Into the belly of the brainstem: The neurons of the “extended” enteric nervous system

Animals must solve the complex problem of how to produce coordinated patterns of muscular activity to generate precise, purposeful movements. To simplify and constrain the range of possible movements, the nervous system is thought to contain low-level microcircuits that specify basic templates or “patterns” of motor neuron activity to synergistically control groups of muscles to move the body. This level of control is thought to be organized by “premotor” neurons, which integrate and route central nervous system control signals directly onto motor neurons. However, surprisingly little is known about the genetic cell types of premotor neurons and how they are organized to produce the coordinated activation of muscles. Mammalian ingestive behaviors, including chewing, swallowing, licking, and lapping, represent a class of movements that exemplify the challenge of coordinated motor control, as they require synchronization of dozens of muscles to produce patterned orofacial movements (1). In newly published data from the Brunet Research Team, Sungeelee et al. (2) define a class of jaw premotor neurons that integrate input from sensory neurons and forebrain regions and that innervate motor neurons for multiple orofacial muscles. Intriguingly, these neurons express some of the same developmental patterning genes as neurons of the autonomic nervous system, as well as the peripheral neurons of the enteric nervous system that control the gut. This important finding suggests that there is a common developmental origin for neuronal circuits that control ingestion and digestion that unexpectedly transcends the classic division between the control of smooth versus skeletal muscle and voluntary versus involuntary movements.

This common developmental origin likely reflects the possibility that evolution chose to recruit the same developmental program multiple times to construct neural circuits for extracting nutrients from the environment as organisms became increasingly sophisticated (2). Like the evolutionary ancient digestive movements of the gut, ingestive movements to bring food into the body are also often highly rhythmic. In mammals, many species exhibit rhythmic lapping to ingest liquids and rhythmic chewing to mechanically mash solid food (3). The anatomical locus for the central nervous system’s generation of rhythmic licking/lapping in rodents has long been thought to be largely contained within the intermediate reticular formation (IRt) in the brainstem, due to the observation that this region contains many tongue premotor neurons and neurons that are rhythmically active during licking (4, 5). Recent work from the Brunet Research Team elegantly demonstrated that tonic optogenetic stimulation of a genetically-defined subset of these neurons produces rhythmic lapping in mice (6), suggesting that the IRt may participate in both neural rhythm generation and motor pattern generation. Motor pattern generation for chewing is thought to involve the principal trigeminal sensory nucleus (NVsnpr) (7); however, the nature of interactions among multiple neural networks involved in this behavior remain unclear (1, 7), likely because chewing is substantially more complex than licking. Specifically, chewing involves a sequence with multiple rhythmic phases: preparation, reduction, and preswallowing, and different muscle activity patterns are involved in these distinct activities (8). Additionally, while lapping involves mostly positioning the tongue and does not substantially load the muscles, chewing requires generating substantial force and modulating the force based on tactile and proprioceptive feedback, which may require a more flexible central pattern-generating network architecture (9). Correspondingly, the premotor neurons identified in the new work from Sungeelee et al. (2), located in the supratrigeminal region of the pons (Sup5), comprise an essential node in the control of ingestive behavior poised to provide this essential flexibility.

In newly published data from the Brunet Research Team, Sungeelee et al. (2) define a class of jaw premotor neurons that integrate input from sensory neurons and forebrain regions, and that innervate motor neurons for multiple orofacial muscles.

Intriguingly, these premotor neurons, as well as the IRt neurons that generate rhythmic lapping, and a third group of neurons in the peritrigeminal (p5) area, were all discovered recently by the Brunet Research Team (2, 6) to express the homeodomain transcription factor Phox2b. Sungeelee et al. (2) demonstrate that the Sup5^Phox2b neurons project to trigeminal, facial, and hypoglossal motor pools including motor neurons for antagonistic muscles, as well as preganglionic neurons that control salivation. As a population, the Sup5^Phox2b neurons are also bidirectionally connected with forebrain areas implicated in the regulation of feeding behaviors, including the central amygdala and bed nucleus of the stria terminalis. These neurons also receive inputs from other feeding-associated forebrain areas, areas of the brainstem implicated in orofacial motor control, and jaw proprioceptors that sense muscle stretch in the trigeminal mesencephalic nucleus (mes5) (Fig. 1). Both optogenetic stimulation and inhibition of Sup5^Phox2b neurons interrupt volitional lapping and chewing of solid food, and bulk optical imaging of calcium dynamics via fiber photometry shows increases in activity during bouts of chewing that appear to scale with jaw force. Notably, optogenetic stimulation does not produce rhythmic movements, consistent with a previous study investigating bilateral-projecting sup5 neurons comprising multiple genetic cell types (10). These results suggest that the Sup5^Phox2b neurons may be downstream of putative neural rhythm generators for chewing and lapping. Specifically, the broad input/output connectivity of the Sup5^Phox2b neurons and their direct communication with sensory proprioceptive signals suggest a possible preparatory role in ingestion by setting muscular tone, saliva secretion, and/or regulating bite force necessary for ingestion (2). The observation that multiple ingestive behaviors are interrupted by optogenetic manipulations of these neurons suggests that their activity is necessary for these behaviors.

Fig. 1.

The brainstem “extension” of the enteric nervous system. The major brainstem motor (gold) and sensory (purple) nuclei for the control of ingestion, brainstem regions that contain Phox2b+ premotor neurons (green), and brain regions that provide long-range inputs to these neurons (blue), are shown in a sagittal view of the mouse brain (gray). Collectively, brainstem premotor neurons, motor neurons, and sensory neurons that express Phox2b in development have been suggested to comprise the “extended enteric nervous system” (2). These regions, their interconnections, and their output projections are outlined (red). Anatomical abbreviations: area postrema (AP); nucleus of the solitary tract (NST); dorsal motor nucleus of the vagus (Mo10); hypoglossal (Mo12), ambiguus (Amb), facial (Mo7), trigeminal (Mo5) motor nuclei; intermediate reticular formation (IRt); peritrigeminal area (p5); supratrigeminal region (Sup5); deep cerebellar nuclei (DCN); trigeminal mesencephalic nucleus (mes5); superior colliculus (SC); central amygdala (CeA); bed nucleus of the stria terminalis (BNST); parasubthalamic nucleus (PSTh); cortex (Cx). This connectivity diagram is based on previous studies (2, 6, 11, 12).

The fact that IRt premotor neurons, p5 premotor neurons, and Sup5 premotor neurons all express Phox2b suggests that development of these neurons may be organized according to their physiological function, rather than their anatomical location. Importantly, Phox2b is also expressed in the cells that form the peripheral autonomic nervous system and is essential for the development of the sympathetic, parasympathetic, and enteric ganglia (13, 14) that innervate smooth muscle. Additionally, Phox2b is necessary for the development of brainstem visceromotor neurons that innervate these autonomic ganglia (11), brainstem primary sensory areas that receive visceral signals (12, 14), and several pools of orofacial skeletal motor neurons, termed “branchial” motor neurons (11), that are derived from the branchial arches in ancestral fishes and innervate the jaw, face, and upper airway musculature (Fig. 1). Collectively, the branchial motor neurons are referred to as the “special visceral” efferent pathway due to their unique developmental origin among skeletal motor neurons (15). The observations in the two recent studies from the Brunet Research Team (2, 6) that multiple groups of hindbrain premotor neurons follow similar developmental trajectories and also express Phox2b are remarkable because they point to a shared developmental origin for ingestive and digestive movements, likely reflecting the evolutionary history of these functions. One apparent exception appears to be the hypoglossal motor neurons that control the tongue, whose development is not dependent on Phox2b expression (11). This observation is consistent with evidence that the tongue evolved more recently in terrestrial vertebrates (16). Nonetheless, groups of Phox2b premotor neurons seamlessly interface with these motor neurons for ingestive behaviors that require tongue/jaw coordination. As the authors point out, it appears to have been evolutionarily favorable to recruit the same developmental programs as organisms gained more complex nutrient acquisition capabilities, and to adapt the corresponding neural circuits accordingly to utilize more sophisticated motor plants like the tongue, rather than to invent completely new control systems.

Based on the observations by Sungeelee et al. (2) and the many years of discoveries on the role of Phox2b in embryonic development by the Brunet Research Team (12, 13), the authors posit a new definition of the “extended” enteric nervous system consisting of neurons that express Phox2b in development and that regulate nutrient acquisition (Fig. 1). This system includes the neurons of peripheral enteric nervous system, the sympathetic ganglia that innervate these neurons, the branchial motor neuron pools, the visceral sensory neurons of the brainstem, and the recently identified groups of Phox2b+ premotor neurons for ingestion. The study by Sungeelee et al. (2) suggests that neural circuits that implement motor control of ingestion are highly evolved for this specific function and raises the question of whether and which neural strategies are shared with other central pattern generating circuits and motor control circuits in general. Beyond the immediate findings, Sungeelee et al. (2) demonstrate the scientific utility of using developmental gene markers to gain access to specific populations of neuronal cell types to interrogate their function. This strategy is likely to become increasingly tractable and powerful in the modern age of single-cell transcriptomics, as the identification of genetic entry points to target specific populations of neurons in vertebrates becomes increasingly specific and high-throughput.