Pre- and postsynaptic modulations of hypoglossal motoneurons by α-adrenoceptor activation in wild-type and Mecp2^−/Y mice

Abstract

Hypoglossal motoneurons (HNs) control tongue movement and play a role in maintenance of upper airway patency. Defects in these neurons may contribute to the development of sleep apnea and other cranial motor disorders including Rett syndrome (RTT). HNs are modulated by norepinephrine (NE) through α-adrenoceptors. Although postsynaptic mechanisms are known to play a role in this effect, how NE modulates the synaptic transmissions of HNs remains poorly understood. More importantly, the NE system is defective in RTT, while how the defect affects HNs is unknown. Believing that information of NE modulation of HNs may help the understanding of RTT and the design of new therapeutical interventions to motor defects in the disease, we performed these studies in which glycinergic inhibitory postsynaptic currents and intrinsic membrane properties were examined in wild-type and Mecp2^−/Y mice, a mouse of model of RTT. We found that activation of α₁-adrenoceptor facilitated glycinergic synaptic transmission and excited HNs. These effects were mediated by both pre- and postsynaptic mechanisms. The latter effect involved an inhibition of barium-sensitive G protein-dependent K⁺ currents. The pre- and postsynaptic modulations of the HNs by α₁-adrenoceptors were not only retained in Mecp2-null mice but also markedly enhanced, which appears to be a compensatory mechanism for the deficiencies in NE and GABAergic synaptic transmission. The existence of the endogenous compensatory mechanism is an encouraging finding, as it may allow therapeutical modalities to alleviate motoneuronal defects in RTT.

hypoglossal motoneurons (HNs) control tongue movements and play a role in mastication, swallowing, sucking, and maintenance of upper airway patency (4, 20, 39). Defects in these neurons may contribute to the development of sleep apnea and other cranial motor disorders including Rett syndrome (RTT). The HNs receive excitatory and inhibitory synaptic inputs (1, 7, 19, 47, 51, 69). The main source of GABAergic and glycinergic inhibitory inputs to the HNs derives from the nucleus of Roller (2, 62), which affect the HNs via activation of postsynaptic GABA_A and glycine receptors (2, 5, 52, 62). These inhibitory synaptic inputs control hypoglossal responses to other synaptic inputs, shape the temporospatial pattern of neuronal activity during reflex and rhythmic behaviors, and contribute to the generation of inspiratory motoneuronal synchrony (5, 49, 53).

The strength of these synaptic inhibitions of HNs is modulated by brainstem norepinephrine (NE) systems by α-adrenoceptors (18). The α-adrenoceptors contain two main types including α₁ and α₂, each comprising at least three subtypes (13, 42). The α_1A-, α_1B-, and α_1C-adrenoceptors are expressed in cranial and spinal motoneurons (14, 34). The α_2A- and α_2C-receptors are also expressed in motoneurons at relatively low levels (45, 57), while α_2B is expressed only in thalamic neurons (44, 45, 57). The β₁- and β₂- but not β₃-adrenoceptors are expressed in the brain (45), although evidence for their expression in motoneurons remains controversial (44, 48, 50). In the brainstem and the spinal cord, activation of α₁-adrenoceptor produces a large variety of excitatory postsynaptic effects in motoneurons (52). In addition, there is also compelling evidence suggesting that the α₁-adrenoceptor mediates facilitation of GABAergic transmission of neurons in the hippocampus (40), frontal cortex (8), cerebellum (27), nucleus ambiguous (9), and olfactory bulb (43). In the hypoglossal nucleus, the α₁-adrenoceptor is involved in NE-mediated depolarization associated with an increase in input resistance, augmentation of spike frequency, and a decrease in afterhyperpolarization (6, 46, 49, 60). In contrast to these postsynaptic effects, how presynaptic α₁-adrenoceptors act on synaptic transmission of HNs is poorly understood. Such presynaptic effects are likely to exist, as experimental evidence suggests that α₁-adrenoceptor agonists increase glycinergic transmission of motoneurons in the nucleus ambiguus and the spinal cord (3, 9, 20, 24).

The synaptic modulation of HNs may be affected in certain diseases. RTT is a neurodevelopmental disorder caused by disruptions of the MECP2 gene (21, 58). The NE content is deficient in people with RTT and the Mecp2^−/Y mice, mouse models of RTT (29, 61, 65). Our laboratory has also demonstrated that dopamine-β-hydroxylase (DBH), the critical enzyme converting dopamine to NE, is significantly reduced without a loss of the DBH-containing neurons in the locus coeruleus (LC; Ref. 73), and such alternations affect the intrinsic membrane properties and CO₂ chemosensitivity of the LC neurons (71, 72). Several recent studies have demonstrated defects in the γ-aminobutyric acid (GABA) system in RTT patients and the mouse models of RTT (12, 28, 30, 41, 55, 74). Indeed, knockout of Mecp2 in a subset of forebrain GABA_Aergic neurons in mice reveals many features of RTT (12). The GABA_Aergic synaptic transmission in LC neurons and neurons in the ventrolateral medulla is drastically depressed in Mecp2-null mice (12, 30, 41). This as well as the fact that the GABAergic synaptic transmission is augmented by the NE system suggests that the joined disruptions of NE and GABA systems may abrogate, to a large degree, the GABAergic system. Such defects should have profound impact on the inhibitory synaptic transmissions, especially in the motorneuronal system including HNs, as motor defects are characteristic in people with RTT (11, 22, 53). Under the condition of GABAergic defects, the motoneuronal system may rely more on glycinergic system, because the glycinergic system plays an important role in inhibitory synaptic transmission in motoneurons (5, 59, 67), suggesting the necessity to know how the Mecp2 disruption affects the glycinergic system. Therefore, we performed whole cell patch-clamp studies to determine what happens to the glycinergic synaptic currents in Mecp2-null mice and how the modulation of glycinergic synaptic transmission by α₁-adrenoceptor is affected.

EXPERIMENTAL PROCEDURES

Animal.

The Mecp2^−/Y male mice were produced by cross-breeding the Mecp2^+/− females with the WT C57BL/6 males (Jackson Laboratories, Bar Harbor, ME). The offspring were routinely genotyped with a PCR to confirm the absence of Mecp2 gene (30, 71, 73). The Mecp2^−/Y males (13–15 days postnatal) were used in the present study as a mouse model of RTT, while their male littermates served as the WT control. Because of the low availability of these mice, nonlittermate, normal C57BL/6 mice (all males) were also used to increase the sample sizes of the WT group. All experimental procedures using animals were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Georgia State University Institutional Animal Care and Use Committee.

Slice preparation.

All electrophysiological experiments were performed on slices from mice. The mice were anesthetized by inhalation of saturated isoflurane and then decapitated. The brainstem was removed and quickly submerged in the ice-cold oxygenated sucrose-rich artificial cerebrospinal fluid (sucrose aCSF) buffer containing the following (in mM): 200 sucrose, 3 KCl, 2 CaCl₂, 2 MgCl₂, 26 NaHCO₃, 1. 25 NaH₂PO₄, and 10 d-glucose. The solution was bubbled with 95% O₂-5% CO₂ (pH 7. 40). The transverse medullary slices (210–250 μM) were made on a Vibratome 1000 (Vibratome, St. Louis, MO). Slices were transferred to normal aCSF in which the sucrose was substituted with 124 mM NaCl and allowed to recover at 33°C for 1 h. Slices were then transferred to a recording chamber and perfused with oxygenated aCSF containing the following (in mM): 124 NaCl, 2.5 KCl, 1.0 NaH₂PO₄, 1.3 MgSO₄, 10-d-glucose, 26 NaHCO₃, and 2 CaCl₂ (pH 7. 40) with 95% O₂-5%CO₂ bubbling through it at a rate of 3 ml/min and maintained at 32–35°C (osmolarity ∼310 mosM).

Whole cell patch-clamp recordings.

Whole cell patch-clamp recordings were performed as described previously (30, 32, 33). During the recording, the slice was maintained fully submerged in the recording chamber and perfused with oxygenated aCSF. HNs were visualized by IR-differential interference contrast microscopy using a ×40 water immersion objective (Carol Zeiss). The patch pipettes were pulled from borosilicate glass using the Sutter pipette puller (Model P-97, Novato, CA) and measured 3–5 MΩ when filled with an intracellular patch solution. Tight-seal (>1 GΩ) whole cell recording was obtained from the cell body of HNs. Cells were accepted for further analysis only when their membrane potential was more negative than −60 mV and the action potential amplitude >50 mV. The input resistance of individual neurons was calculated by the ratio of membrane potential at the 400 ms of command current.

For whole cell voltage-clamp experiments, patch pipettes were filled with the internal solution containing the following (in mM): 125 KCl, 10 NaCl, 1 CaCl₂, 2 MgCl₂, 10 HEPES, 0.5 EGTA, 2 Mg-ATP, 0.3 GTP, and 5 N-(2, 6-dimethylphenylcarbamoylmethyl) triethylammonium bromide (QX314; pH 7.4; ∼310 mosM). The slice was perfused with aCSF with 95% O₂-5% CO₂, neurons were voltage-clamped at a holding potential of −70 mV, and spontaneous inhibitory postsynaptic currents (sIPSCs) were recorded with an Axopatch 200B amplifier (Molecular Devices, Union City, CA). Signals were filtered at 2 kHz and digitized at 20 kHz. The series resistances (15∼20 mΩ) and holding current were regularly monitored during recording, and neurons were rejected if the resistance changed 20% or holding current altered >100 pA. Concerning the largest currents 200 pA, the maximal space clamp error is 4 mV. To isolate glycine receptor-mediated sIPSCs, 6-cyano-7-nitroquinoxaline-2, 3-dione (CNQX; 20 μM), dl-2-amino-5-phosphonopentanoic acid (dl-AP5; 10 μM) and bicuculline (10 μM) were added to the aCSF. The miniature (m)IPSCs were recorded at a holding potential of −70 mV in the presence of CNQX, dl-AP5, and tetrodotoxin (TTX; 1 μM). The theoretical chloride reversal potential was ∼2.1 mV. For whole cell current-clamp recording, the internal solution contained the following (in mM): 120 potassium gluconate, 20 KCl, 2 MgCl₂, 1 CaCl₂, 0.5 EGTA, 10 HEPES, 2 Mg ATP, and 0.3 GTP pH 7.4 (300–310 mosM). In some experiments, the G-protein inhibitor guanosine 5′-O-(2-thiodiphosphate) (GDP-β-S; 1 mM) was placed in the recording pipette for both voltage- and current-clamp experiments (15). dl-AP5, CNQX, bicuculline, strychnine, phenylephrine (PE), and TTX and were purchased from Tocris Cookson (Ellisville, MO). All compounds were aliquoted and stored at −20°C as a concentrated solution (×1,000) and diluted into the aCSF immediately before use.

Data analysis.

Signals were further filtered at 1-kHz offline and were analyzed using Clampfit 10.3 software (Molecular Devices) and Mini Analysis software (Synaptosotf, Fort Lee, NJ). The cumulative probability distributions were compared by the Kolmogorov-Smirnov test. Data are expressed as means ± SE. Statistical significance was Student's t-test and Mann-Whitney test. Difference was considered significant when P < 0.05.

RESULTS

Modulation of postsynaptic membrane properties of HNs by postsynaptic α₁-adrenoceptors in WT mice.

To determine the postsynaptic effects of α₁-adrenoceptor activation, we performed whole cell current-clamp studies in HNs. As described in a previous study in rats (64), the hypoglossal nucleus in mice contained a heterogeneous population of neurons that can be classified into two main subtypes on the base of their firing patterns: neurons with a decrementing or adapting firing pattern (type D) and neurons with an incrementing or accelerating firing pattern (type I). At P8-P15, most of neurons are type I (64). Similarly, most HNs in our current studies belonged to type I in this age group.

Consistent with previous studies of cranial and spinal motoneurons (16, 63, 68), bath application of PE (3∼5 min) produced reversibly depolarization (Fig. 1A) and an increases in spike frequency in the presence of 10 μM bicuculline (Fig. 1, B and C). On average, the membrane potential was −65 ± 2.7 mV (n = 6 cells) before and −57 ± 3.57 mV (n = 6) during PE (40 μM) application, respectively (P < 0.01, Student's t-test; Fig. 1D). The depolarization was associated with an increase in input resistance (144 ± 18.7 MΩ, n = 6, before and 188.6 ± 20 MΩ, n = 6, during PE, respectively, P < 0.01, Student's t-test; Fig. 1E). The action potential frequency in response to the same current injection increased from 22.9 ± 3.2 Hz (n = 6) to 58.8 ± 11 Hz (n = 6) with PE treatment as well (P < 0.05, Student's t-test; Fig. 1F).

Fig. 1.

α₁-Adrenoceptor agonist augmented postsynaptically hypoglossal motoneuron (HN) excitability in wild-type (WT) mice. A: records of membrane potential, input resistance and spike frequency in response to current injection in an HN in baseline control (top), during phenylephrine (PE; 40 μM) application (middle), and washout PE (bottom). B: frequency-time (f–t) relationship for the same neuron at highest current intensity. C: frequency-current (f–I) relationship for the same neuron. D and E: bar graphs showing the effect of PE on membrane potential (D) and input resistance (E) when perfused with artificial cerebrospinal fluid (aCSF) alone and aCSF plus 1 μM of tetrodotoxin (TTX). F: spike frequency to the current injection under normal aCSF perfusion. G–I: bar graphs showing the effect of PE on membrane potential (G), input resistance (H), and spike frequency (I) under the conditions that electrodes were filled with guanosine 5′-O-(2-thiodiphosphate) (GDP-β-S) and slices were perfused with aCSF plus barium. *P < 0.05; **P < 0.01; NS, not significant; n, number of neurons examined.

The PE-induced excitation existed when the perfusate contained 1 μM TTX but was blocked when patch electrodes were filled with 1 mM GDP-β-S. In the presence of TTX, the membrane potential was −62 ± 9.95 mV (n = 6) before and −52 mV ± 8.2 mV (n = 6) during PE application (P < 0.01, Student's t-test; Fig. 1D), while the input resistance was 140 ± 13.7 MΩ (n = 6) in control and 169 ± 14.3 (n = 6) in the presence of PE (P < 0.001, Student's t-test; Fig. 1E). When applied 1 mM GDP-β-S in the recording pipette for >10 min, the membrane potentials were not evidently changed (−69.4 ± 2.3 mV before vs. −69.2 ± 2.7 mV during PE; P = 0.38; Fig. 1G) and neither were the input resistance (112 ± 16.9 MΩ before vs. 115 ± 17.9 MΩ during PE; P = 0.13; Fig. 1H) and the action potential frequency (35.8 ± 3.9 Hz before vs. 40 ± 2.3 Hz during PE; P = 0.12; Fig. 1I) in five HNs (Student's t-test). These results indicate that activation of α₁-adrenoceptors augments HN excitability by G protein-dependent postsynaptic mechanisms.

Involvement of G protein-coupled and barium-sensitive K⁺ channels.

The α₁-adrenoceptors are metabotropic and can inhibit G protein-coupled K⁺ currents including inwardly rectifying K⁺ (GIRK) channels (10) and TASK channels/the resting leak K⁺ currents (49). Indeed, previous studies have suggested that at least two ionic mechanisms were involved in PE-induced motor neuron excitation, which are barium sensitive and barium insensitive, in rat facial and HNs (20, 37, 49) (for reviews, see Ref. 52). We have shown that PE depolarized the HNs associated with increased input resistance suggesting barium-sensitive mechanism may be involved in HNs. Thus the effects of PE on membrane potentials, input resistance, and action potential frequency were examined in the presence of 300 μM BaCl₂, a concentration known to inhibit selectively the inward rectifier K⁺ channels (30, 35, 36). Under this condition, the effects of PE on membrane potentials, input resistance and action potential frequency were almost totally eliminated in five HNs (Fig. 1, G–I). The membrane potentials were −62.3 ± 3 mV before PE and −61.5 ± 2.8 mV during PE (P > 0.05, Student t-test; Fig. 1G), the input resistance was 205 ± 45 MΩ before PE and 208 ± 42 MΩ during PE (P = 0.17, Student's t-test; Fig. 1H) and action potential frequency was 35.6 ± 3.9 Hz before PE and 36 ± 5.3 Hz during PE (P = 0.8, Student's t-test; Fig. 1I).

To test whether the barium-sensitive K⁺ channels is responsible for PE-mediated excitation on HNs, whole cell K⁺ currents were recorded using a ramp voltage protocol (−150 to 10 mV) at a holding potential of −70 mV in the presence of 20 μM CNQX, 30 μM dl-AP5, 10 μM bicuculline, 1 μM strychnine, and 1 μM TTX. The inward currents were measured at −150 mV before and during PE application in five cells. The currents averaged −0.70 ± 0.05 and −0.45 ± 0.02 nA before and during PE application, respectively (P < 0.01; Fig. 2C). The PE-induced inhibition of whole cell currents was also barium sensitive, which was blocked in the presence of 300 μM BaCl₂ (Fig. 2D). Taken together, these results provide further evidence suggesting that G protein-coupled and barium-sensitive K⁺ channels are involved in the postsynaptic effect of PE in HNs.

Fig. 2.

Involvement of barium-sensitive inward K⁺ currents in the postsynaptic effects of α₁-adrenoceptor agonist. A and B: sample traces showing the effect of PE (40 μM) on ramp evoked whole cell currents at holding potential −70 mV in aCSF (A) and aCSF plus barium (B). C and D: bar graphs summarizing PE induced-inhibition on the membrane currents in aCSF (C) and aCSF plus barium (D). All experiments were performed in the present of 20 μM 6-cyano-7-nitroquinoxaline-2, 3-dione (CNQX), 30 μM dl-2-amino-5-phosphonopentanoic acid (dl-AP5), 10 μM bicuculline, 1 μM strychnine, and 1 μM TTX. **P < 0.01; n, number of neurons examined.

Modulation of postsynaptic membrane properties of HNs by postsynaptic α₁-adrenoceptors in Mecp2^−/Y mice.

Accumulating evidence indicates that the NE system is defective in people with RTT (75, 76) and the mouse models of RTT(29, 56, 65, 73). Such a defect may have an impact on cranial motor neurons, as motor disorders are seen in most Rett patients (11, 22, 53). Thus we performed experiments in Mecp2^−/Y mice to elucidate whether the postsynaptic modulation of HN excitability by α₁-adrenoceptor agonist is altered in Mecp2^−/Y mice. We examined the effect of PE on membrane potentials, input resistance and action potential frequency with current injection in eight neurons. As shown in Fig. 3A, PE application clearly depolarized the HN from an Mecp2^−/Y mouse and increased spike frequency to the current injection. On average, the membrane potentials, input resistance, and spike frequency before and during PE application were −71.9 ± 1.72 vs. −56.4 ± 3.3 mV (Fig. 3D; P = 0.001, Student' s t-test), 144.4 ± 17.4 vs. 196 ± 27.9 MΩ (Fig. 3E; P < 0.05, Student's t-test), and 37 ± 4.5 Hz vs. 74.9 ± 8 Hz (Fig. 3F; P < 0.01, Student's t-test), respectively. These results are similar to those obtained from WT (Fig. 1), suggesting that the effects of postsynaptic α₁-adrenoceptor activation on HN membrane potentials, input resistance, and spike frequency are retained in Mecp2^−/Y mice.

Fig. 3.

Postsynaptic effects of α₁-adrenoceptor agonist on HN excitability in Mecp2^−/Y mice. A: records of membrane potential, input resistance and spike frequency in response to current injection from an Mecp2^−/Y HN in control (top) and during PE (40 μM) application (bottom). B: frequency-time (f–t) relationship for the same neuron for the highest current intensities. C: frequency-current (f–I) relationship for the same neuron. D–F: bar graphs showing the effects of PE on membrane potential (D), input resistance (E), and spike frequency (F) of HNs from Mecp2^−/Y mice. *P < 0.05; **P < 0.01; n, number of neurons examined.

Effects of presynaptic α₁-adrenoceptor activation on glycinergic synaptic transmission in hypoglossal motoneurons of WT mice.

Inward spontaneous postsynaptic currents were studied in HNs with high concentrations of Cl⁻ (∼140 mM) applied to both pipette and bath solutions at a holding potential of −70 mV in voltage clamp (see experimental procedures). As shown in Fig. 4A, ionotropic glycine receptor-mediated sIPSCs were isolated by selective inhibitions of the GABA_A currents with 10–30 μM bicuculline, the non-N-methyl-d-aspartate (NMDA) currents with 20 μM CNQX, and the NMDA currents with 30 μM dl-AP5. The sIPSCs were completely suppressed when strychnine was applied (0.5–1 μM; n = 3; Fig. 4A).

Fig. 4.

Activation of α₁-adrenoceptor facilitated glycinergic synaptic transmission in the WT HNs. A: sample traces showing spontaneous inhibitory postsynaptic currents (sIPSCs) recorded in the presence of 30 μM bicuculline, 20 μM CNQX, and 30 μM d-AP5 (left). sIPSCs were completely abolished when perfused with bicuculline, CNQX, dl-AP5, and 1 μM of strychnine (right). B: sample traces showing sIPSCs recorded under control (top) and during bath application of PE (40 μM; bottom). Note that sIPSCs trace at large time scale obtained from the area indicated by 2 dashed lines. C and D: cumulative distribution of the interevent interval and amplitude curves obtained from the same neuron as in B. PE shifted the interevent interval curve (C) at left, but had no significant effect on the amplitude distribution curve (D). E and F: bar graphs summarizing the effects of PE on sIPSCs frequency and amplitude expressed as percentage of control ± SE in normal recording solution (E) and recording solution added GDP-β-S (F). *P < 0.05; **P < 0.01; n, number of neurons examined.

In the first series of experiments, we examined the effect of the highly selective α₁-adrenoceptor agonist PE on sIPSCs of WT neurons. The sIPSCs were recorded every 3–5 min. For control, we recorded sIPSCs for more than 10 min to confirm that the baseline was stable. The effect of PE was then bath applied for 3–5 min, followed by washout recording after another 10 min. For data analysis, only the last 3–5 min of results from each period were analyzed with 200–10,000 events. Only sIPSCs with short rising time (<1 ms) were used for data analysis. The amplitude of sIPSCs was averaged, and frequency was calculated using minianalysis program. The concentration of 40 μM was chosen as PE at 30–50 μM affects both GABA_A and glycine receptors mediated-responses in cerebellar and nucleus ambiguous neurons (9, 27). Bath application of PE increased the frequency (146 ± 16.8% of control; P < 0.01, Student's t-test) but had no significant effect on amplitude (109 ± 7.3% of control; P = 0.12, Student's t-test) of sIPSCs in eight HNs (Fig. 4, B–E). Since the α₁ receptor is expressed on postsynaptic membranes of several brainstem motoneurons including HNs (20, 52), we applied 1 mM GDP-β-S, a nonhydrolyzable GTP analog and G-protein inhibitor (15), to the pipette solution to block G-protein signaling in the postsynaptic neurons (17, 25, 30, 31, 66). The GDP-β-S was applied for >10 min before PE application. Under this condition, the PE-mediated facilitation on sIPSCs frequency was not reduced at all (189 ± 27.5% of control; P = 0.013, Student t-test) in six cells (Fig. 4F).

Subsequently, we tested whether the PE effect on sIPSCs depends on WT neuronal firing activity. We recorded mIPSCs in the presence of bicuculline, CNQX, dl-AP5, and 1 μM TTX in WT neurons (Fig. 5A). Bath application of PE also increased the frequency (156 ± 14.3% of control; P < 0.001, Student's t-test) but not the amplitude (108.6 ± 4.9% control; P = 0.26, Student's t-test) of mIPSCs in seven cells (Fig. 5D). These results thus indicate that activation of α₁-adrenoceptor facilitates glycinergic synaptic inputs in HNs and such an effect is likely to be mediated by enhancing spontaneous glycine release from presynaptic terminals.

Fig. 5.

α₁-Adrenoceptor agonist facilitated spiking-independent glycinergic synaptic transmission in the WT HNs. A: sample traces showing the effect of PE on mIPSCs recorded in the presence of CNQX, dl-AP5, bicuculline, and 1 μM TTX, under control (top) and during 40 μM PE application (bottom). B and C: cumulative distribution of the interevent interval and amplitude curves obtained from the same neuron as in A. PE shifted the interevent interval curve (B) at left but had no significant effect on the amplitude distribution curve (C). D: bar graphs summarizing the effect of PE on mIPSCs frequency and amplitude expressed as percentage of control ± SE. **P < 0.01; n, number of neurons examined.

Modulation of glycinergic synaptic transmission of HNs in Mecp2^−/Y mice.

The effects of PE on glycinergic sIPSCs were examined in the presence of bicuculline, CNQX and d-AP5. Bath application of PE (40 μM) increased the frequency of sIPSCs (Fig. 6, A–D) in eight HNs from 3 Mecp2^−/Y mice. Surprisingly, the effect of PE on sIPSC frequency was much stronger in Mecp2^−/Y mice than in WT (299.5 ± 47.6 vs. 146 ± 16.8%, respectively; P < 0.01, Mann-Whitney test). Unlike the WT neurons, PE also augmented significantly the amplitude of the glycinergic sIPSCs in Mecp2^−/Y mice (by 214.4 ± 44.3 vs. 109 ± 7.3% in WT; P < 0.05, Mann-Whitney test). Although the frequency effect is likely to be mediated by presynaptic mechanisms, the effect of PE on sIPSC amplitude could be both pre- or postsynaptic. For this reason, we employed GDP-β-S to block the postsynaptic α_1-receptor-mediated signaling. With GDP-β-S, the frequency and amplitude were 253 ± 39% (P < 0.01, compared with control) and 102 ± 16% (P > 0.05), respectively (n = 5 cells, Student's t-test; Fig. 6E), suggesting that postsynaptic mechanisms are involved in the modulation of sIPSC amplitude.

Fig. 6.

α₁-Adrenoceptor agonist facilitated glycinergic synaptic transmission in HNs of Mecp2^−/Y mice. A, sample traces showing sIPSCs recorded under control (top) and during PE application (bottom). B and C: cumulative distribution of the interevent interval and amplitude analyzed from the same neuron as in A. PE shifted interevent interval at left (B) and amplitude at right (C). D: bar graph summarizing the effect of PE on sIPSCs frequency and amplitude expressed as percentage of control ± SE in Mecp2^−/Y HNs. E: bar graph showing the effect of PE on frequency and amplitude of sIPSCs recorded with electrodes filled with GDP-β-S. *P < 0.05; **P < 0.01; n, number of neurons examined.

The effects of PE on mIPSCs in Mecp2^−/Y mice were also tested in five neurons in the presence AMPA, kainite, NMDA, GABA_A receptors antagonists, and TTX (1 μM). The frequency and amplitude of mIPSCs were 143 ± 15.9% (P < 0.05, Student's t-test) and 119.6 ± 15.1% of control (P = 0.2, Student's t-test), respectively (Fig. 7, A–D). These were not significantly different from those in the WT neurons (P = 0.29 and 0.9 for amplitude and frequency, respectively (Mann-Whitney test). Thus these results suggest that the modulation of glycinergic synaptic transmission by α₁-adrenoceptors is augmented in Mecp2^−/Y mice, and the augmentations are spiking dependent involving likely both pre- and postsynaptic mechanisms.

Fig. 7.

α₁-Adrenoceptor facilitated spiking-independent glycinergic transmission in Mecp2^−/Y. A: sample traces showing the effect of PE on mIPSCs recorded in the of presence 1 μM TTX, under control (left), and during 40 μM of PE application (right). B and C: cumulative distribution of the interevent interval and amplitude curves from same neuron as in A. Interevent interval (B) but not the amplitude (C) curve was shifted at left. D: bar graphs summarizing the effect of PE on mIPSCs frequency and amplitude expressed as percentage of control ± SE from Mecp2-null mice. *P < 0.05; n, number of neurons examined.

DISCUSSION

The present study has shown that presynaptic α₁-adrenoceptor activation facilitates glycinergic transmission and excites HNs, and this presynaptic facilitation is altered in Mecp2-null mice through both pre- and postsynaptic mechanisms. In addition, our data demonstrated that postsynaptic modulation HNs by α₁-adrenoceptors involving the G protein-dependent and barium-sensitive K⁺ channels.

Postsynaptic α₁-adrenoceptor modulation of HN excitability in WT mice.

Previous studies have demonstrated that activation of α₁-adrenoceptors produces depolarization associated with an increase in input resistance in cranial and spinal motoneurons (16, 20, 37, 49, 52, 63, 68). These excitatory effects are postsynaptic and are not derived from network drive since they are retained in the presence of TTX (52). Consistent with these findings (52), we found that the PE-induced depolarization and increase in input resistance persist in the presence of TTX. Also, the activation of postsynaptic α₁-adrenoceptors is critical, as the PE-mediated depolarization, increase in input resistance, and spiking were all blocked when the recording pipette was filled with GDP-β-S. The complete blockade seems to be an overestimate, as it is very difficult to inhibit totally the G protein-dependent signaling by intracellular dialysis of GDP-β-S through the recording pipette. Such an effect seems to be a result of both the GDP-β-S and the glycinergic IPSCs that are augmented by the presynaptic α₁-adrenoceptors.

Several mechanisms have been described for PE mediated-excitatory effects in the central nervous system (52). The PE-induced depolarization is due to inhibition of K⁺ currents in facial motoneurons (37) and Ba²⁺-sensitive K⁺ currents in juvenile and young adult HNs (49). In addition, the PE-mediated excitation of HNs also involves the activation of Ba²⁺-insensitive Na⁺ currents (49) and inhibition of TASK channels (60). In the present study, we have shown that the PE effects on HN membrane excitability are attributable to the inhibition of K⁺ currents that are sensitive to GDP-β-S and micromolar concentration of Ba²⁺, suggesting that the G-protein-coupled and Ba²⁺-sensitive K⁺ channels play a role in the postsynaptic effect of PE.

Presynaptic α₁-adrenergic modulation of HN glycinergic transmission in WT mice.

We have shown that PE modulates glycinergic sIPSCs in HNs. It has been reported that in several brain regions NE facilitates GABA and glycinergic synaptic transmissions by activation of presynaptic α₁-adrenoceptors (3, 9, 23, 26, 27). Indeed, our results indicate that PE augments frequency of the sIPSCs and mIPSCs only, without changing their amplitude. Consistent with these results, the PE-mediated facilitation of sIPSCs was not changed when postsynaptic metabotropic α₁-adrenoceptors were blocked by GDP-β-S. Thus our results suggest that activation of presynaptic α₁-adrenoceptors involved in the facilitation of an action potential-dependent glycinergic transmission at HNs.

Pre- and postsynaptic modulations of HNs in Mecp2^−/Y mice.

RTT is a neurodevelopmental disorder caused by Mecp2 gene mutations (21, 58). Defects in the brainstem NE system have been suggested as an underlying cause for the breathing irregularities of the Mecp2^−/Y mice (65, 72). Potential mechanisms for defects in the NE system have been reported, such as inadequate tyrosine hydroxylase and dopamine β-hydroxylase expressions (54, 61, 71), increased neuronal excitability (61, 71, 73), disrupted CO₂ chemosensitivity (72), and defect of GABAergic synaptic transmission 3 (30). In the present study, we have found that the PE-dependent facilitation of glycinergic synaptic transmission is enhanced in Mecp2^−/Y mice. PE augments both frequency and amplitude of sIPSCs in Mecp2^−/Y neurons, both of which are significantly greater than in WT neurons. The enhanced ISPC frequency appears presynaptic, while the augmentation of IPSC amplitude is likely postsynaptic since it was blocked when electrodes were filled with GDP-β-S.

HNs neurons receive both GABAergic and glycinergic synaptic inputs (7, 38, 47, 69). A defect in GABAergic transmission has been reported in ventrolateral medulla of Mecp2^−/Y mice (12, 30, 41). Similar findings have also been made in hippocampal (70) and LC neurons (26). Interestingly, the glycinergic system in ventrolateral medulla remains normal in Mecp2^−/Y mice (41). Data from the present study suggest that the adrenoceptor-mediated facilitation of glycinergic transmission indeed is stronger in Mecp2^−/Y mice than in the WT. Our data suggest that both pre- and postsynaptic mechanisms are involved in the facilitation of glycinergic synaptic inputs in HNs.

It is possible that the enhanced glycinergic transmission by presynaptic α₁-adrenoceptors serves as a compensatory mechanism. GABAergic synaptic currents are dysfunctional in several types of neurons of the Mecp2^−/Y mice (12, 30, 41). Thus enhanced glycinergic receptor activity may reflect compensation for reduced GABAergic activity. The Mecp2^−/Y mice show a marked deficiency in NE content and function (29, 61, 65). For instance, the NE content is lower in Mecp2^−/Y mice than the wild type in brainstem (61, 65) and medulla (65). It was hypothesized that breathing disturbances in Mecp2^−/Y mice may be due to a deficiency in noradrenergic modulation of the respiratory network (65). The increased sensitivity of presynaptic glycinergic receptors to α₁-adrenoceptor agonist may allow the Mecp2^−/Y HNs to maintain inhibitory synaptic inputs with limited NE in the presynaptic terminals. Therefore, the augmentation of glycinergic transmission by α₁-adrenoceptors may be a compensatory mechanism in the Mecp2^−/Y HNs. We speculate that when the Mecp2 knockout affects the body, the system may try to respond to minimize the defect. Thus our findings open up the possibility for the development of pharmacotherapies to increase the glycinergic transmission by α₁-adrenoceptor agonists for the treatment of motor disorders of RTT.

In conclusion, activation of α₁-adrenoceptor facilitates glycinergic synaptic transmission and excites HNs. These effects are mediated by pre- and postsynaptic mechanisms. The latter effect is likely to be mediated by an inhibition of the G protein-coupled K⁺ channels. The pre- and postsynaptic modulations of HNs by α₁-adrenoceptors are not only retained in Mecp2^−/Y mice but also markedly enhanced, which may be a compensation mechanism for the deficiencies in NE modulatory system and GABAergic synaptic transmission. These findings are encouraging, as they suggest that endogenous compensation mechanisms, and others that may exist as well, for an alleviation of the defects in the central nervous system caused by Mecp2 gene disruption provides a new hope for therapeutical interventions to RTT.

GRANTS

This work was supported by National Institute of Neurological Disorders and Stroke Grant 1R01-NS-073875 and the International Rett Syndrome Foundation.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: X.-T.J. and C.J. conception and design of research; X.-T.J., N.C., W.Z., and Z.W. performed experiments; X.-T.J. and X.J. analyzed data; X.-T.J. and C.J. interpreted results of experiments; X.-T.J. prepared figures; X.-T.J. drafted manuscript; X.-T.J., Z.W., and C.J. edited and revised manuscript; X.-T.J., Z.W., and C.J. approved final version of manuscript.