. Paukert, UTHSCSA or purchased from Jackson Laboratory). All mice were housed in the institutional animal facilities on a 12 hr light/dark cycle. Mice of both sexes aged P8–25 were used for all experiments.

Transverse brainstem slices containing the MNTB were prepared from P9–18 mouse pups. After rapid decapitation of the mice, the brains were quickly removed from the skull and immediately immersed in ice-cold low-calcium artificial cerebrospinal fluid (aCSF) containing (in mM): 125 NaCl, 2.5 KCl, 3 MgCl₂, 0.1 CaCl₂, 25 glucose, 25 NaHCO₃, 1.25 NaH₂PO₄, 0.4 ascorbic acid, three myoinositol, and 2 Na-pyruvate, pH 7.3–7.4 when bubbled with carbogen (95% O₂, 5% CO₂; osmolarity of 310–320 mOsm). Then, 200-µm-thick sections were collected using a Vibratome (VT1200S, Leica, Germany). Slices were incubated in a chamber that contained normal aCSF bubbled with carbogen at 35°C for 30 min and then were kept at room temperature. The normal aCSF was the same as the low-calcium aCSF, except that 3 mM MgCl₂ and 0.1 mM CaCl₂ were replaced with 1 mM MgCl₂ and 2 mM CaCl₂.

Whole-cell patch-clamp recording was carried out on postsynaptic principal neurons and presynaptic calyx of Held terminals in the MNTB using an EPC-10 amplifier controlled by PATCHMASTER software (HEKA Elektronik, Lambrecht/Pfalz, Germany). Slices were visualized using an infrared differential interference contrast microscope (AxoExaminer, Zeiss, Oberkochen Germany) with a 63 × water immersion objective and a CMOS camera (Hamamatsu Photonics, Hamamatsu, Japan). During experiments, slices were perfused with normal aCSF solution at 2 ml/min at room temperature.

Slices used for patch-clamp analysis or fresh brainstem slices (~200 µm thick) were fixed with 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS) for 20 min. Free-floating slices were blocked in 4% goat serum and 0.3% (w/v) Triton X-100 in PBS for 1 hr and then were incubated with primary antibody overnight at 4°C. The following primary antibodies were used: mouse anti-Olig1 (1:500; Millipore, MAB5540), mouse anti-MAP2 (1:200; Millipore, MAB3418), mouse anti-NeuN (1:200; Millipore, MAB377), rabbit anti-GFAP (1:500; DAKO, Z033429), mouse anti-CC1 (1:200, Millipore, OP80), mouse anti-NeuN (1:600, Millipore, MAB377), anti-Olig2 (1:100, Abcam, 109186), rat anti-PDGFRa (1:300, abcam, AB90967), rabbit anti-BDNF (1:100; Bioss, BS4989R), mouse anti-TrkB (1:50; Santa Cruz, sc-136990), and guinea pig anti-VGluT1 (1:1000; Millipore, AB5905). Tissues were then incubated with different Alexa-conjugated secondary antibodies (1:500; Invitrogen) for 2 hr at room temperature. After three rinses with PBS, slices were coverslipped using mounting medium with 4′,6-diamidino-2-phenylindole (DAPI; Vectashield; Vector Laboratories) to counterstain cell nuclei. Stained slices were viewed on a confocal laser-scanning microscope (Zeiss LSM-510) at 488, 568, and 633 nm using 40 × or 60 × oil immersion objective.

Animals were anesthetized and intracardially perfused with normal saline. Brains were removed and 400-μm-thick samples of brainstem MNTB area were dissected out followed by primary fixation in 1% glutaraldehyde/4% paraformaldehyde. Further processing was performed by the UTHSCSA Electron Microscopy Lab. Briefly, each brainstem was post-fixed with 1% Zetterqvist’s buffered osmium tetroxide, dehydrated, and embedded in PolyBed resin at 80°C in an oven. Tissue containing the MNTB, which is innervated by calyces of Held, was cut into 90 nm ultrathin sections and placed on copper grids. The sections were then stained with uranyl acetate and Reynold’s lead citrate. The samples were imaged on a JEOL 1400 electron microscope using Advanced Microscopy Techniques software. The calyx of Held terminals contacting cell bodies of MNTB principal neurons were recognizable as a cluster of cells located medially in the superior olivary complex (Taschenberger et al., 2002). A total of 166–196 synapses were analyzed from five animals for each group (control and Bdnf cKO). The number of docked vesicles per the active zone was measured for each synapse at a final magnification of 80,000×. The active zone was defined as the dark presynaptic density contacting the postsynaptic density. Docked and clustered vesicles were defined as those within 10 nm and 200 nm of the presynaptic active zone, respectively (Sätzler et al., 2002; Taschenberger et al., 2002).

ABR recordings were performed as described (Kim et al., 2013). Briefly, mice were anesthetized with 4% isoflurane and maintained with 2% isoflurane during recording (1 l/min O₂ flow rate). ABR recordings were carried out in a sound attenuation chamber (Med Associates, Albans, VT). Body temperature of mice was maintained at 37°C using a heating pad. Subdermal needle electrodes for recording were placed on the top of the head (active), ipsilateral mastoid (reference), and contralateral mastoid (ground). The electrical potential differences between the vertex and the mastoid electrodes were amplified and filtered (100–5000 Hz), and a recording window of 10 ms starting at the onset of click sound stimulation was distally sampled at 40-μs intervals. Acoustic stimuli were generated by the Auditory Evoked Potentials Workstation (Tucker-Davis Technologies [TDT], Alachua, FL). Closed-field click stimuli were delivered to the left ear using a series of square waves (0.1 ms duration) through TDT Multi-Field Magnetic Speakers placed 10 cm away from the left ear canal. A repetition rate of sound stimuli of 16/s was transmitted through a 10 cm length of plastic tubing (Tygon; 3.2 mm outer diameter). Sound intensities ranged from 90 to 20 dB with 5 dB decrements and responses to 512 sweeps were averaged. The lowest sound intensity that produced a reproducible waveform was interpreted as the threshold. Free-field pure tone stimuli were taken at frequencies of 8, 12, 16, 24, and 32 kHz at 70 to 20 dB in decrements of 5 dB.

Bdnf cKO mice (Cnp^Cre: Bdnf^fl/fl, n = 3) and control mice (Cnp^Cre: Bdnf^fl/+, n = 2) were used for FACS experiments. Cnp^Cre: Rosa-GCAMP6f-GFP^+/- (n = 3) and Cnp^Cre heterozygous controls (n = 3) were used. A cell suspension was generated from the brainstem using enzymatic (papain, 48 U/mL, P1325, Sigma) and mechanical titration. Dissected brainstem was incubated in dissociation media (145 mM NaCl, 5 mM KCl, 20 mM Hepes, 1 mM Na-pyruvate, 2 mM EDTA, pH 7.2) with papain and DNase (10 μM, 10104159001, Sigma) for 20 min. The tissue was spun down at 300 x g for 5 min and resuspended in 1 ml of dissociation media with DNase (10 μM). Mechanical titration was performed using a 1000 ul pipette tip followed by a 200 ul pipette tip. After dissociation, the suspension was filtered and spun down at 300x g for 7 min. The cells were resuspended in 400 μl of dissociation media. 50 μl were set aside for a no primary antibody control. 1 μl anti-O1 antibody (MAB1327, R and D systems) was added to the remaining 350 μl cell suspension (dilution 1:350) and cells were incubated on ice for 25 min. Cells were washed with 1 ml of dissociation media and spun at 300x g for 7 min. Cells were resuspendend in 400 μl dissociation solution with secondary antibody (Alexa 488 anti-mouse IgM, 1:500) and incubated for 25 min. Cells were washed with 1 ml of dissociation media, spun at 300x g for 7 min, and resuspended in 400 μl of dissociation solution. Sorting was performed on a BD FACSARIA III (BD Biosciences) in the Flow Cytometry Facility at UT Health San Antonio with funding from University and the NIH (NCI P30 CA054174). BD FACS Diva software 8.0.1 was used to visualize forward scatter and side scatter to determine cell population and perform doublet discrimination. 488-labeled O1+ cells were selected, yielding a population of 7,000–25,000 cells. In Cnp^Cre: Rosa-GCAMP6f-GFP^+/- mice, GFP+ cells produced a population of 50,000–75,000 cells. Negative cells were also collected with a total of 200,000–1,000,000 cells. Cells were kept on ice prior to RNA isolation.

RNA isolation was performed using an RNAqueous kit with no modifications to procedure (AM1931, Thermofisher). This kit includes DNase treatment. RNA was quantified using a nanodrop (ND-1000, Thermofisher). Purity was assessed by A260/A280 ratios with values ranging from 1.83 to 1.98. 100 ng of RNA was utilized for each reverse transcription (RT) reaction. RT was performed using Superscript III First-strand synthesis with 1 μl Oligo (dT) primers, 1 μl 10 mM dNTP mix (180180–051, Invitrogen). cDNA was stored at −20C for up to 48 hr. qPCR was performed using the PowerUp SYBR green master mix (A25742, Thermofisher). Primers were used at 0.25 μM. Reactions were manually loaded into optical plates with covers (plates:4309849, covers:4360954, Applied Biosystems). The plate was spun for 30 s. qPCR was performed using 7900HT Fast Real-Time PCR system (Applied Biosystems), data was analyzed using SDS v2.4 (Applied Biosystems) and the determined CT was used for analysis. No outliers were removed. No template controls (NTC) as well as no RT controls did not produce determinable CT values. Gapdh was chosen as a reference. Technical replicates were done in triplicate. Delta CT (mean_Gene – mean_GAPDH) was used to determine Delta Delta CT (Delta CT_Pos-Delta CT_Neg) or (Delta CT_cko-Delta CT_CTL), relative gene expression was calculated 2^(-deltadeltaCT) and normalized to 100%. Mouse primers used follows:

• Olig2 Forward: 5’-CAAATCTAATTCACATTCGGAAGGTTG

• Olig2 Reverse: 5’-GACGATGGGCGACTAGACACC

• Kcc2 Forward: 5’-GGGCAGAGAGTACGATGGC

• Kcc2 Reverse: 5’-TGGGGTAGGTTGGTGTAGTTG

• BDNF Forward: 5’-TCGTTCCTTTCGAGTTAGCC

• BDNF Reverse: 5’-TTGGTAAACGGCACAAAAC

• GAPDH Forward: 5’-AGTATGACTCCACTCACGGCAA

• GAPDH Reverse: 5’-TCTCGCTCCTGGAAGATGGT

All statistical analyses were performed in GraphPad Prism. For electrophysiology, the n equals the number of individual whole-cell recordings. All electrophysiological experiments were performed in at least seven independent slices from at least seven individual animals. The in vivo ABR test was performed in at least 20 individual controls and at least 15 individual Bdnf cKO mice. Data were analyzed off-line and displayed with Igor Pro (Wavemetrics, Lake Oswego, OR). α values were set to 0.05, and all comparisons were two-tailed. To compare two groups, unpaired t-test or Mann-Whitney U test was carried out. Differences were considered statistically significant when p-values were <0.05 by a Student’s t-test or Mann-Whitney U test (GraphPad Prism). Data are shown as the mean ± standard error of the mean (s.e.m.)

The authors declare that all data generated or analyzed in this study are available within the article.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We would like to thank Drs. Klaus Nave and Manzoor Bhat for providing the Cnp^cre mouse line. This work was supported by a grant from the National Institute on Deafness and Other Communication Disorders (NIDCD; R01 DC03157) to J H Kim.

Animal experimentation: All animal procedures were performed in accordance with the guidelines approved by the University of Texas Health Science Center, San Antonio (UTHSCSA) Institutional Animal Care and Use Committee protocols (#140045x).

© 2019, Jang et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.