A muscarinic, GIRK channel-mediated inhibition of inspiratory-related XII nerve motor output emerges in early postnatal development in mice

Abstract

In neonatal rhythmic medullary slices, muscarinic acetylcholine receptor (mAChR) activation of hypoglossal (XII) motoneurons that innervate the tongue has a net excitatory effect on XII inspiratory motor output. Conversely, during rapid eye movement sleep in adult rodents, XII motoneurons experience a loss of excitability partly due to activation of mAChRs. This may be mediated by activation of G-protein-coupled inwardly rectifying potassium (GIRK) channels. Therefore, this study was designed to evaluate whether muscarinic modulation of XII inspiratory motor output in mouse rhythmic medullary slices includes GIRK channel-mediated inhibition and, if so, when this inhibitory mechanism emerges. Local pressure injection of the mAChR agonist muscarine potentiated inspiratory bursting by 150 ± 28% in postnatal day (P)0–P5 rhythmic medullary slice preparations. In the absence of muscarine, pharmacological GIRK channel block by Tertiapin-Q did not affect inspiratory burst parameters, whereas activation with ML297 decreased inspiratory burst area. Blocking GIRK channels by local preapplication of Tertiapin-Q revealed a developmental change in muscarinic modulation of inspiratory bursting. In P0–P2 rhythmic medullary slices, Tertiapin-Q preapplication had no significant effect on muscarinic potentiation of inspiratory bursting (a negligible 6% decrease). However, preapplication of Tertiapin-Q to P3–P5 rhythmic medullary slices caused a 19% increase in muscarinic potentiation of XII inspiratory burst amplitude. Immunofluorescence experiments revealed expression of GIRK 1 and 2 subunits and M1, M2, M3, and M5 mAChRs from P0 to P5. Overall, these data support that mechanisms underlying muscarinic modulation of inspiratory burst activity change postnatally and that potent GIRK-mediated inhibition described in adults emerges early in postnatal life.

NEW & NOTEWORTHY Muscarinic modulation of inspiratory bursting at hypoglossal motoneurons has a net excitatory effect in neonatal rhythmic medullary slice preparations and a net inhibitory effect in adult animals. We demonstrate that muscarinic modulation of inspiratory bursting undergoes maturational changes from postnatal days 0 to 5 that include emergence of an inhibitory component mediated by G-protein-coupled inwardly rectifying potassium channels after postnatal day 3 in neonatal mouse rhythmic medullary slice preparations.

INTRODUCTION

Airway resistance is determined by in part by the tone of airway muscles, including the muscles of the tongue. The hypoglossal (XII) motoneurons that innervate the main protruder muscle, genioglossus, of the tongue undergo state-dependent changes in excitability, including that their activity is markedly reduced during rapid eye movement (REM) sleep. The mechanisms that contribute to the state-dependent decrease in XII motoneuron excitability remain to be fully elucidated but recent data suggest a combination of noradrenergic disfacilitation as well as active cholinergic inhibition (1, 2). Although noradrenergic modulation of XII motoneurons has been well characterized (3–6), the mechanisms through which muscarinic modulation influences inspiratory bursting at XII motoneurons during development remain to be fully elucidated.

In neonatal rat in vitro rhythmic slice preparations, both exogenous application of muscarine (7) and endogenous cholinergic drive (8) potentiates XII inspiratory burst amplitude. In XII motoneurons, muscarine has multiple postsynaptic effects, including 1) depolarizing the voltage of half-activation of the hyperpolarization-activated, mixed cation current (I_h), which is hypothesized to facilitate XII motoneuron depolarization; 2) changes in the current-firing rate frequency relationship with depolarizing current injections attributed to inhibition of small conductance Ca²⁺-activated potassium channel (SK), and 3) decreased action potential height and increased action potential half width, suggestive of muscarine blocking voltage-gated sodium channels (7, 9). The first two mechanisms are hypothesized to increase XII motoneuron excitability, whereas the latter might decrease XII motoneuron excitability. In addition to blocking SK, muscarine appears to modulate other potassium (i.e., outward inhibitory) currents (9). In neonatal preparations, muscarine also acts presynaptically to depress inhibitory input (10) and inhibit glutamatergic (11) neurotransmission to XII motoneurons via an M2 receptor mechanism.

In adult rats in vivo, exogenous application of muscarinic agonists decreases output from the genioglossus muscle (12). Moreover, endogenous muscarinic modulation of XII motoneurons during REM sleep markedly depresses XII motoneuron excitability, via activation of what is most likely an inhibitory, GIRK channel-mediated mechanism (1) that may also include a contribution from presynaptic inhibition of glutamate signaling (13). Thus, the modulatory actions of muscarinic signaling on XII inspiratory motor output appear to switch completely during postnatal development: from excitation in brainstem slices from neonatal mice or rats to a powerful inhibition in adult rats. Whether these differences are due to different types of preparations and sources of modulation (i.e., exogenous muscarine application in slice preparations from neonatal rodents vs. endogenous release of acetylcholine during sleep in the adult preparation) or postnatal changes in the actions of muscarinic signaling on inspiratory motor output is not clear.

Here we test during early postnatal development whether muscarinic modulation of XII inspiratory motor output in vitro matures to include a GIRK-mediated inhibition. We used rhythmic medullary slice preparations from postnatal days (P)0–5 mice and compared the effects of muscarinic modulation on XII inspiratory burst behavior in the absence and presence of a blocker of GIRK channels.

We then performed a series of immunofluorescence experiments to begin to explore the molecular underpinnings of any developmental changes in muscarinic actions. There are five subtypes of muscarinic acetylcholine receptors (mAChRs) that mediate excitatory (M1, M3, and M5) or inhibitory (M2 and M4) actions. There are also four subtypes of GIRK channels (GIRK1–4) that could mediate the muscarinic-mediated inhibition of output observed in adults during sleep. We therefore used immunofluorescence to assess when during postnatal development these different proteins first appeared.

Despite the presence of immunolabeling for M1, M2, M3, M5 receptors and GIRK1 and GIRK2 subunits from P0, a muscarine-sensitive, GIRK-mediated inhibition of XII motor output was not evident until P3. These data are consistent with observations in other XII transmitter systems that the modulatory control over XII output is changing rapidly during the early postnatal period. The fact that mAChRs and GIRK channel subunits were present at P0 but GIRK-mediated inhibition was not seen until P3 suggests that the emergence of this inhibitory mechanism after birth reflects maturation of transduction mechanisms through which mAChRs modulate GIRK channels between P0 and P3.

METHODS

Ethical Approval

All experimental procedures were approved by the Midwestern University Institutional Animal Care and Use Committee and were carried out in compliance with the guidelines outlined by the NIH regarding the care and use of animals used for experimental procedures (14).

Animals and Preparations

CD-1 mice were obtained from Charles River to maintain a breeding colony at Midwestern University. Mice were housed at the Midwestern University’s Animal Facility with ad libitum access to food and water on a 12-h light/dark cycle. For experimental procedures, neonatal mice from postnatal days (P)0 to P5 of either sex were used.

Rhythmic medullary slice preparations were generated as described previously (4, 15). Briefly, neonatal mice were anesthetized via inhaled isoflurane and once withdrawal reflexes were absent, mice were immediately decerebrated. The neuraxis was extracted in a dissection chamber containing cold (5°C–10°C) artificial cerebral spinal fluid (in mM) (120 NaCl, 3 KCl, 1.25 NaH₂PO₄, 1 CaCl₂, 2 MgSO₄, 26 NaHCO₃, 20 d-glucose) that was bubbled with a gas mixture of 95% oxygen and 5% carbon dioxide (“carbogen”) to saturate the solution with oxygen and to achieve a pH of 7.4. The neuraxis was then glued to an agar block and sectioned on a vibratome (Leica VT1000S or Leica VT1200S, Leica Microsystems, Illinois). Serial 100- to 200-µm sections were collected from the rostral to caudal direction until the compact division of nucleus ambiguus and the rostral portions of the inferior olive appeared in transilluminated slices. At this point, a final, 600-µm slice was collected where the rostral border was 100–150 µm from the facial nucleus, and the caudal surface just captured the obex (700–750 µm from the caudal pole of the facial nucleus). This slice contained the preBötC, rostral ventral respiratory group, most of the XII motor nucleus and the rostral XII nerve rootlets (15). Since these slices did not encroach on the rostral border of preBötC, rhythmic output was typically more robust, which permitted us to test pharmacologic manipulation of GIRK channels and muscarinic modulation on XII motoneuron inspiratory output. The slice was transferred to a sylgard-lined recording chamber (∼10 mL volume) and pinned with the caudal surface up. The slice was superfused with artificial cerebral spinal fluid using a peristaltic pump to permit recirculation at a rate of 10–12 mL/min, at room temperature (21°C–22°C, bath application experiments for Tertiapin-Q or ML297) or warmed to 24°C–27°C (all local application experiments), bubbled with carbogen. The aCSF composition was as above, except that extracellular potassium concentration was increased from 3 to 9 mM for experiments. Slices were allowed to equilibrate in the recording chamber for at least 30 min before experimental manipulation. Elevated potassium concentrations are not required for inspiratory bursting but permit stable inspiratory bursting behavior for several hours required to complete the experimental protocols outlined herein (16, 17).

Nerve Recordings

A XII nerve rootlet(s) was captured by a glass suction electrode (ID 70–90 um) to record inspiratory-related bursting output from the rhythmic slice preparation. XII nerve rootlet activity was amplified (10,000×), band-pass filtered (300 Hz to 1 kHz, Model 1700 Differential AC Amplifier, AM Systems, Carlsborg, WA), sampled at 2 kHz, and displayed on a computer monitor where it was full-wave rectified and integrated with a moving averager (τ = 50 ms) using a PowerLab data acquisition system and software for integration (Powerlab 16/35, ADInstruments, Colorado) and LabChart software (Version 8, ADInstruments) or a Digidata and Axoscope 9.2 software (pCLAMP Suite, Molecular Devices, Sunnyvale, CA) using a CWE moving averager (CWE Incorporated, Pennsylvania) and saved using a Digidata 1322 A/D board (Molecular Devices). Data were analyzed offline using LabChart 8 software including the Peak Analysis plug-in or Clampfit 9.2 software (pClamp Suite) and collated in Microsoft Excel.

Immunofluorescence

Neonatal mice (P0–P5) were deeply anesthetized with isoflurane. Once withdrawal reflexes were confirmed to be absent, the chest cavity was opened and the left ventricle was punctured with a butterfly needle and the right atrium was nicked to permit transcardiac perfusion (1–2 mL/min) of cold 4% paraformaldehyde (PFA) in phosphobuffered (PB) saline. Brains were subsequently dissected and placed into 4% PFA in PB saline overnight.

Brains were then washed three times with PB saline (5 min each) and placed into a saturated sucrose solution. Once all water was exchanged for the sucrose solution (indicated by the brains sinking to the bottom of the conical tube), they were removed, embedded in optimum cutting temperature compound (OCT, Tissue-Tek), and frozen using a dry ice-liquid nitrogen slurry for sectioning on a cryostat (Leica CM 1860, Leica Biosystems, Deer Park, Il). Brains were sectioned in a caudal to rostral direction at −18°C to −20°C until the XII nucleus was observed. Twenty-micrometer slices were collected into 96-well plates and stored in PB saline until processed.

Immunofluorescence was performed on free-floating sections using antisera against one of the following targets at the indicated dilutions: M1 (anti-CHRM1, Cat. No. AMR-001, 1:200, Alomone Labs, Jerusalem, Israel), M2 (anti-CHRM2, Cat. No. AMR-002, 1:200, Alomone), M3 (anti-CHRM3, Cat. No. AMR-006, 1:200, Alomone), M5 (anti-CHRM5, Cat. No. AMR-005, 1:200, Alomone), GIRK1 (anti-GIRK1, Cat. No. APC-005, 1:200, Alomone), GIRK2 (anti-GIRK2, Cat. No. APC-006, 1:300, Alomone), GIRK3 (anti-GIRK3, Cat. No. APC-038, 1:300, Alomone), GIRK4 (anti-KCNJ5, Cat. No. APC-027, 1:200, Alomone). Choline acetyltransferase (anti-ChAT, Cat. No. AB144, 1:500, Chemicon) was coapplied and used as a positive control to identify XII motoneurons. All immunofluorescence experiments involved double-labeling of one mAChR or GIRK target and ChAT.

Briefly, to limit nonspecific labeling and to increase antibody penetration, slices were incubated for 60 min in PB saline containing 10% normal donkey serum (NDS) (Millipore 566460), 0.05% Triton-X, pH 7.3–7.4. PB saline containing primary ABs (either a muscarinic or GIRK target plus anti-ChAT) were applied overnight (14–16 h) at room temperature on a rocker plate. Slices were then washed three times with PB saline (5 min each) and then incubated in the secondary ABs at 1:200 (AF488 AffiniPure donkey anti-rabbit IgG, Cat. No. 711-545-152; Cy5 AffiniPure donkey anti-goat IgG, Cat. No. 705-175-147, Jackson ImmunoResearch Inc., West Grove, PA) in PB saline and 1% NDS on a rocker plate for 2 h. Slices were then washed an additional three times in PB saline (5 min each) and mounted using VECTASHIELD HardSet antifade mounting medium with DAPI (Vector Laboratories), coverslipped, and sealed with nail polish. Slides were imaged using a confocal microscope (Leica DMI 4000, Leica Biosystems) using both ACS APO ×10 (NA 0.3) and ×63 (NA 1.3) objectives with the following parameters and lasers: gain 800 V; smart offset −0.9%; laser intensity 405 nm = 20%, 488 nm = 20%, and 635 nm = 20%. Images were stitched together using automated Smooth stitching function from LASX software and adjusted for brightness and contrast using Fiji (18).

Drugs and Their Application

Muscarine (muscarine chloride hydrate, Sigma-Aldrich, Missouri), a mAChR agonist, and Tertiapin-Q (Tocris, Bio-Techne Corporation, Minnesota), a GIRK channel antagonist, were prepared as concentrated stock solutions in water (stock concentrations: Tertiapin-Q, 100 µm, muscarine, 1 mM). ML297 (Tocris), a GIRK channel antagonist, was prepared as a concentrated stock solution (200 mM stock concentration) in dimethyl sulfoxide (DMSO). Drugs were diluted with aCSF containing 9 mM K⁺ to the final concentration for bath or local application. The final concentration of DMSO did not exceed 0.005%. Previous data have established that DMSO concentrations up to 0.5% did not affect network inspiratory burst output (19) and DMSO up to 0.1% in the bath did not affect XII motoneuron membrane properties (7).

For bath application, drugs were added to the circulating solution and allowed to equilibrate for 10 min before the effects were evaluated. Due to the rapid dilution of locally applied drugs into the tissue, drugs applied locally were applied at a 10-fold higher concentration than in the bath, which has previously been shown to have an equivalent effect (7, 20). Drug injection pipettes were pulled from triple barrel borosilicate glass (Cat. No. 3B120F-6, World Precision Instruments, Florida) and broken back to an OD diameter of 4–6 µm/barrel. Microinjection of drugs was accomplished via pressure ejection using a picospritzer (Picospritzer III, Parker Hannifin, Ohio). The drug pipette was targeted to the ventrolateral portion of the XII nucleus, which tends to be the location of XII motoneurons that innervate the genioglossus muscle (21) and have stronger inspiratory modulation (5).

Data Analysis and Statistical Comparisons

Inspiratory burst amplitude is reported relative to control burst amplitude and expressed as means ± SD. Absolute values of inspiratory burst amplitude do not have biological meaning. For bath application experiments, inspiratory burst amplitude, burst area, and period during control were calculated over a 5-min period immediately preceding the drug application. During drug application, inspiratory burst amplitude was calculated as the average burst amplitude over a 5-min period after the drug had circulated for 10 min. For local application experiments, control burst amplitude was similarly determined as the average burst amplitude over a 5-min recording period before the subsequent drug application. During the local drug application, inspiratory burst amplitude was calculated as a five-point rolling average, which had the effect of smoothing the variance in inspiratory burst amplitude. Inspiratory burst period and area are reported in absolute terms as well as relative to the control values as the means ± SD. For local drug application protocols, data were excluded from analysis if drug injection into the XII caused more than a 20% change in burst period as a change in period indicates drug spread to the preBötC, which could indirectly change inspiratory burst pattern.

Statistical comparisons were made on raw data for burst period and on normalized data for burst amplitude and for burst area. Comparison of means from two groups was performed using paired t tests. For data with three or more groups, means were compared using a one-way analysis of variance with repeated-measures (no missing values) or a mixed-effects analysis (when some values were missing) followed by a Tukey’s multiple comparisons post hoc test (GraphPad Prism 8 or 9). Statistical significance was set at P ≤ 0.05.

RESULTS

To evaluate whether there is a significant contribution of GIRK channel activity in shaping baseline XII inspiratory motor bursting, we applied the GIRK antagonist, Tertiapin-Q, to the solution perfusing rhythmically active brainstem slices isolated from mice between P0 and P5. As shown for an example P4 mouse slice, in some slices, there was a hint of an excitatory effect in the small upward shift in baseline ∼60 s after drug application indicative of a small increase in baseline tonic XII motoneuron activity (i.e., activity that is not coordinated with inspiratory activity) shortly after Tertiapin-Q application. This baseline shift was not commonly observed. In addition, there were no obvious effects of Tertiapin-Q on baseline inspiratory burst amplitude neither area nor period at 10 nM for the P4 slice shown in Fig. 1A. This example is consistent with the group data from 10 preparations where bath application of Tertiapin-Q (10 µM) unexpectedly did not affect inspiratory burst amplitude (98 ± 9% of control amplitude, Fig. 1B, P = 0.37, n = 10), inspiratory burst area (97 ± 22% of control amplitude, Fig. 1B, P = 0.71, n = 9), or inspiratory burst period (15.7 ± 6.3 s vs. 16.6 ± 5.9 s, Fig. 1B, P = 0.09, n = 10). In this set of Tertiapin-Q bath application experiments, there was no obvious difference in effect with postnatal age (in contrast to the effects of Tertiapin-Q and muscarine application, see Fig. 3 and discussion).

Figure 1.

Effects of blocking GIRK channels with Tertiapin-Q (TQ) on inspiratory bursting in the rhythmic slice preparation. A: rectified, integrated recordings of XII nerve activity from a P4 slice showing inspiratory burst behavior during bath application of Tertiapin-Q (10 nM, arrowhead denotes addition of Tertiapin-Q to the perfusion solution) and at an expanded time scale as indicated at the arrowheads (A1 and A2). B: group data showing relative changes in XII burst amplitude, burst period, and burst area during bath application of Tertiapin-Q to the perfusion solution [10 nM, n = 10 (amplitude and period), n = 9 (area), P > 0.05, paired t test]. C: rectified, integrated recordings of XII nerve activity from a P5 slice showing inspiratory burst behavior during local application of Tertiapin-Q (100 nM, 60 s) to the XII nucleus and at an expanded time scale as indicated at the arrowheads (C1 and C2). D: group data showing relative changes in XII burst amplitude, burst period, and burst area during local application of Tertiapin-Q to the XII nucleus (100 nM, 60 s, n = 7, P > 0.05, paired t test). Box and whisker plot: medial and interquartile range with 9th and 91st confidence intervals, individual experiments: gray circles. P, postnatal day. The number (n) reported for each experiment reflects the number of animals used.

Bath application will influence all GIRK channels in the brainstem slice, including the preBötC, which could obscure or incorrectly ascribe (e.g., the minor shift in baseline activity in Fig. 1A could derive from a premotoneuron pool, the preBötC, etc.) a role of GIRK channels at XII motoneurons. To minimize this potential confounder and more directly assess the effects of GIRK channel block on XII motoneuron activity, we locally applied Tertiapin-Q using pressure microinjection directly within the XII nucleus. In the example recording from a P5 brainstem slice in Fig. 1C, inspiratory burst amplitude was 105% of baseline, inspiratory burst area was 104% of baseline, and burst period was 103% of baseline in minutes immediately following local application of Tertiapin-Q (100 nM) into the XII nucleus. Similarly, and unexpectedly (see discussion), group data showed that local application of Tertiapin-Q had no effect on inspiratory burst amplitude or inspiratory burst area. In the first 5 min after Tertiapin-Q application, burst amplitude averaged 106 ± 5% of baseline (Fig. 1D, P = 0.32, n = 7) and burst area averaged 107 ± 9% of baseline (Fig. 1D, P = 0.08, n = 7). In this set of Tertiapin-Q local application experiments, there was no obvious difference in effect with postnatal age (in contrast to the effects of Tertiapin-Q and muscarine application, see Fig. 3 and discussion).

Figure 3.

Effects of muscarinic modulation on inspiratory bursting including a contribution of GIRK channels in the rhythmic slice preparation. A: rectified, integrated recordings of XII nerve activity showing inspiratory burst behavior during local application of muscarine (mAChR agonist, 100 µM) alone (left trace) and then following local application of Tertiapin-Q (TQ; 100 nM, red line) followed by muscarine (black line, 100 µM, right trace) in a P1 rhythmic medullary slice. Rectified, integrated recordings of XII nerve activity shown at an expanded time scale as indicated at the arrowheads (A1–A4). B: rectified, integrated recordings of XII nerve activity showing inspiratory burst behavior during local application of muscarine (mAChR agonist, 100 µM) alone (left trace) and then following local application of Tertiapin-Q (100 nM, red line) followed by muscarine (black line, 100 µM, right trace) in a P5 rhythmic medullary slice. Rectified, integrated recordings of XII nerve activity shown at an expanded time scale as indicated at the arrowheads (B1–B4). C: group data showing the potentiation of XII inspiratory peak burst amplitude mediated by muscarine alone (100 µM, control conditions; black circles with line of best fit, linear regression equation, y = 8.529x + 136.4, r² = 0.1477, P > 0.05) and after preapplication of Tertiapin-Q (100 nM, gray squares with line of best fit, linear regression equation, y = 15.65x + 123.3, r² = 0.2916, P < 0.05) in rhythmic medullary slice preparations as a function of postnatal age between P0 and P5 (n = 21). Group data showing the potentiation of XII inspiratory peak burst amplitude mediated by muscarine alone (100 µM) and after preapplication of Tertiapin-Q (100 nM) in rhythmic medullary slice preparations from P0 to P2 (D, n = 11, P > 0.05, paired t test) and from P3 to P5 mice (E, n = 10, *P < 0.05, repeated-measured analysis of variance with Tukey’s multiple comparisons post hoc test). Box and whisker plot: medial and interquartile range with 9th and 91st confidence intervals, individual experiments gray circles and lines. P, postnatal day. The number (n) reported for each experiment reflects the number of animals used.

The lack of an effect of Tertiapin-Q on baseline inspiratory activity could simply reflect that Tertiapin-Q-sensitive GIRK channels are closed under basal conditions in the rhythmic slice. We therefore tested whether activation of GIRK channels with the agonist ML297 could modify baseline activity. As shown for an exemplary P1 brainstem slice (Fig. 2A), bath application of ML297 (10 µM) was without effect on burst amplitude or burst period but decreased burst area. After 10 min in ML297, inspiratory burst amplitude was 101% of baseline and burst period was 98% of baseline whereas inspiratory burst area was 87% of baseline. Similarly, and unexpectedly (see discussion), in eight preparations, ML297 had no significant effect on inspiratory burst amplitude (96 ± 8% of control amplitude, Fig. 2B, P = 0.28, n = 8) or inspiratory burst period (11.8 ± 4 s vs. 12.8 ± 4 s, control vs. bath ML297, Fig. 2B, P = 0.10, n = 8). ML297, however, did significantly decrease inspiratory burst area to 86 ± 8% of baseline (Fig. 2B, P = 0.005, n = 7). There was no obvious difference in ML297 effects with postnatal age. Attempts to locally apply ML297 into the XII nucleus at 10× the concentration used in the bath could not be completed because ML297 is not sufficiently soluble in aqueous solution to reach the necessary concentration of 100 µM for local application (it precipitated).

Figure 2.

Effects of activating GIRK channels with ML297 on inspiratory bursting in the rhythmic slice preparation. A: rectified, integrated recordings of XII nerve activity showing inspiratory burst behavior during bath application of ML297 (10 µM) and at an expanded time scale as indicated at the arrowheads (A1 and A2). Arrowhead indicates addition of ML297 to the perfusion solution. B: group data showing relative changes in XII burst amplitude, burst period, and burst area following bath application of ML297 [10 µM, n = 8 (amplitude and period), n = 7 (area), P > 0.05 for burst amplitude and burst period, *P < 0.05 for burst area, paired t test]. Box and whisker plot: medial and interquartile range with 9th and 91st confidence intervals, individual experiments: gray circles. The number (n) reported for each experiment reflects the number of animals used.

Our observations that inhibition or activation of GIRK channels had very minor effects on basal XII inspiratory burst behavior do not exclude a role for GIRK channels in modulating output. Additional conditions necessary for GIRK channel activation may not be present in vitro. Since our main objective was to assess whether GIRK channels are an effector of muscarinic signaling in XII motoneurons during early postnatal development, we assessed the effect of mAChR activation on XII motoneuron burst activity before and after inhibiting GIRK channels with Tertiapin-Q in slices from animals ranging in age from P0 to P5. The most obvious effect of mAChR activation in neonatal mice is excitation. Thus, the presence of a muscarine-sensitive inhibitory GIRK mechanism predicts that blocking of GIRK channels with Tertiapin-Q will potentiate the muscarine-mediated excitation of XII inspiratory output. In addition, if a muscarine-sensitive inhibitory GIRK mechanism emerges or increases during postnatal development, we predict that the Tertiapin-Q potentiation of the muscarinic excitation will emerge at some point or increase postnatally (i.e., the potentiation of XII burst amplitude by Tertiapin-Q, which corresponds to a muscarine-sensitive, GIRK channel-mediated inhibition, will increase with development).

As shown in the example traces from a P1 and P5 slice in Fig. 3, A and B (left traces), we first measured the effects on XII inspiratory burst behavior of locally applying muscarine (30 s, 100 µM) into the XII nucleus. These effects were compared with a second injection 15 min later of muscarine (30 s, 100 µM) that was applied immediately after a 60-s preapplication of Tertiapin-Q (Fig. 3, A and B, right traces). To first assess developmental changes in the muscarinic potentiation of XII inspiratory burst amplitude and any Tertiapin-Q enhancement of the muscarinic potentiation of burst amplitude, we plotted as a function of postnatal age the peak potentiation of burst amplitude evoked by muscarine alone and in the presence of Tertiapin-Q (Fig. 3C). Under control conditions (i.e., no Tertiapin-Q), the relationship between muscarinic potentiation and postnatal age was described by the linear regression equation, y = 8.529x + 136.4 (r² = 0.1477). The slope of this relationship was not significantly different from zero (P = 0.09), although there is a trend for an increase in the muscarinic potentiation of XII burst amplitude between P0 and P5. In the presence of Tertiapin-Q, however, the slope of the relationship between the potentiation of XII inspiratory burst amplitude by muscarine and age between P0 and P5 was significantly greater than zero (equation: y = 15.65x + 123.3, r² = 0.2916, P = 0.012), indicating a positive relationship between the Tertiapin-Q-mediated potentiation of muscarinic actions and development. To test whether the two highest data points at P4 and P5 (233% and 371%; muscarinic potentiation of inspiratory bursting after Tertiapin-Q application) had an outsized effect on the regression analysis, we removed these two data points and reran the regression analysis. The slope remained statistically significantly greater than zero (y = 8.163x + 130.4, r² = 0.3775, P = 0.0051).

To assess the degree to which this Tertiapin-Q potentiation of the muscarine effect changed developmentally, we examined the control and Tertiapin-Q responses for individual animals at each age. None of the animals aged P0–P2 showed a potentiating effect of Tertiapin-Q on the muscarinic potentiation of XII burst amplitude; burst amplitude for each animal in Tertiapin-Q and muscarine was the same or numerically less than in muscarine alone. However, a small numerical potentiation was apparent in some of the P3 slices. We therefore grouped data from P0 to P2 and from P3 to P5 for further analysis.

In preparations from mice P0–P2, there was no effect of preapplied Tertiapin-Q on the muscarinic potentiation of inspiratory burst amplitude. This is illustrated in the example trace in Fig. 3A from a P1 medullary slice preparation where the initial muscarinic potentiation of inspiratory bursting (125% of baseline inspiratory burst amplitude) was unchanged compared with the muscarinic potentiation of inspiratory bursting after GIRK channels were blocked with Tertiapin-Q (129% of baseline inspiratory burst amplitude (Fig. 3A). In this example preparation, as with several preparations, there was a significant widening of the baseline during the muscarine application. This baseline widening could reflect increased activity (either recruitment or increased firing) of noninspiratory XII motoneurons and/or increased activity of inspiratory XII motoneurons, i.e., XII motoneurons that are inspiratory modulated that also increase their firing during the noninspiratory phase. The group data from nine preparations similarly showed that Tertiapin-Q had no effect. The muscarine-induced potentiation of burst amplitude was 145 ± 25% in control and 139 ± 22% after Tertiapin-Q (n = 11, P = 0.20, paired t test). In contrast, preapplication of Tertiapin-Q increased the muscarinic-mediated potentiation of inspiratory bursting in rhythmic slice preparations from mice P3 to P5. This is shown for a single P5 slice in Fig. 3B in which muscarine potentiated inspiratory burst amplitude by 155% when applied alone but 185% when applied after blocking GIRK channels with Tertiapin-Q. Group data confirmed this effect. Blocking GIRK channels with Tertiapin-Q caused an average 19% increase in the muscarinic potentiation of inspiratory burst amplitude from the initial response of 175% ± 54% to 194 ± 68% following Tertiapin-Q (n = 10, P = 0.047, repeated-measured analysis of variance with Tukey’s multiple comparisons post hoc test). We waited an additional 15 min and repeated the control muscarine (30 s, 100 µM) injection to assess whether the effects of Tertiapin-Q were reversible. The potentiation of XII inspiratory burst amplitude by muscarine alone 15 min after the Tertiapin-Q/muscarine trial of 212 ± 96% (n = 10) was the same as in Tertiapin-Q (P = 0.25, repeated-measured analysis of variance with Tukey’s multiple comparisons post hoc test).

Tertiapin-Q effects of binding GIRK1 subunits are long lasting (22), and this high binding affinity is a likely explanation for the lack of reversibility of the Tertiapin-Q effects on the muscarinic modulation of inspiratory bursting. Nevertheless, the possibility must also be considered that the potentiating actions of Tertiapin-Q were not due to the actions of Tertiapin-Q per se, but the cumulative actions of multiple muscarine doses; i.e., the muscarinic potentiation of XII burst amplitude increases with subsequent doses. To test this possibility, in a separate set of experiments, we locally applied muscarine repeatedly (100 µM, 30 s) at 15 min intervals up to four times. The potentiation was very consistent between consecutive applications, averaging 151 ± 26% (n = 6), 163 ± 45% (n = 6), 154 ± 25% (n = 6), and 163 ± 28% (n = 4) for the first, second, third, and fourth applications, respectively (P = 0.25, mixed-effects analysis with Tukey’s multiple comparisons post hoc test, P2–P3, data not shown).

Previous data indicate that vehicle aCSF injection does not affect inspiratory burst amplitude (23). To confirm this, in a subset of five experiments, we compared the muscarinic potentiation of inspiratory bursting without a prior vehicle injection to the muscarinic potentiation of inspiratory bursting with a preceding 60-s vehicle aCSF injection. In these five experiments, muscarinic potentiation of XII inspiratory burst amplitude was similar under both conditions (192 ± 6% vs. 185 ± 55%, P = 0.53, paired t test).

These developmental changes in the ability of Tertiapin-Q to modulate the effects of muscarine on inspiratory bursting could be due to changes in the expression of mAChR subtype(s) that inhibits GIRK channels, a developmental increase in the expression of GIRK channels, or increases in the effectiveness of mAChRs to activate GIRK channels via maturation of intracellular signaling cascades. As an initial step into exploring the first two options, we identified which muscarinic receptors were present in the XII nucleus between P0 and P5 by using double-labeling, immunofluorescence approaches. We used choline acetyltransferase (ChAT) to label XII motoneurons and antisera against M1, M2, M3, and M5. We did not examine M4 receptor expression because neither protein nor mRNA for the M4 receptor subtype has been found in XII motoneurons (10, 24). Example immunofluorescence images are shown in Fig. 4, A–D, with immunolabeling for the mAChR subtype shown in the left, ChAT in the middle, and the overlay on the right. Micrographs for M1, M2, M3, and M5 immunolabeling in Fig. 4, A–D are from P3, P2, P2, and P4 mice, respectively. Labeling was confirmed from three mice for each target. Although we do not show images from P0 mice, the first point is that we confirmed immunolabeling for M2 and M3 receptor subtypes from P0 onward and M1 and M5 receptor subtypes from P1 onward (we did not test for M1 and M5 receptor subtypes in P0 XII motoneurons). Second, labeling for M1 was strongest, followed by M5. M3 labeling was weak. The signal for M2 receptor subtype was strong but, unlike M1, M3, and M5, it did not appear to be cytoplasmic. The margins of some XII motoneuron cell bodies appear to be delineated by M2 receptor immunolabeling. Labeling for M2 was also strong in the neuropil, suggesting a dendritic or presynaptic distribution.

Figure 4.

Immunofluorescence images showing double immunolabeling for ChAT (motoneuron marker) and muscarinic acetylcholine receptor (mAChR) subtypes, M1–M3 and M5 (A–D) or GIRK channel subunits 1 and 2 (E and F) in neonatal XII motoneurons. A: M1 AChR distribution (green), choline acetyltransferase (ChAT) (magenta), and overlay (gray tones indicate double-labeling) from P3 CD1 mouse. B: M2 AChR distribution (green), choline acetyltransferase (ChAT) (magenta), and overlay from P2 CD1 mouse. C: M3 AChR distribution (green), choline acetyltransferase (ChAT) (magenta), and overlay (white indicates overlay) from P2 CD1 mouse. D: M5 AChR distribution (green), choline acetyltransferase (ChAT) (magenta,) and overlay from P4 CD1 mouse. E: GIRK1 distribution (green), choline acetyltransferase (ChAT) (magenta), and overlay from P2 CD1 mouse. F: GIRK2 distribution (green), choline acetyltransferase (ChAT) (magenta), and overlay from P3 CD1 mouse. Scale bar, 50 µm. P, postnatal day.

We next assessed which of the GIRK channel subunits, GIRK1, GIRK2, GIRK3, and GIRK4, are expressed by XII motoneurons between P0 and P4. Images in Fig. 4, E and F, from a P2 and P3 mouse, respectively, show expression of both GIRK1 and GIRK2 by XII motoneurons. GIRK1 labeling is weakly present throughout the cytosol whereas GIRK2 labeling appears to be primarily on the nuclear membrane (confirmed by DAPI counterstaining, data not shown). Neither GIRK3 nor GIRK4 immunolabeling was found in XII motoneurons. These negative data were supportive with a positive control showing positive labeling for GIRK3 and GIRK4 in the cerebellum (data not shown). Labeling data were confirmed in three mice for each target which revealed that XII motoneurons express GIRK1 and GIRK2 from P0 onward.

DISCUSSION

The present results indicate that 1) muscarinic modulation of inspiratory bursting at XII motoneurons in brainstem slices from neonatal mice includes a GIRK channel-mediated inhibition that emerges during postnatal development after P3 and 2) mouse XII motoneurons express four mAChRs (M1, M2, M3, and M5) as well as GIRK1 and GIRK2 subunits from P0 to P5 with no obvious differences in immunolabeling over this same developmental window that the GIRK-mediated inhibition emerges. It therefore seems possible that emergence of this muscarine-sensitive, GIRK-mediated inhibitory mechanism that emerges at P3 and enhances the muscarinic potentiation of XII inspiratory output when blocked reflects maturation of the transduction mechanism through which mAChRs couple to GIRK channels.

Muscarine-Sensitive, GIRK-Mediated Inhibitory Mechanism Emerges Developmentally

Our pharmacological data in neonatal slices indicate that, compared with the potent inhibitory action of mAChRs attributed to activation of GIRK channels in adult rats, activation of the inhibitory GIRK channel mechanism has relatively minor impact on inspiratory burst amplitude in neonatal mice. Thus, it is very likely that the discrepancy between minor effects observed here in neonate compared with the potent inhibition in adult reflects that the responses are not yet mature. It is also possible that the discrepancy between our data in vitro neonates and adults reflect pharmacological limitations.

Tertiapin-Q has an IC₅₀ at GIRK1/2 heterotetramers of 5.4 nM (25) and requires <60 s to achieve that block (22, 26, 27), although Tertiapin-Q may not completely block GIRK currents (28). Our application of Tertiapin-Q at 10 nM (or the equivalent 100 nM locally) limited possible off-target actions but a consequence is that we may have underestimated the degree to which Tertiapin-Q inhibited the muscarine effect and, therefore, the degree of muscarinic modulation of GIRK channels.

Tertiapin-Q can block ROMK1/ROMK2 (K_ir1.1) channels (25, 26) and BK (KCa1.1) channels although it requires >60 s to block these latter channels (22, 26, 27, 29, 30). Both ROMK1/ROMK2 channels (expressed at low levels) (31) and BK channels are expressed in XII motoneurons (32), Thus, although we cannot exclude a contribution from either ROMK1/ROMK2 channels or BK channels, because Tertiapin-Q was applied for 60 s and at relatively low concentration, the Tertiapin-Q-mediated effects we observed are likely due its inhibition of GIRK channels.

Similarly, the only effect we observed of bath applied ML297 (10 µM) was a decrease in inspiratory burst area. These data were collected from P0 to P4 preparations, and we did not observe any correlation between burst amplitude in the presence of ML297 and postnatal age. ML297 acts directly to activate GIRK channels (33), with an EC₅₀ of ∼160 nM for GIRK1/2 subunits (33, 34). We were surprised that ML297 only affected XII burst area as it was applied at a concentration more than 50× the reported EC₅₀. ML297 can also partially inhibit the potassium voltage-gated channel subfamily H member 2 (hERG) channels at an IC₅₀ of ∼10 µM (33). mRNA for KCNH2 (the gene encoding for hERG channels) is expressed in XII motoneurons as described in the Allen Brain Atlas (35). Thus, if ML297 is activating GIRK channels and inhibiting other potassium channels such as hERG, this could explain the limited response to ML297. Finally, ML297 is soluble in nonpolar solvents and we were unable to dissolve it at concentrations needed for local pressure-injection experiments. Therefore, although ML297 has been used successfully in expression systems (36), or in thick spinal cord sections for patch clamp recordings (37), without a positive control to confirm that ML297 has penetrated into the slices and affected the respiratory network, the lack of effect on inspiratory burst amplitude in the present study is difficult to interpret.

Motoneuron Intrinsic and Modulatory Properties Undergo Dramatic Change Postnatally

XII motoneurons undergo anatomical and intrinsic excitability changes over the first postnatal weeks that may influence responses to neuromodulation (38, reviewed in Ref. 39). In addition, data from other modulatory systems indicate that the response to neuromodulators is not yet mature early in postnatal development, including maturational changes in modulation of inspiratory bursting (4, 40–42).

Muscarinic Modulation of XII Motoneuron Excitability Includes Both Excitatory and Inhibitory Components

In neonatal rhythmic brainstem preparations, muscarinic modulation evokes a net potentiation of inspiratory bursting (7, 8). In individual XII inspiratory motoneurons, muscarinic modulation can include activation of a tonic inward current, coupled with a decrease in synaptic inspiratory current magnitude (43). This decrease in inspiratory current magnitude is likely mediated by presynaptic M2 receptors (11). Activation of GIRK channels by postsynaptic mAChRs in animals older than P3 could also contribute as the reduction in input resistance associated with GIRK channel opening could partially shunt the inspiratory drive current. How this observed decrease in inspiratory current amplitude from individual XII motoneurons fits with the net increase in inspiratory burst amplitude with muscarinic modulation observed in neonatal mouse brainstem slices requires further investigation. The observed tonic inward current could help shift XII motoneurons toward firing threshold and thereby increase the total number of XII motoneurons exhibiting inspiratory modulation with activation of mAChRs. Certainly, the increased baseline activity observed in many recordings in the present results (Fig. 3) with inspiratory modulation may be due to this tonic inward current.

Conversely, in adult rats, blocking mAChRs revealed that endogenous muscarinic modulation has a net inhibitory effect on inspiratory activity, which was attributed to activation of GIRK channels (1), although presynaptic inhibition (13) may also contribute. Muscarine can also directly activate GABA_A receptors (44). Since the glycine, and likely GABA_A, receptor-mediated chloride currents shift from depolarizing to hyperpolarizing with maturation in XII motoneurons (45–47), muscarinic activation of GABA_A receptors could contribute to neonatal potentiation and adult inhibition of inspiratory bursting at XII motoneurons. This effect would be tempered, however, by presynaptic inhibition of GABAergic neurons by muscarine (10). It is unknown when during postnatal maturation muscarinic modulation develops a net inhibitory effect on inspiratory burst transmission.

GIRK Channel Modulation Emerges Developmentally

Data from other brain regions indicate that GIRK channels are more easily activated with postnatal maturation including activation of putative GIRK currents in 67% of neocortical pyramidal neurons at P3 and in 100% of neurons at P8–P10 (48). Overall, these authors concluded that GIRK channels were available in abundance but that the limiting factor is the ability of the modulator to activate GIRK channels (48). Similarly, the amplitude of evoked presumptive GIRK currents in hippocampal CA3 neurons increased from zero current at P0 to increasing current amplitude from P3 to P30 (49).

Our immunofluorescence data indicate that there are GIRK1/2 subunits present from P0 onward, but there was little effect on muscarinic modulation of GIRK channels with Tertiapin-Q until postnatal day 3 or later (Fig. 3). GIRK2 is required to form a functional heterotetrametric GIRK1/2 channel inserted into the plasma membrane (50). Our immunofluorescence data indicated GIRK2 labeling primarily associated with the nuclear membrane. Other inwardly rectifying potassium channels (e.g., K_ir2.2) are expressed on the nuclear membrane, although their physiologic function remains speculative (51). Data from the CA1 region of the hippocampus indicate that GIRK2 subunits are primarily located in the endoplasmic reticulum at P5 and distribution progressively shifted to the plasma membrane at P15 and P60 (52). If GIRK2 is not yet strongly expressed on the cellular membrane at P0–P5 in XII motoneurons, smaller GIRK currents would be likely to result.

GIRK currents may be activated by many different neuromodulators; the efficacy can vary with the neuromodulatory system and/or postnatal age (48, 53). Whether different modulators are more effective at activating GIRK-mediated currents in XII motoneurons in early postnatal life remains to be investigated.

Mechanisms Underlying Developmental Changes in Muscarinic Modulation of XII Motor Output and the Contribution of GIRK Channels

The maturational shift in muscarinic modulation of inspiratory bursting via GIRK channels may reflect that there are changes in 1) the expression level or distribution of muscarine receptor subtypes within the XII motoneuron membrane; 2) the expression level or trafficking of GIRK channels; or 3) intracellular signaling cascades through which mAChRs activate GIRK channels.

The five muscarine receptor subtypes can be subdivided into two broad categories. M2 and M4 couple to G_i or G_o G-proteins, whereas M1, M3, and M5 receptors typically couple to G_q G-proteins. Previous data suggest that neonatal XII rat motoneurons express M1, M2, and M5 whereas adult XII motoneurons additionally express M3 receptors (10, 54). Our immunofluorescence results indicated that there was expression of M1, M2, M3, and M5 mAChR subtypes from P0 onward (Fig. 4), although we did not quantify whether there were changes in expression level or cellular distribution across postnatal maturation. XII motoneurons express high levels of mAChRs compared with other brain regions, and show minor changes in expression pattern from P1 to P17 (55). Whether over that developmental period there are changes in specific mAChR expression patterns is not known. In spinal motoneurons, immunofluorescence data indicate a reorganization of M2 receptors such that they become more tightly clustered around synapses and particularly the large cholinergic terminals (i.e., C-boutons) over the first two postnatal weeks and into adulthood (56).

The G_βγ subunit of G_i canonically associated with M2 is the classical activator of GIRK channels (57–61). GIRK channels also require phosphatidylinositol 4,5-bisphosphate (PIP₂) as a cofactor to be activated (62), which is cleaved by activation of phospholipase C (PLC) by G_q G-proteins. Therefore, activation of G_q could decrease the GIRK current (58, 62). Indeed, muscarinic activation of sympathetic neurons led to a rapid 90%+ decrease in PIP₂ levels (63). Overall, GIRK-mediated currents evoked by mAChR activation in XII motoneurons will reflect the interplay between M2-mediated activation as well as M1/M3/M5-mediated reduction in GIRK currents. An increase in M2 receptor expression and/or a decrease in M1, M3, and/or M5 in XII motoneurons with postnatal maturation could explain the increased GIRK channel activation by mAChRs reported in adult rats (1).

Furthermore, any maturational shifts in the levels of G-proteins and proteins of the associated signaling cascades will affect the ability of mAChRs to activate GIRK channels. Although there do not seem to be maturational changes in G_q-protein (64) or G_βγ subunit genes (65), data support increased PLC activity with muscarinic activation with postnatal maturation (66) in other brain areas. Future research will need to evaluate whether there are changes in the expression or activity patterns of G-protein signaling cascade proteins in XII motoneurons across postnatal maturation that will influence the transduction of muscarine receptor activation to signaling effects.

Physiological Significance

There are several recently developed clinical trials testing the efficacy of a mixed noradrenergic agonist and muscarinic antagonist drug combination to treat symptoms of obstructive sleep apnea. Overall, the drugs have positively improved obstructive sleep apnea symptoms (67–73). The assumption from these studies is that these drugs are acting directly at XII motoneurons to increase their excitability. The diversity of muscarinic effects and changes with postnatal maturation discussed here, as well as sex differences (74), necessitate further mechanistic understanding of how muscarinic modulation can influence airway tone. This increased mechanistic understanding may facilitate development of more specific treatment options for sleep apnea patients. This is pertinent because drug side effects reported include urinary difficulties and dry mouth (68, 72), which may reduce enthusiasm for the current drug formulations.

In conclusion, our data establish that a small component of muscarinic modulation of inspiratory bursting at XII motoneurons in neonatal rhythmic brainstem slice preparations includes activation of GIRK channels and that this effect develops with postnatal maturation. We speculate that should these maturational changes continue to mature this will contribute to the differences reported in muscarinic modulation of inspiratory activity at XII motoneurons across postnatal maturation.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This research was supported by grants to A.L.R. from NIH/National Heart, Lung, and Blood Institute (R15HL148870) and to G.D.F. from Canadian Institutes of Health Research (CIHR; 130306) and Natural Sciences and Engineering Research Council (NSERC, 402532).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

G.D.F. and A.L.R. conceived and designed research; S.L.R., J.C.W., T.B., C.B., and A.L.R. performed experiments; S.L.R., J.C.W., T.B., C.B., and A.L.R. analyzed data; S.L.R., J.C.W., T.B., G.D.F., and A.L.R. interpreted results of experiments; S.L.R., J.C.W., and A.L.R. prepared figures; A.L.R. drafted manuscript; S.L.R., J.C.W., T.B., C.B., G.D.F., and A.L.R. edited and revised manuscript; S.L.R., J.C.W., T.B., C.B., G.D.F., and A.L.R. approved final version of manuscript.