An X on the Map for Sleep Apnea's Holy Grail: Drug Therapy.

Obstructive sleep apnea was supposed to be an easy target for drug therapies. Characterized decades ago, the pathophysiology of sleep apnea requires sleep state–dependent reductions in pharyngeal dilator activity (1, 2). Specifically, in obstructive sleep apnea, increased activity of upper airway dilator muscles across wakefulness ensures airway patency (3). In sleep, normal sleep state–dependent reductions in muscle tone result in relaxation of upper airway dilators, leading to collapse of the upper airway with resultant obstructive apneas (complete collapse) and hypopneas (partial collapse). One of the critical upper airway dilator muscles for airway patency is the genioglossus (tongue) muscle, which is innervated by the hypoglossal nerve (4). Armed with the understanding that sleep apnea involves a clear sleep state–dependent effect on this important dilator muscle and its nerve, researchers needed only to determine the neurochemical source of sleep-dependent inhibition of the hypoglossal nerve and then replace the neurochemical across sleep in the form of drug therapy, and we would obviate the need for cumbersome positive airway pressure machines. Excitatory receptor subtypes were then identified in the hypoglossal nucleus, and relevant agonists for the receptors were explored for effects on obstructive sleep-disordered breathing in animal models and in humans (5). Identified targets, however, largely had the untoward effect of promoting wakefulness and/or were no match for powerful sleep state–dependent hyperpolarization of hypoglossal motoneurons (5). Failure to identify a universally effective pharmacologic therapy for obstructive sleep apnea prompted development of a second mechanical therapy for sleep apnea, electric hypoglossal nerve stimulation (6, 7). Although this therapy has been effective in some individuals, it is expensive; it requires implantation of a foreign body and does not fully alleviate sleep-disordered breathing events in all individuals (8).

Enter chemogenetics. If the ideal receptor is not present on hypoglossal neurons to allow sufficient excitation across sleep to maintain a patent upper airway, it is now possible to place an exogenous receptor genetically onto target neurons. The ideal designer receptor would be one not found in the brain, or elsewhere, so that off-target effects of the drug could be avoided, and one that is receptive to an otherwise inert ligand that readily crosses the blood–brain barrier and has no active metabolites, which could also induce undesired effects. In this issue of the Journal, Fleury Curado and colleagues (pp. 102–110) tested the effectiveness of using a chemogenetic approach to treat upper airway flow limitation (9). Specifically, the researchers delivered an excitatory designer receptor by way of an adeno-associated viral vector injected into the tongue of a mouse model. The vector was then carried retrogradely to hypoglossal motoneurons and established as a new receptor (9). It is important to note that there is no widely available animal model of obstructive sleep apnea, and for this reason, the group used a model of obesity hypoventilation with inspiratory airflow limitation but without apnea (10). This is not an insignificant consideration, as it is entirely possible that complete airway collapse, as in an obstructive apnea, requires far greater hypoglossal activation to resolve the hysteresis resulting from increased surface tension at the site of collapse. With that caveat, the group was able to show quite convincingly that a viral vector could be injected into the tongue to deliver a designer receptor retrogradely to hypoglossal motoneurons and that administration of a selective agonist for the excitatory receptor could indeed markedly increase tonic muscle activity. In addition to increasing the genioglossal tonic and phasic muscle electromyographic activity, the research team demonstrated that administration of the designer ligand increased metabolic activity specifically in the genioglossus muscle in the model and increased upper airway cross-sectional area. Notably, the latter two effects were measured in anesthetized animals. Thus, this chemogenetic approach is potent enough to counter the significant inhibitory effects of anesthesia on muscles, lending further support for the feasibility of this approach in sleep apnea and its use also in patients with sleep apnea requiring anesthesia. Whether the designer receptor remains localized long term exclusively to hypoglossal motoneurons, or whether this may jump transynaptically to additional brain regions, will require longer-term studies and potentially refinement of the viral vector and/or receptor-ligand pair.

Nonetheless, there are clear advantages with this general direction as a therapeutic approach for obstructive sleep apnea. Unlike hypoglossal nerve stimulation, the vector carrying the target receptor to the desired motoneurons can be placed specifically in protractor components of the genioglossus muscle, and as a less invasive procedure and as a procedure potentially done under local anesthesia, exogenous delivery of the designer receptor could be done gradually over time, allowing titration of the therapy to successful resolution of obstructive sleep-disordered breathing. Additionally, this is a therapy that would not be limited to one nerve but in theory could be injected serially into additional pharyngeal muscles to augment the effect. Moreover, for such localized effects, very little virus is needed to deliver the designer receptor to its target motoneuron if delivered in this present study as retrograde transfer from muscles. A wonderful next step would be to bring back the Hendricks bulldog model of sleep apnea (11) for longer-term studies on safety, viral localization, and effectiveness of therapy on apneic events.

Given the overall promise of chemogenetics in general for treating a vast array of disorders, newer receptors and ligands are being developed. Recently, a chimeric ion channel was developed to conduct cations in response to the antismoking drug, varenicline, which readily crosses the blood–brain barrier and can be given to activate the designer receptor at concentrations far lower than that needed to activate nicotinic receptors (12). This approach may allow improved excitability of hypoglossal motoneurons across REM sleep. At the same time, the techniques to assess activation used in the present paper by Fleury Curado and colleagues (positron emission tomography and magnetic resonance imaging) to measure effect on airway dimensions may be helpful to determine optimal delivery of chemogenetic receptors. At long last, it does seem hopeful that we are palpably closer to an elusive drug therapy for obstructive sleep apnea.