Inhibitory mechanisms control active expiration by limiting parafacial expiratory drive

Abstract

Activity of parafacial neurons that control active expiration is heavily dependent on tonic and CO₂/H⁺-dependent excitatory and inhibitory inputs from yet poorly defined sources. Contrary to the idea that CO₂/H⁺ disinhibits parafacial expiratory neurons, the recent work of J. D. Silva et al. (Silva JD, Oliveira LM, Souza FC, Moreira TS, Takakura AC. J Neurophysiol 123: 1933–1943, 2020) suggests that GABAergic raphe neurons preferentially limit expiratory activity during high CO₂. Here, I discuss these findings and propose a model where GABAergic raphe neurons function as CO₂/H⁺-dependent breaks on expiratory drive.

Breathing is composed of three phases, namely, inspiration, postinspiration, and expiration. Each of these phases is controlled by specialized but integrated neural networks that together produce a dynamic output necessary to meet highly variable metabolic demands. Importantly, to increase respiratory output during times of high metabolic demand, it is necessary to augment all phases of breathing including expiration. Under resting conditions, expiration is a passive process mediated primarily by elastic recoil of the lungs. However, during strenuous physical activity or in response to metabolic ques like hypoxia or hypercapnia, expiration becomes an active process involving recruitment of accessory, abdominal, and intercostal muscles, to facilitate compression of the thoracic cavity.

The ability to regulate breathing in response to CO₂/H⁺ is referred to as respiratory chemoreception. Several brain regions are thought to function as respiratory chemoreceptors including the nucleus of the solitary tract (NTS), the medullary raphe, and the retrotrapezoid nucleus (RTN). The RTN is located in the ventral parafacial region just medial to the parafacial respiratory group (pFRG) that regulates active expiration (AE) by sending excitatory projections to premotor neurons in the caudal ventral respiratory group (cVRG) that drive expiratory motor neurons (1, 2).

Evidence suggests that activity of the pFRG and consequently expiratory drive is dependent on excitatory and inhibitory synaptic drive from multiple levels of the respiratory circuit. For example, the pFRG receives CO₂/H⁺-dependent glutamatergic input from nearby chemosensitive RTN neurons (3), thus serving to couple expiratory activity to metabolic demand. Although one study found that blockade of glutamate receptors in the pFRG with kynurenic acid suppresses CO₂/H⁺-evoked AE (3), another reported that blocking ionotropic glutamate receptors in the pFRG did not eliminate CO₂/H⁺-evoked active expiration (4). Therefore, the extent to which glutamatergic signaling in the pFRG contributes to AE remains unclear. Excitatory drive from the RTN to the pFRG is also negatively regulated by cholinergic projections from the rostral pedunculopontine tegmental nucleus to the RTN, resulting in inhibition by activation of muscarinic cholinergic receptors M2/M4 (5).

During normal breathing, when AE is absent, activity of pFRG neurons is suppressed by inhibitory inputs. In rats, bilateral injections of bicuculine and strychnine into the pFRG disinhibit the region resulting in AE under normocapnic conditions (1, 5). Since AE is recruited in response to hypercapnia, it is possible that inhibitory input to parafacial expiratory neurons is withdrawn during exposure to high CO₂/H⁺. Although the source(s) of inhibitory input to parafacial expiratory neurons are not fully understood, a recent publication in the Journal of Neurophysiology describes a novel projection from medullary raphe regions to the pFRG, which appears to limit AE in conditions of high CO₂ (6).

Medullary raphe regions including the magnus (RMg) and obscurus (ROb) are important respiratory centers. Serotonergic neurons in these medullary raphe regions function as respiratory chemoreceptors by stimulating breathing in response to CO₂/H⁺. Moreover, a subset of GABAergic raphe neurons is inhibited by CO₂/H⁺ (7), and so conceivably may contribute to CO₂/H⁺-dependent disinhibition of parafacial expiratory neurons. Therefore, it was hypothesized that neurons in the RMg or ROb modulate expiratory activity by projecting to and regulating activity of parafacial expiratory neurons (6).

To test this possibility, it was first established that RMg and ROb neurons project to the parafacial region (6). Numerous serotonergic and GABAergic neurons in the RMg and ROb were back-labeled by fluorogold injections into the lateral parafacial region where parafacial expiratory neurons are located. Despite the spread of fluorogold injections labeling regions adjacent to the pFRG (6), a follow-up study using more precise targeting methods confirmed the existence of inhibitory projection from medullary raphe to the pFRG (8). Overall, these results provide an anatomical support for the possibility that serotonergic and GABAergic raphe neurons contribute to AE.

Next, to understand mechanisms by which medullary raphe might contribute to AE, inhibition of the RMg or ROb by injections of muscimol was tested (6). Inhibition of the RMg or ROb potentiates expiratory activity only under hypercapnic conditions, suggesting that the RMg and ROb provide a source of inhibition to the pFRG when respiratory drive is enhanced by high CO₂/H⁺ (6). Since injecting muscimol into the RMg and ROb silences both 5-HT and GABAergic populations projecting to the pFRG, parsing out individual functional roles of these populations in the context of AE was the logical next step. This was done by selectively depleting serotonin neurons of the RMg or ROb using a saporin anti-SERT strategy while preserving other neural populations including GABAergic neurons, thus allowing them to study contributions of GABAergic raphe neurons to AE in relative isolation. Like before, application of muscimol into the RMg or ROb in serotonin neuron-lesioned rats increases active expiratory activity only under high CO₂/H⁺ conditions. Lack of difference between expiratory responses of serotonin-lesioned and control rats implicates GABAergic but not serotonergic raphe neurons in AE. However, the serotonergic network in this region does modulate respiratory behavior. To minimize the potential off-target effects, optogenetic or chemogenetic manipulation of GABAergic raphe neurons would allow future work to interrogate functional mechanisms pertaining to AE without compromising serotonergic modulation.

Considering that some GABAergic raphe neurons are inhibited by CO₂/H⁺ (7) and since GABAergic raphe neurons project to the pFRG (6, 8), it is tempting to speculate that these raphe regions provide source of CO₂/H⁺-dependent disinhibition to the pFRGs. If this were the case, muscimol inhibition of the RMg or ROb is expected to potentiate AE under normocapnic conditions but not during high CO₂/H⁺ when GABAergic raphe neurons would be least active. However, pharmacological inhibition of the RMg or ROb potentiates AE under high CO₂/H⁺ and not normocapnia. Therefore, I hypothesize that GABAergic raphe neurons projecting to the pFRG are CO₂/H⁺ insensitive, given CO₂/H⁺-activated GABAergic raphe neurons have not been described, and likely limit AE through tonic inhibition of the pFRG. In this case, CO₂/H⁺-dependent excitatory drive must overcome this tonic inhibition to elicit AE.

The pFRG receives inhibitory input from sources other than the RMg and ROb (8). GABAergic neurons in the medial aspect of the NTS (mNTS) also project to the parafacial region and injection of muscimol to the mNTS increases AE in anesthetized rats under normocapnia but not during exposure to high CO₂/H⁺ (5, 8). Populations in the NTS have intrinsic chemosensitivity and while most are activated by CO₂/H⁺, a smaller population is CO₂/H⁺ inhibited (9). The neurochemical phenotype of CO₂/H⁺-inhibited NTS populations has yet to be determined; however, they likely recapitulate the neurochemistry of other CO₂/H⁺-inhibited populations and thus are also GABAergic (7). Therefore, the mNTS could be a source of CO₂/H⁺-dependent disinhibition to the pFRG and a key modulator of AE.

Peripheral inputs are also potential regulators of AE. Inhibitory pump cells in the NTS project to the RTN/pFRG region. This population receives information regarding lung volume through vagal afferents and can trigger expiratory activity in response to lung inflation, a behavior known as the Breuer–Hering reflex. Vagal denervation in vivo, which presumably eliminates excitatory drive to pump cells and diminishes inhibitory drive to the pFRG region, facilitates recruitment of AE in response to hypercapnia or hypoxia (10). However, the majority of mNTS neurons are spontaneously active (11) so it remains possible NTS pump cells projecting to the pFRG contribute to AE despite the animal models undergoing vagotomy (5, 6).

Based on the anatomical and physiological data currently available, control of active expiration appears to be achieved through a balancing act between excitatory and inhibitory inputs to the pFRG, which are dynamically regulated in response to metabolic demands. Figure 1 depicts a simplified functional model for AE in response to hypercapnia where either disinhibition or excitation passed a given threshold can drive expiratory activity. Disinhibition of the pFRG is sufficient to drive AE (1). However, whether disinhibition occurs as a result to changing CO₂/H⁺ levels remains to be definitively tested. Recently, inhibitory projections from the RTN region to the pFRG have been described in mice (8), highlighting the possibility that inhibitory populations in the RTN may function as a local source of CO₂/H⁺-dependent disinhibition to the pFRG. Consistent with this, the parafacial region contains CO₂/H⁺-inhibited neurons (12). Therefore, neurons in the RTN region could function as a source of both CO₂/H⁺-dependent excitation and disinhibition to pFRG to drive AE as a response to hypercapnia.

Figure 1.

Working model of the network basis for regulation of active expiration under control conditions and during hypercapnia. Under resting condition, the activity of pFRG neurons and active expiration is limited by inhibitory input from medullary raphe regions, the mNTS, and a dearth of excitatory input from other respiratory centers. Exposure to high CO₂ (i.e., hypercapnia) appears to inhibit activity of GABA/glycinergic neurons in the mNTS to favor disinhibition of the pFRG while at the same time CO₂/H⁺-activated RTN neurons increase excitatory drive to the pFRG. When the shift in excitatory to inhibitory synaptic input to the pFRG is sufficient to overcome tonic inhibitory input from medullary raphe regions, active expiration will ensue. cVRG, caudal ventral respiratory group; mNTS, medial portion of the nucleus of the solitary track; pFRG, parafacial respiratory group; PPTg, pedunculopontine tegmental nucleus; RMg, raphe magnus; ROb, raphe obscurus; RTN, retrotrapezoid nucleus.

Overall, parafacial expiratory activity is limited under control conditions by tonic inhibitory drive from GABAergic populations in nearby respiratory centers and the lack of excitatory input from chemoreceptor regions, particularly the RTN. In this microcircuit, metabolic stress would invert the roles of inhibitory and excitatory drive from the mNTS and RTN, likely by activating excitatory and silencing inhibitory projections to the pFRG. Although these mechanisms drive AE in response to high CO₂/H⁺, the pFRG simultaneously receives additional inhibitory input from GABAergic neurons in the RMg and ROb, presumably as a means of maintaining respiratory stability during high respiratory output. Though our anatomical understanding of the circuit responsible for AE has expanded in recent years, thanks in large part to the work of Drs. Ana C. Takakura and Silvia Pagliardini, an electrophysiological characterization of the populations involved in this behavior is still lacking. In particular, characterizing whether inhibitory populations in the RTN and mNTS projecting to the pFRG are silenced by CO₂/H⁺ will be an important step in understanding how AE is coupled to metabolic demand at a populational level.

Lastly, understanding the roles of inhibitory neurons in control of breathing may be relevant in the context of disease. For example, loss of inhibitory neural activity in a mouse model of Dravet syndrome (a type of epilepsy with a high mortality rate) resulted in unstable breathing and premature death (13). Therefore, understanding how inhibitory neurons contribute to respiratory control including at the level of parafacial expiratory neurons may provide novel insight into mechanisms underlying disordered breathing in disease.

GRANTS

This work was supported by National Science Foundation award no. 1702132 and by National Heart, Lung, and Blood Institute (NHLBI) Grant 13062834 and NHLBI F31 award no. 1F31HL156470-01.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

J.S-P. prepared figures; drafted manuscript; edited and revised manuscript; and approved final version of manuscript.