The role of enteroendocrine cells in visceral pain

Abstract

While visceral pain is commonly associated with disorders of the gut-brain axis, underlying mechanisms are not fully understood. Dorsal root ganglion (DRG) neurons innervate visceral structures and undergo hypersensitization in inflammatory models. The characterization of peripheral DRG neuron terminals is an active area of research, but recent work suggests that they communicate with enteroendocrine cells (EECs) in the gut. EECs sense stimuli in the intestinal lumen and communicate information to the brain through hormonal and electrical signaling. In that context, EECs are a target for developing therapeutics to treat visceral pain. Linaclotide is an FDA-approved treatment for chronic constipation that activates the intestinal membrane receptor guanylyl cyclase C (GUCY2C). Clinical trials revealed that linaclotide relieves both constipation and visceral pain. We recently demonstrated that the analgesic effect of linaclotide reflects the overexpression of GUCY2C on neuropod cells, a specialized subtype of EECs. While this brings some clarity to the relationship between linaclotide and visceral analgesia, questions remain about the intracellular signaling mechanisms and neurotransmitters mediating this communication. In this Fundamental Neurochemistry Review, we discuss what is currently known about visceral nociceptors, enteroendocrine cells and the gut-brain axis, and ongoing areas of research regarding that axis and visceral pain.

Graphical Abstract

In this review we discuss the novel interaction between enteroendocrine cells (EECs) and spinal afferent nociceptors and their role in visceral pain. We focus on the role of the intestinal membrane receptor guanylyl cyclase C (GUCY2C) in visceral nociception and its potential as a therapeutic target.

Introduction

Chronic visceral pain, especially in the context of gastrointestinal (GI) disorders, remains an area of immense clinical concern and has substantial implications for global public health. A large proportion of the global population experiences this pain, with diseases of the gut-brain interaction (DGBIs) like irritable bowel syndrome (IBS) affecting an estimated 10–15% of individuals worldwide (Canavan et al., 2014). Beyond its prevalence, the intricacies surrounding its etiology encompass molecular, genetic, psychological, and even socio-environmental dimensions (Drossman & Hasler, 2016). The complexities underlying visceral pain result in challenges to its treatment, often leading to suboptimal patient outcomes. Particularly in the realm of GI-associated visceral pain, therapeutic modalities sometimes provide only transient relief, underscoring an unmet need for innovative therapeutic interventions and a deeper understanding of the root mechanisms (Zhou et al., 2019). In the context of visceral pain, understanding the etiologies and origins of the painful stimuli may ultimately lead to better therapeutics and treatments.

The response to a painful stimulus requires both a localization and appreciation of the intensity of that stimulus. Acute visceral stimulation tends to be poorly localized, generally with a dull or aching intensity. For instance, upper abdominal pain of visceral origin may radiate to the shoulder blade and suggest cholecystitis (Oppenheimer & Rubens, 2019). Visceral pain is often described as “referred” or “diffuse” pain (Gebhart & Bielefeldt, 2016). This type of pain is distinct from somatic pain – which is generally well-localized at the source. The characteristic features of visceral pain emerge from its cellular, physiologic, and anatomic underpinnings.

Visceral organs generally receive innervation from two nerves – either the vagus nerve and a spinal nerve, or two spinal nerves (Figure 1) (Bielefeldt & Gebhart, 2022; Simon & Mertens, 2009). These spinal nerves generally pass through the pre- and para-vertebral ganglia, and have cell bodies in the dorsal root ganglion (DRG) (Godel et al., 2017). In the spinal cord themselves, these nerves synapse in the superficial dorsal horn lamina I and III, interiomedial lateral cell column, and sacral parasympathetic nucleus (Kirazlı et al., 2016). Additionally, visceral structures such as the gastrointestinal (GI) tract have their own intrinsic nervous system which regulates secretion, mobility, and blood flow (Browning & Travagli, 2014). This intrinsic system interacts with extrinsic visceral innervation – yet the form of this interaction remains incompletely understood (Cervi et al., 2014).

Figure 1: Enteroendocrine cells communicate with DRG neurons, vagal afferents, and the enteric nervous system.

Vagal and spinal afferent neurons emerge from the brain and spinal cord respectively, and innervate enteroendocrine cells. The intestinal epithelium is heterogeneous, containing both adsorptive enterocytes as well as secretory cell subtypes. EECs form connections with DRG neurons which project from the spinal cord dorsal horn, relaying nociceptive signals. In addition, EECs form communicate with vagal afferents to relay homeostatic and nutrient signals. Intrinsically, the gut is innervated by the enteric nervous system with the submucosal plexus and the myenteric plexus. EECs communicate with the enteric nervous system, both the enteric neurons and enteric glia, to regulate gut motility.

The majority of fibers involved in the transmission of visceral pain are thinly myelinated a-delta fibers or unmyelinated C fibers (Arcilla & Tadi, 2023; Santiago et al., 2000). A relatively low number of axons innervate visceral structures compared with somatic innervation; approximately 5 to 15% of afferent input to the spinal cord arises from visceral structures (Bielefeldt & Gebhart, 2022). In part, this reflects the large arborization of visceral afferents and the convergence of these axons on common ganglia. This organizational structure permits a small number of axons to innervate visceral organs and also underlies why visceral pain signals are poorly localized and generally referred. While at the organ level the anatomical and physiological bases of visceral pain are generally understood, much less is known about the peripheral cellular and molecular mechanisms transmitting and regulating noxious signaling, particularly those involving the gut-brain axis.

Here, we review recent work that suggests a novel mechanism of visceral nociception involving a synaptic-like interaction between specialized intestinal epithelial cells and spinal visceral afferents. To provide the necessary context, we briefly describe the properties of DRG neurons involved in visceral nociception, enteroendocrine cells (EECs), and neuropod cells. We then focus on the novel role of guanylyl cyclase C (GUCY2C) in visceral nociception and its therapeutic potential, and conclude with a perspective on remaining challenges and future directions.

DRG Nociceptors

A variety of stimuli can excite visceroceptive spinal DRG neurons, including gallbladder, urinary bladder, and colon distention. In the case of IBS, patients report tenderness to abdominal palpation. Spinal DRG neurons generally are innervated from higher order centers and are therefore subject to descending modulation from higher order brain centers. Further, spinal neurons typically receive somatic input as well. At higher order centers in the nervous system, visceroceptive stimuli are generally processed in the thalamus at the ventroposterolateral nucleus. Experiments of noxious visceral stimuli have largely focused on electrophysiologic and behavioral responses to noxious mechanosensitive stimuli (Eijkelkamp et al., 2007; Ness & Gebhart, 1990). Visceral mechanoreceptors are generally differentiated into low and high threshold receptors and it is likely that both classes contribute to visceral pain states (Gebhart & Bielefeldt, 2016). Studies of mechanoreceptors involve balloon distention of hollow organs which allows experimental control of the location, intensity, and quality of the stimulus. Less emphasis has been placed on thermal and chemical visceral nociception, with most focusing on activation of TRPV1 receptors by capsaicin.

Little is known about the peripheral terminals of visceroceptive neurons, and it is assumed that these endings are generally unencapsulated. However, these fibers may form synapses with hybrid visceral structures with neuron-like properties at their peripheral terminals (Bohórquez et al., 2015). For instance, neuropod cells, a specialized type of EEC, exhibit neuron-like properties and synapse with DRG and vagal neurons (Bellono et al., 2017; Kaelberer et al., 2018). Through these interactions, spinal cord afferents may receive initial input to transmit painful signals to the CNS. Yet, the exact role that EECs play in visceral pain, how they communicate with visceral afferents, and what neurotransmitters are involved, among other questions, remain unknown. However, it is increasingly clear that these cells play a critical role in the pathophysiology of visceral pain.

Ion Channels in Visceral Pain

Visceral pain reflects chronic noxious stimulation of visceral nociceptors through inflammatory mediators, ectopic activity, and/or noxious stimuli (Sikandar & Dickenson, 2012). Noxious stimuli of DRG nociceptors can act on several different receptors to induce sensitization, including transient receptor potential vanilloid receptors (TRPV) and ion channels (Farmer & Aziz, 2009). A higher proportion of nociceptive visceral DRG neurons express TRPV1, a nonselective cation channel involved in the somatosensory system (Robinson & Gebhart, 2008). TRPV1 is upregulated in DRG neurons from patients with a variety of visceral hypersensitivity syndromes including inflammatory bowel syndrome and painful bladder syndrome (Mukerji et al., 2006; Schiltenwolf & Akbar, 2008). In models of inflammatory states, DRG neurons are hypersensitive to stimuli, firing a greater number of action potentials (APs) and having a lower threshold for the generation of APs. These DRG neurons express a variety of ion channels, including voltage gated sodium channels (Navs). In particular, Nav1.8 is commonly expressed in nociceptive DRG neurons. Nav1.8 knockout mice exhibit reduced response to painful stimuli and to DRG neuron hypersensitivity, which typically follows GI inflammation (Hillsley et al., 2006; Laird et al., 2002). Additionally, visceroceptive neurons express voltage gated potassium channels (Kvs) which, among diverse functions, are responsible for the repolarization phase of the AP. In models of visceral hypersensitivity, DRG neurons exhibit reduced potassium currents (Beyak, 2010). In addition to Navs and Kvs, visceral afferent neurons express ligand-gated ion channels which open upon agonist binding. Two major classes of ligand-gated channels are expressed by visceral DRG neurons: ionotropic serotonin (5-HT) receptors and purinergic P2X receptors. GI hypersensitivity is associated with increased 5-HT₃ levels, and 5-HT₃ receptor antagonists are used to treat irritable bowel syndrome (Gershon & Tack, 2007). ATP, the physiologic agonist of P2X receptors, is released from cells upon mechanical stimulation and is well-known to modulate visceral nociception (Burnstock, 2009).

Ion channels are integral transmembrane proteins that allow a regulated flow of ions across the plasma membrane, allowing for changes in the polarity of the cell’s membrane potential. Both ligand gated ion channels (LGICs) and transient receptor potential (TRP) ion channels play an important role in the communication between EECs and the nervous system. The 5-hydroxytryptophan₃ receptor (5-HT₃R) is a cation selective LGIC formed by a pentamer of subunits which can be homomeric or heteromeric (Alexander et al., 2011). It is found in both peripheral and central nervous systems, as well as the enteric nervous system; the role of 5-HT₃R in gut motility will be discussed later in the review (Thompson & Lummis, 2006). 5-HT₃R is activated by binding of its ligand, serotonin (5-HT), which leads to an influx of cations and an excitatory response (Thompson & Lummis, 2006). Alosetron and cilansetron are two 5-HT₃R antagonists have been FDA approved for the treatment of irritable bowel syndrome with diarrhea (Chey & Cash, 2005; T. Lembo et al., 2001). Side-effects of these drugs include constipation, and therefore is not prescribed for individuals with IBS with constipation (IBS-C) (Andresen et al., 2008). Clinical trials have shown that 5-HT₃R antagonist treatment leads to a decrease in visceral pain as well as diarrhea relief (Andresen et al., 2008). The mechanism of alosetron’s visceral pain relief is not fully understood, but recent studies suggest that it reduces the activity of EECs (Bayrer et al., 2023).

The TRP family of ion channels are evolutionarily conserved and are responsible for sensing a wide variety of stimuli (Samanta et al., 2018). Vanilloid TRP (TRPV) and ankyrin TRP (TRPA) have been characterized on EECs and visceral nociceptors (Holzer, 2011; Samanta et al., 2018). TRPV1 is present in the digestive tract primarily along the DRG afferent neurons and enteric neurons (Holzer, 2011). It is upregulated in neurons in mouse models of colitis as well as patients with inflammatory bowel disease (Miranda et al., 2007; Yiangou et al., 2001). Because TRPV1 is upregulated in IBS patients and can be sensitized by proalgesic and inflammatory mediators, it has become a target for treating visceral pain (Holzer, 2008). However, most TRPV1 antagonists have the serious side-effect of altering core body temperature, limiting utility to treat visceral pain (Yue et al., 2022). TRPA1 is activated by pungent compounds such as allyl isothiocyanate (AITC; mustard seed oil) and cinnamaldehyde (cinnamon flavonoid) (Samanta et al., 2018). It was previously implicated in visceral hypersensitivity in the GI tract, although this was primarily attributed to its presence on colonic afferents (Brierley et al., 2009). Recent studies suggest that TRPA1 localized to EECs senses luminal irritants which is then communicated to spinal afferents (Bellono et al., 2017). Application of the TRPA1 agonists AITC or cinnamaldehyde on EECs produced secretion of serotonin (Nozawa et al., 2009). Additionally, activation of TRPA1-positive EECs by bacterial metabolites induced serotonin secretion, activating enteric and vagal neurons (Ye et al., 2021). This suggests that TRPA1 plays a role not only in the pain signaling of DRG neurons but also in sensing luminal contents.

Enteroendocrine Cells

EECs communicate signals from the gut to the brain and other organs, and are of particular interest for treating DGBIs. Historically EECs have been classified based on their hormone expression, which can change over the course of their lifetime and varies along the GI tract (Beumer et al., 2018). Gastric inhibitory polypeptide (GIP) is typically found in the duodenum, cholecystokinin (CCK) in the duodenum and jejunum, glucagon-like peptide (GLP-1) in the jejunum, ileum, and colon, and peptide YY (PYY) in the distal ileum and colon (Psichas et al., 2015). EECs throughout the small and large intestine exhibit a narrow surface facing the lumen and a broader surface at the interface with the submucosa from which dense secretory vesicles are released (Ku et al., 2003). These cells detect nutrients in the intestinal lumen which leads to release of hormones, including GLP-1, GIP, CCK, and PYY (Gribble & Reimann, 2016). Once released into the subepithelial space, gut hormones can signal to vagal, enteric, and spinal sensory neurons (Figure 1) (Barton et al., 2023; Kaelberer et al., 2018; Treichel et al., 2022).

One of the best understood gut-brain pathways is between EECs and vagal afferents. Gut-brain signaling to vagal afferents mediates satiety and nutrient sensing. In that context, CCK is released from EECs and can act on vagal CCK₁ receptors (Moran et al., 1987). Application of CCK on vagal CCK₁ receptors induces satiety and reduces meal size, an effect which is enhanced by the presence of leptin (Peters et al., 2006). Additionally, intraduodenal infusion of nutrients suppresses feeding, an effect which is reduced in vagotomized mice (Greenberg et al., 1990; Yox et al., 1991). Vagal afferents are sensitive not only to nutrients but also to stomach and intestine stretch. Isolated cervical vagal afferents are sensitive to circulating CCK, which can activate tension receptors on the afferents. Additionally, detection of gastric load by vagal afferents reduced the concentration of CCK required to reduce feeding in rats (Davison & Clarke, 1988). This indicates that satiety signaling is the result of the integration of gastric distension signaling with direct CCK signaling (Schwartz et al., 1991). Williams et al. identified a set of specialized vagal afferents expressing GLP-1 receptors which detect intestinal stretch. GLP-1 receptor-expressing afferents are distinct from those which sense nutrients, highlighting unique roles of vagal afferents (Williams et al., 2016).

Also, EECs communicate with enteric neurons, primarily by serotonin. The gastrointestinal system is the only internal organ to have evolved an independent nervous system, termed the enteric nervous system (ENS) (Furness et al., 1995; Kunze et al., 1995). EECS and the ENS have extensively been studied in relation to gut motility. Since the identification of serotonin in EECs, it has been implicated in peristalsis and propulsion. However, there has been some debate as to whether serotonin derived from EECs is required for gastric motility. The largest store of serotonin in the body is found in enterochromaffin cells (ECs), a subtype of EECs, making it a primary focus for exploring regulation of peristalsis by enteric neurons. Knocking out tryptophan hydroxylase 1 (TPH1) in the intestine, which is required for serotonin synthesis and is upregulated in ECs, did not interfere with gastric motility (Li et al., 2011). Additionally, serotonin release from the mucosa was not required for cyclical generation of colonic migrating motor complexes (Keating & Spencer, 2010). In contrast, activation of mucosal reflexes led to serotonin release and an increase in the amplitude and frequency of colonic migrating motor complexes (Heredia et al., 2009). These studies suggest that serotonin release from EECs may regulate gut motility through communication with the ENS, but is not be required for motility. This is supported by a more recent study investigating the role of EECs in mechanosensation. A subpopulation of EECs express Piezo2, a mechanosensitive ion channel that are involved in touch and proprioception (Alcaino et al., 2018; Wu et al., 2017). Piezo2-expressing EECs can be activated, resulting in the release of serotonin onto enteric neurons and the regulation of colonic motility (Treichel et al., 2022). The exact signaling cascade that results in serotonin release has not been fully elucidated.

It has been shown that EECs expressing the hormone PYY form a direct connection with enteric glia via a presynaptic pseudopod termed a neuropod (Bohórquez et al., 2014). EECs express receptors for neurotrophic factors produced by enteric glia including NGF-β and artemin, and supplementing organoid media with these factors increased neuropod growth (Bohórquez et al., 2014). With the context that EECs communicate to the ENS through serotonin, this suggests a bidirectional communication between the ENS and EECs. Studies have also suggested that the communication between EECs and the ENS may be heterogeneous. Activation of specific subpopulations of EECs results in varying rates of gut transit, suggesting that each EEC subtype may communicate with different downstream extrinsic or enteric neurons to control gut physiology (Bai et al., 2022). This aligns with single cell RNA sequencing of enteric neurons, which defines distinct subpopulations with heterogeneity of hormone and neurotransmitter receptors (Drokhlyansky et al., 2020). Therefore, signal transmission between enteric neurons and EECs likely changes depending on both the pre- and post-synaptic cell.

Finally, EECs communicate with primary spinal afferents from DRG neurons as well. Recent studies of serotonergic ECs show that they communicate directly with primary afferent fibers that extend into the intestinal villi (Bellono et al., 2017). These ECs sense different stimuli through unique signaling pathways, enabling them to detect irritants, microbial metabolites, and stress response-related catecholamines. In addition, stimulating ECs with epinephrine led to serotonin release that was dependent on voltage-gated calcium channels, although the physiological importance of this was not determined. The relationship between ECs and DRG neurons recently was expanded by Bayrer et al, who showed that activation of ECs induced hypersensitivity to gut distention (Bayrer et al., 2022). Further, prolonged EC activation led to an increase in anxiety-like behaviors which was reduced by serotonin inhibitors (Bayrer et al., 2022). This furthers our understanding of the connection between DGBIs and psychosocial comorbidities.

Enteroendocrine Cells and Visceral Pain

Studies suggest that in IBS, the intestinal epithelial barrier permeability, gut microbiota, food antigens, and stomach bile can result in the dysregulation of the immune system, ENS, and gut-brain axis, ultimately resulting in the chronic noxious stimulation of visceral nociceptors. The immune system plays an important role in visceral hypersensitivity, as the release of inflammatory agents can result in persistent activation and peripheral sensitization of nociceptive afferents (Farmer & Aziz, 2009). As a result, there is often an increase in pain sensitivity at the site of injury known as hyperalgesia. Mast cells are found in connective tissue throughout the body, including the intestinal mucosa, and play an important role in the initiation of inflammatory signals (Fong & Crane, 2023). Rats lacking mast cells failed to develop visceral hypersensitivity in a 2,4,6-trinitrobenzenesulfonic acid (TNBS) model, which induces chronic visceral hypersensitivity (CVH) in WT mice (Hughes et al., 2009; Ohashi et al., 2008). However, rats lacking mast cells do not have significantly different pain thresholds compared to wild type, suggesting that although mast cells are important to the development of CVH, they do not play a role in modulating visceral pain (Ohashi et al., 2008). In addition, the interaction between mast cells and enterochromaffin cells (ECs) has been of interest because both produce serotonin (Farmer & Aziz, 2009). Specifically, interleukin-33 (IL-33) stimulates the release of serotonin from ECs increasing gut motility (Chen et al., 2021). IL-33 is upregulated in mast cells and is released following tissue damage, suggesting an alternative explanation for DGBIs that exhibit dysregulation of serotonergic signaling.

In addition to stimulation via mast cells, ECs are also stimulated by microbial metabolites. The microbial metabolites of spore-forming bacteria have been shown to stimulate 5-HT production in colonic ECs (Yano et al., 2015). Given that ECs communicate with the ENS as well as vagal afferents, this suggests a mechanism by which a perturbed gut microbiome leads to chronic serotonin release and noxious stimulation of nociceptors. This is supported by data which indicates that treatment of enteric neurons with mucosal supernatants from human IBS patients resulted in an increase in neuron spiking (Buhner et al., 2009). Additionally, this spiking was abolished by co-treatment with a 5-HT3 receptor antagonist (Buhner et al., 2009). More recently it has been shown that isovalerate, a short chain fatty acid microbial metabolite, stimulates ECs and activates vagal afferents further supporting this hypothesis (Bellono et al., 2017). However, direct manipulation of the microbiome to modulate nociception has not yet been shown.

As ECs are implicated in visceral nociception, it would follow that a change in the quantity of ECs occurs in individuals with IBS. However, in mouse models and humans there is not a clear consensus on whether different subtypes of EECs are altered in IBS. A study of TNBS-treated rats showed a decrease of colonic EECs expressing chromogranin A, PYY, Neurog-3, and NeuroD1 but an increase in EECs expressing serotonin and somatostatin (El‑Salhy et al., 2016). This is paradoxical given that chromogranin A (ChgA) is often coexpressed with serotonin in EECs, and therefore is used as a EC marker (Bellono et al., 2017). Another study was conducted by the Gilja lab, this time investigating changes in EECs in patient samples from all IBS subtypes. They found a decrease in EECs expressing Neurog-3, ChgA, and serotonin across all IBS subtypes (El‑Salhy & Gilja, 2017). These data indicate that visceral nociception could stem from a loss of EECs and EEC development. However, other studies have also shown no change in EECs between IBS and normal patients. A study of IBS with diarrhea patients showed no change in the number of EECs in the ileum, ascending colon, or rectum compared to healthy samples (Park et al., 2006). However, they did find a negative correlate between the number of EECs present and the maximum rectal pressure that could be tolerated, indicating that EEC signaling increases visceral nociception (Park et al., 2006). A study of the jejunum of patients with IBS also found no change in the number of EECs compared to healthy individuals, further suggesting that EEC signaling, and not EEC development, is perturbed in IBS (Mohammadian et al., 2020).

Neuropod Cells

Recently, a subset of EECs was identified with an axon-like basal process that extends into the basal mucosa, termed a neuropod (Bohórquez et al., 2014). This process contains neurofilaments, a structural component typically observed in axons (Bohórquez et al., 2014). In addition, 73.5% of peptide-secreting vesicles within these cells were contained within the neuropod suggesting a function in hormone secretion and signaling, as well as an anatomical location for hormone secretion (Bohórquez et al., 2014). Neuropod cells were first characterized as contacting enteric glia, the support cells of the enteric nervous system, and have since been shown to communicate with DRG neurons (Figure 2) (Barton et al., 2023; Bellono et al., 2017; Bohórquez et al., 2014). This further emphasizes the connection between EECs and the nervous system (Bohórquez et al., 2014). Bohorquez et al. followed up on this work to determine that the neuropod contacts neurons within the small intestine and colon, the first time such a connection was identified. They identified both pre- and post-synaptic proteins within the neuropod cells and confirmed synaptic connections by Rabies tracing (Bohórquez et al., 2015).

Figure 2: Neuropod cell signaling to DRG neurons transmitting pain signals.

Neuropod cells express a variety of ion channels and signal to DRG neurons, which transmit to spinal cord afferents and ultimately higher order brain centers.

Functionally, neuropod cells synapse with vagal nodose neurons and transmit responses to sugar (Kaelberer et al., 2018). The ability of neuropod cells to activate neurons has been studied using optogenetics, a technique that allows for the activation or inhibition of specific cell types by photostimulation. By expressing the activating channelrhodopsin in CCK-positive intestinal epithelial cells, Kaelberer et al. where able to investigate the effect of EEC activation on vagal nodose neurons. While neuropod cells express CCK, so do other subtypes of EECs and as such the observed effects may not be limited to neuropod cells (Barton et al., 2023; Egerod et al., 2012; Kaelberer et al., 2018). Activation of neuropod cells by optogenetics led to an increase in vagus nerve firing rate in vivo, an effect which was replicated by the application of sucrose to neuropod cells. This was further investigated ex vivo using co-cultures of vagal afferents with CCK-positive neuropod cells expressing channelrhodopsin. Optogenetic activation of neuropod cells excited co-cultured vagal neurons; this excitation was reduced by kynurenic acid, a broad-spectrum ionotropic glutamate receptor blocker suggesting that the interaction between neuropod cells and vagal neurons is mediated by glutamate (Majláth et al., 2016). Buchanan et al. later showed that not only do neuropod cells sense nutrients, but they discriminate between sugars and artificial sweeteners (Buchanan et al., 2022). Neuropod cells express several nutrient receptors, including sodium glucose transporter 1 (SGLT1) and the taste receptor T1R3. When SGLT1 was inhibited, vagal responses to sugars were reduced. However, when T1R3 was inhibited, vagal response to the artificial sweetener sucralose was reduced suggesting different mechanisms for sensing sugar versus sweeteners. In that context, the neurotransmitter for sugar preference is glutamate, and duodenal infusion of glutamate receptor inhibitors decreased the preference for sucrose over sucralose in mice. Thus, neuropod cells are involved in the preference for sugar versus sweetener suggesting that they play a role in calorie sensing.

GUCY2C as a Therapeutic Target for Visceral Pain

We recently showed that neuropod cell GUCY2C controls visceral hypersensitivity, suggesting a new therapeutic target for visceral pain that universally occurs with IBS (Barton et al., 2023). Currently, treatment of IBS often addresses bowel function independently of visceral hypersensitivity. Visceral pain treatment is limited to anti-inflammatory drugs, which often are not effective, and opioids, which have adverse side-effects. Two guanylyl cyclase-C (GUCY2C) agonists have been FDA-approved to treat constipation for IBS with constipation (IBS-C) and chronic idiopathic constipation (CIC) (Miner et al., 2017; Rao et al., 2012). GUCY2C is an intestinal membrane receptor that converts GTP to the second messenger cGMP. GUCY2C has an extracellular ligand-binding domain and an intracellular catalytic domain (Kuhn, 2016). Upon ligand binding, the catalytic domain is activated, converting GTP to cGMP. Cyclic GMP acts as a second messenger to activate cGMP-dependent protein kinase II (PKGII) which phosphorylates the cystic fibrosis conductance regulator (CFTR), an ion channel in the intestinal epithelium. Activation of the CFTR leads to chloride secretion in the intestinal lumen, generating an electrochemical gradient that leads to further ion secretion (Ahsan et al., 2017; Foulke‑Abel et al., 2016; Pattison et al., 2016). Simultaneously, cGMP inhibits the sodium-hydrogen exchanger (NHE3) which prevents dissipation of that gradient. Activation of the CFTR and inhibition of NHE3 lead to an osmotic gradient which generates fluid secretion from the enterocytes, an effect which is leveraged to treat constipation.

Originally, GUCY2C was identified as the receptor for the heat-stable enterotoxin (ST), the causative agent of traveler’s diarrhea (Hasegawa et al., 1999; Schulz et al., 1990). Since then, the endogenous ligands guanylin and uroguanylin were identified, and two agonists have been developed as FDA-approved treatments for constipation (Currie et al., 1992; Hamra et al., 1993). Linaclotide, an ST analog, and plecanatide, a uroguanylin analog, both improved IBS-C and CIC in randomized placebo controlled trials (Shah et al., 2018). Beyond constipation, in clinical trials linaclotide also significantly relieved abdominal pain, compared to placebo (Chey et al., 2012; A. J. Lembo et al., 2011; Rao et al., 2012). Additionally, linaclotide reduces visceral pain in models of bladder pain and endometriosis, suggesting that the mechanism of visceral pain relief is applicable to pathologies beyond constipation (Ge et al., 2019; Grundy et al., 2018). The canonical GUYC2C signaling mechanism described above can be used to explain the constipation relief via fluid and ion secretion. However, the mechanism of visceral pain relief is not fully understood.

Neuropod Cell GUCY2C and Visceral Pain

We recently identified a unique population of intestinal epithelial cells which overexpress functional GUCY2C, termed GUCY2C^High cells (Barton et al., 2023). These cells are rare (<1% of intestinal epithelial cells) and transcriptomic profiling revealed they were enriched in transcripts for CCK, Pyy, and Chga, which are EEC markers (Bellono et al., 2017; Gribble & Reimann, 2016). This was confirmed by Gene Set Enrichment Analysis (GSEA), which revealed that GUCY2C^High cells are similar to fetal intestine chromaffin cells and fetal stomach neuroendocrine cells. Together, these data suggest that GUCY2C^High cells are a subtype of EECs. SynGO analysis revealed that GUCY2C^High cells are enriched in presynaptic transcripts, suggesting that GUCY2C^High cells are neuropod cells. Flow cytometry demonstrated that GUCY2C^High cells selectively express CCK (an EEC marker) and Syn1 (a presynaptic marker), which are established markers of neuropod cells (Gribble & Reimann, 2016; Kaelberer et al., 2018). These data indicate that GUCY2C^High cells are a subset of neuropod cells.

Since neuropod cells form synapses with peripheral and central neurons, we explored this relationship using co-cultures of DRG neurons and intestinal crypts containing GUCY2C^High cells (Bellono et al., 2017; Kaelberer et al., 2018). DRG neurons in co-culture with intestinal crypts behaved in a neuropod cell-dependent manner. DRG neurons which did not synapse with neuropod cells fired a single AP upon stimulation for 1s at 3 times rheobase and had a characteristic rheobase on the order of 100 pA, as expected for healthy DRG neurons in vitro. In contrast, DRG neurons which synapsed with neuropod cells had a lower rheobase (less than ½ that of isolated DRG neurons) and fired multiple APs upon stimulation, indicative of hyperexcitability. Further, when linaclotide, a GUCY2C agonist, was added to the co-culture, DRG neurons synapsing with neuropod cells exhibited an increased rheobase and decreased number of APs, phenocopying the behavior of DRG neurons in the absence of neuropod cells. Linaclotide had no effect on DRGs in the absence of neuropod cells.

To determine if linaclotide was acting through GUCY2C, similar electrophysiology experiments were conducted on co-cultures of DRG neurons and crypts from GUCY2C knockout (GUCY2C^−/−) mice. DRG neurons synapsing with GUCY2C^−/− neuropod cells still had a relatively low rheobase and fired multiple APs upon stimulation as expected. However, when linaclotide was applied to the co-culture, DRG neurons synapsing with GUCY2C^−/− neuropod cells exhibited no change in rheobase or number of APs on prolonged stimulation. Therefore, the effects of linaclotide on reducing hyperexcitability are mediated by GUCY2C. Previous studies suggested that the analgesic effect of linaclotide was mediated by extracellular cGMP (Castro et al., 2013). To investigate this further, co-cultures of DRG neurons and intestinal crypts containing GUCY2C^High cells were treated with extracellular cGMP. Extracellular cGMP had no effect on either the rheobase or number of APs on produced by stimulation. Yet, linaclotide, applied to the same co-cultures following cGMP application, decreased DRG neuron hyperexcitability. These observations indicate that extracellular cGMP does not mediate the analgesic effect of linaclotide, and alternative signaling molecules will be investigated going forward.

Since linaclotide relieves visceral pain as well as constipation in patients with IBS-C, we explored the roll of GUCY2C in visceral pain. We used three measurements of visceral pain produced by colorectal distension: abdominal withdrawal reflex (AWR; behavioral); visceromotor reflex (VMR; physiological); and spinal cord phospho-ERK (biochemical) (Castro et al., 2013; Najjar et al., 2021; Xu et al., 2018). With each of these techniques, there was a significant increase in sensitivity to visceral pain produced by colorectal distension in GUCY2C^−/− (global knock out), compared to wild type, mice. In addition, linaclotide reduced visceral pain in WT, but not GUCY2C^−/−, mice. These observations suggest that silencing GUCY2C signaling in the intestine produces spontaneous visceral hypersensitivity and increased responses to colorectal distension that is refractory to linaclotide. Importantly, silencing GUCY2C signaling specifically in neuropod cells also produced visceral hypersensitivity that was refractory to linaclotide. We also wanted to test the ability of linaclotide to relieve visceral pain in an IBS model. To do this we utilized a TNBS model, which induces chronic visceral hypersensitivity (CVH) in mice (Hughes et al., 2009). Activation of GUCY2C signaling has been shown to relieve visceral pain in TNBS models of CVH (Eutamene et al., 2010; Shailubhai et al., 2015). CVH mice treated with linaclotide had lower AWR scores and reduced pERK. However, silencing GUCY2C signaling specifically in neuropod cells in CVH mice produced a response that was refractory to linaclotide, quantified by AWR and pERK. Together, these observations demonstrate that neuropod cells specifically mediate the effects of GUCY2C agonists modulating nociceptive signaling between the gut and brain.

Physiological and Cellular Models of Visceral Pain Regulation by GUCY2C

The working hypothesis suggests that silencing GUCY2C leads to DRG neuron hyperexcitability (lower rheobase, repetitive APs) and visceral pain, which is ameliorated by activation of GUCY2C signaling by linaclotide (Barton et al., 2023). Linaclotide functions primarily in the small intestine, with the majority of cGMP production occurring in the duodenum and jejunum (Busby et al., 2010). Additionally, neuropod cells are clustered in the duodenum and jejunum compared to other areas of the gastrointestinal tract (Barton et al., 2023). Of clinical relevance, patients with IBS-C and CIC exhibit a loss of uroguanylin, the endogenous small intestine GUCY2C ligand, silencing the GUCY2C-cGMP signaling axis (Waldman & Camilleri, 2018; Waldman et al., 2019). These observations suggest the hypothesis that neuropod cells in the small intestine control visceral nociceptive signals ascending from the colorectum.

Current studies raise two major questions regarding cellular mechanisms of GUCY2C-mediated regulation of visceral pain that inform future directions. First, the specific neurotransmitter responsible for mediating the interaction between DRGs and GUCY2C^High neuropod cells remains undefined. In the absence of GUCY2C activity (i.e., uroguanylin insufficiency), we hypothesize that neuropod cells are activated and constantly release an excitatory messenger that acts on DRG neurons to induce hyperexcitability. Regulation of DRG hyperexcitability by neuropod cell GUCY2C signaling could reflect inhibition of synaptic excitatory neurotransmitter vesicle release by intracellular cGMP. The pseudopod structure for which neuropod cells are named are densely packed with vesicles for neurotransmission (Bohórquez et al., 2014). Further, there is precedent for soluble guanylyl cyclase production of cGMP regulating neuron firing (Jouvert et al., 2004; Kaun et al., 2007; Yang & Hatton, 1999). Additionally, guanylyl cyclase D (GUCY2D) silences photoreceptor rod cells through cGMP, which acts on cyclic nucleotide gated channels (CNGCs) (Kuhn, 2016). Furthermore, GUCY2C^High neuropod cells preferentially express several excitatory neurotransmitters. Thus, serotonin is expressed in GUCY2C^High neuropod cells and released from ECs mediating visceral pain, as discussed above (Barton et al., 2023; Bayrer et al., 2023). Serotonin enhances DRG neuron excitability through the 5-HT2C receptor (Salzer et al., 2016). In addition, serotonin sensitized isolated DRG neurons to activation by TRPV1 agonist, an effect that was blocked by a 5-HT3 inhibitor (Cenac et al., 2010). Together, these data suggest a mechanism where GUCY2C^High neuropod cells release excitatory serotonin at baseline leading to DRG hyperexcitability and visceral hypersensitivity. Activation of GUCY2C signaling by linaclotide could reduce serotonin release, decreasing DRG neuron hyperexcitability which relieves visceral pain.

Recently, Bayrer et al. specifically implicated EC serotonin in sensing visceral pain (Bayrer et al., 2023). Intra-colonic application of isovalerate (ISV), a short-chain fatty acid microbial product elevated in individuals with visceral hypersensitivity, activated ECs and increased the sensitivity of DRG neurons (Bayrer et al., 2023). The selective 5-HT₃ receptor antagonist alosetron blocked ISV hypersensitivity, suggesting the interaction is mediated by serotonin. However, the direct effect of ECs on nociceptive neuron firing was not explored. If serotonin is mediating visceral hypersensitivity produced by chronic ECC activation, then hyperexcitability of DRG neurons by neuropod cells in co-cultures should be blocked by alosetron. This experiment would better address the direct impact of serotonin on visceral nociception and the specific receptor involved. In addition to mediating visceral pain, Bayrer et al. suggested that chronic activation of ECs leads to increased anxiety behavior, a phenotype that was abolished by alosetron. This provides a possible mechanism for the high rate of comorbid anxiety in IBS. However, alosetron was administered subcutaneously where it can act globally on peripheral neurons, not exclusively on the DRGs and mucosal afferents associated with the GI tract (Bayrer et al., 2023). Given this, the direct role of serotonin in visceral hypersensitivity and associated anxiety behaviors remains unclear.

Beyond serotonin, glutamate mediates synaptic communication between neuropod cells and the vagus nerve in nutrient sensing (Buchanan et al., 2022). Inhibition of metabotropic and ionotropic glutamate receptors decreased the vagal response to sugar, but not artificial sweetener, suggesting that stimulation of neuropod cells with sugar leads to glutamate release (Buchanan et al., 2022). Further, the preference for sugar compared to sweetener is mediated by glutamate. Glutamate receptor blockers delivered locally into the duodenum reduced preference for sucrose compared to sucralose (Buchanan et al., 2022). Thus, neuropod cells release glutamate following sugar stimulation, which mediates changes in behavior. Interestingly, while neuropod cells use glutamate as a neurotransmitter, transcriptomic profiling of GUCY2C^High neuropod cells, compared to GUCY2C^Med enterocytes, do not reveal enrichment of the glutamate synthesis machinery (Barton et al., 2023). This suggests that an alternative neurotransmitter may be involved in communicating visceral pain.

Alternatively, GUCY2C signaling could release a second inhibitory factor which decreases DRG excitability. Single cell transcriptomics have shown that both excitatory and inhibitory neurotransmitter transcripts can be contained within one cell.(Brunet Avalos & Sprecher, 2021) Therefore, neuropod cells could produce glutamate and serotonin as well as inhibitory factors. Notably, application of GABA receptor agonists attenuates the visceromotor reflex to colorectal distension (Loeza‑Alcocer & Gold, 2021; Loeza-Alcocer et al., 2019). However, transcriptomic analysis reveals that the vesicular inhibitory amino acid transporter (VIAAT, Slc32a1) is not produced by neuropod cells (Barton et al., 2023). VIAAT is a synaptic vesicle protein required for the storage of inhibitory neurotransmitters like GABA and glutamine, and its absence makes it unlikely that they are secreted from neuropod cells (Juge et al., 2009). Interestingly, components of the glutamine and glycine degradation apparatus are enriched in GUCY2C^High neuropod cells supporting the suggestion that an inhibitory neurotransmitter is not mediating GUCY2C signaling effects by neuropod cells (Barton et al., 2023). However, because neuropod cells are not strictly neurons and most neurotransmitter studies have been conducted in neurons, a previously unanticipated mechanism involving the release of a novel inhibitory neurotransmitter from neuropod cells may mediate the effects of GUCY2C signaling.

Beyond defining the neurotransmitters mediating the effects of neuropod cells on neuron excitability, it is important to explore cellular mechanisms of GUCY2C-cGMP signaling regulating neurotransmitter release. Overexpression of GUCY2C in neuropod cells is associated with a proportional increase in cGMP in response to linaclotide stimulation (Barton et al., 2023). In turn, cGMP can act as a second messenger to activate phosphodiesterases (PDEs), protein kinase A (PKA), protein kinase G (PKG), or cyclic nucleotide gated channels (CNGCs). Canonically, GUCY2C signals through PKGII and PDE5 (Ahsan et al., 2017; Pattison et al., 2016). However both of these proteins, along with several other signaling components of GUCY2C, are significantly downregulated in GUCY2C^High neuropod cells suggesting that these cells signal through an alternate mechanism (Barton et al., 2023). Nitric oxide-mediated cGMP signaling regulates synaptic transmission and neuron excitability (Ahern et al., 2002). In DRG neurons, activation of N-type calcium channels is inhibited by cGMP, which can lead to a decrease in synaptic vesicle release (Klyachko et al., 2001). Voltage-gated calcium channel transcripts were identified in bulk RNA sequencing of GUCY2C^High neuropod cells, although transcripts for N-type channels were not significantly different compared to GUCY2C^Med cells (Barton et al., 2023). In addition, studies of ECs revealed an increase in intracellular Ca²⁺ following depolarization (Bellono et al., 2017). This response was attenuated by inhibiting P/Q type Ca_v channels, which are downregulated in GUCY2C^High neuropod cells compared to GUCY2C^Med cells (Bellono et al., 2017). In photoreceptors, nitric oxide leads to an increase in cGMP which acts on CNGCs to enhance Ca²⁺ efflux, leading to increased synaptic transmission (Ahern et al., 2002). Importantly, GUCY2C^High neuropod cells do not express CNGCs, suggesting that cGMP acts by an alternate signaling pathway (Barton et al., 2023).

Perspective and Future Directions

GUCY2C^High cells represent a small but unique proportion of EECs implicated in visceral pain. These cells are the target of visceral pain analgesics and play a key role in the transmission, modulation, and regulation of pain throughout the GI system (Barton et al., 2023). Electrophysiological experiments of DRG neurons and GUCY2C^High cells in co-culture suggests a mechanism by which this specialized cell type may interact with nociceptive DRG neurons and respond to GUCY2C agonists. Yet, a number of areas require further emphasis and research. Firstly, future directions may focus on the neurotransmitter involved in the communication between the DRG neuron and the GUCY2C^High neuropod cell. While there are a variety of possible neurotransmitters, serotonin is the top candidate. Blocking serotonin receptors during electrophysiology experiments of DRG neurons in co-culture with intestinal crypts can begin to determine if serotonin is mediating their interaction. Secondly, the signaling mechanism within the GUCY2C^High cell following GUCY2C activation remains to be identified. Extracellular cGMP is not the terminal messenger, as previous experiments revealed no effect on DRG neurons (Barton et al., 2023). Therefore, additional work on the downstream effects of GUCY2C activation in neuropod cells is warranted. Lastly, a third area of additional research focuses on the anatomical/physiological communication between the colon and the small intestine. In particular, the precise anatomic relationship between cells involved in sensing painful stimuli (such as balloon distention), the GUCY2C^High cells which respond to analgesics, and visceroceptive neurons remains to be defined. In hollow organ balloon models of pain, these cells all occupy different anatomic locations and therefore are part of an intricate network of communication. A similar model could be involved in visceral pain, where the cells sensing distension are located in the colorectum but the cells responsible for relieving visceral pain are in the small intestine.

Whereas in other pathophysiologic contexts, this vast network of cells may appear highly intricate and complicated, within the realm of visceral pain these anatomic relationships fit in quite well. Visceral pain is, by its nature, diffuse and poorly-localized due to the relatively low number of nociceptive afferents and the integration of signals at different points in the pathway. Of course, much still remains to be discovered about visceral pain and the role that EECs play. What is clear is that there exists an emerging role of EECs is in the modulation and regulation of visceral pain. Once considered to be primary sensing afferent neurons with free-endings, visceral afferents appear to now have an intimate relationship with EECs in the context of nociception.