The enteric nervous system in gastrointestinal disease etiology

Abstract

A highly conserved but convoluted network of neurons and glial cells, the enteric nervous system (ENS), is positioned along the wall of the gut to coordinate digestive processes and gastrointestinal homeostasis. Because ENS components are in charge of the autonomous regulation of gut function, it is inevitable that their dysfunction is central to the pathophysiology and symptom generation of gastrointestinal disease. While for neurodevelopmental disorders such as Hirschsprung, ENS pathogenesis appears to be clear-cut, the role for impaired ENS activity in the etiology of other gastrointestinal disorders is less established and is often deemed secondary to other insults like intestinal inflammation. However, mounting experimental evidence in recent years indicates that gastrointestinal homeostasis hinges on multifaceted connections between the ENS, and other cellular networks such as the intestinal epithelium, the immune system, and the intestinal microbiome. Derangement of these interactions could underlie gastrointestinal disease onset and elicit variable degrees of abnormal gut function, pinpointing, perhaps unexpectedly, the ENS as a diligent participant in idiopathic but also in inflammatory and cancerous diseases of the gut. In this review, we discuss the latest evidence on the role of the ENS in the pathogenesis of enteric neuropathies, disorders of gut–brain interaction, inflammatory bowel diseases, and colorectal cancer.

Introduction

Evolution has endowed the gastrointestinal tract with its own dedicated nervous system. In mammalians, the enteric nervous system (ENS) consists of millions of neurons and glial cells that are organized into interconnected ganglia embedded within the gut wall. The ENS has the ability to autonomously command gastrointestinal tissue dynamics and gut homeostasis, devoid of input from the brain or spinal cord, earning it the sobriquet ‘second brain’ [1]. As the greatest division of the autonomic nervous system and rivalling the spinal cord in terms of complexity, ENS components form integrated circuits which independently or in concert with extrinsic parasympathetic and sympathetic innervation regulate a myriad of gut processes including bowel motility, transmucosal movement of fluids, immune responses, and local blood flow [2]. Positioned in the largest sensory organ of the body, the ENS also works alongside the intestinal epithelia, immune system, enteroendocrine system and intestinal microbiome to allow the absorption of nutrients, water, and electrolytes while at the same time to prevent  access to harmful substances present in the lumen [3, 4].

In vertebrates, the majority of enteric neurons and glia develop from neural crest cells originating from the vagal level of the neural tube which invade the foregut and rostro-caudally migrate culminating in the quasi-uniform colonization of the gastrointestinal tract [5, 6]. The constellation of intrinsic primary afferent neurons, interneurons, and motor neurons, with distinct neurochemical coding and functional roles are surrounded by 1–7 times as many enteric glial cells (EGCs) [7] and generally configured within two major ganglionated and interconnected plexus layers: the myenteric and submucosal plexus. Enteric neuron cell bodies are located in the myenteric and submucosal ganglia while their processes extend throughout the external muscle layers, submucosa and mucosa [2]. In addition, EGCs also reside in extraganglionic spaces and are embedded within the intestinal mucosa and muscle layers [7]. Although evidence for functional specialization is rather limited, much like enteric neurons, EGCs can be classified into distinct subtypes based on their morphology and location within the gut wall and along the gastrointestinal tract [8]. Of late, the established classification schemes for ENS cells have been complemented with data from single-cell sequencing analyses that further unravel the identity of enteric neuron and EGC subpopulations [9,10,11,12,13]. The heterogeneous populations of enteric neurons and glia closely interact to coordinate the stereotypic patterns of gut motility and secretion which are key to gastrointestinal homeostasis. To guide some of these processes, the ENS is assisted by different types of enteroendocrine cells, which are scattered throughout the epithelium and monitor the gut lumen [14, 15]. Among these are enterochromaffin cells, a particular serotonin-producing enteroendocrine cell type thought to be crucial for conveying luminal information to the ENS [16,17,18]. For a detailed presentation on functional circuits and ENS signaling the reader is directed to recent literature [19, 20]. Given its central role as an integrating hub for controlling gastrointestinal physiology [21], it is unsurprising that alterations in ENS function are concomitant with disruptions of gut homeostasis, resulting in both gastrointestinal and extra-gastrointestinal diseases [22, 23]. The complexity of ENS architecture is paralleled by the overarching adverse consequences that can result from perturbations of genes critical for ENS development to subtler alterations in its connectivity and its cross-talk with neighboring cellular systems involved in gut function [24, 25]. In this review, we present the latest evidence for ENS involvement in disorders that strike the gastrointestinal tract. We discuss both experimentally established and hypothesized roles for enteric neurons and EGCs in the pathogenesis of enteric neuropathies, disorders of gut–brain interaction (DGBIs), inflammatory bowel disease (IBD), and colorectal cancer (CRC).

Enteric neuropathies

Enteric neuropathies emanate from loss, degeneration, and/or functional impairment of enteric neurons [26,27,28]. They are caused by congenital defects in ENS development, acquired through the effect of infectious agents or toxins, or secondary to other pathological conditions such as diabetes and neurodegenerative disorders [22, 29]. ENS defects vary from subtle alterations in the biochemistry or connectivity within the neural network to a complete loss of ganglia from entire segments of the bowel. The latter of which can result in life-threatening conditions such as Hirschsprung disease (HSCR), where survival of patients depends on early diagnosis and surgical intervention [30, 31].

HSCR is the best known congenital enteric neuropathy affecting 1 in 5000 individuals, in which the absence of ganglia in distal portions of the gut results in difficult expulsion of meconium, causing life-threatening intestinal occlusion, and over time intractable constipation due to the lack of peristalsis in this intestinal region. Collectively, genetic studies of patients with HSCR [32,33,34] and in vivo transgenic animal models [35,36,37] have identified multiple genes involved in the development of the ENS, including the receptor tyrosine kinase (RET) [38, 39] and endothelin receptor type B (EDNRB) [40] and their family members as major players for the HSCR phenotype, together with mutations in SOX10 [41, 42], PHOX2B [43], semaphorins [44, 45], among other genes [46]. Interestingly, recent studies revealed other molecular HSCR candidates [47] and genetic variants, including pathogenic genes, alleles and loci that can exacerbate the susceptibility of HSCR patients in manifesting the disease phenotype [48, 49]. Highly conserved among species, deficiencies along the signaling pathways of these genes may result in failure of ENS progenitors to migrate, proliferate, differentiate or survive within the distal intestine and cause congenital bowel obstruction [5, 50]. Of note, novel work by Chatterjee et al. [51], identified specialized genetic programs active in ENS cells during critical stages of gastrointestinal organogenesis that also control epithelial, endothelial and muscle cell specification. Therefore, such impaired gene networks not only impel ENS precursor cells to colonize the distal gut, causing aganglionosis, but also influence the development of non-neural intestinal tissues [52]. Remarkably, even after the aganglionic intestine has been surgically removed, patients often experience functional abnormalities which fluctuate from severe constipation to fecal incontinence [53]. Although the pathophysiological mechanisms behind these symptoms are not clear, recent work reported histopathological disturbances in the ganglionic bowel of patients who had previously undergone optimal pull-through surgery [54]. Moreover, fatty acid binding protein 7 (FABP7), a marker of immature enteric glia, was significantly upregulated in the myenteric plexus and resulted in a higher ratio of FABP7 to S100 calcium-binding protein B (S100B) expression as compared to controls, signifying a higher proportion of immature EGCs in the ganglionic bowel of HSCR patients [55]. These findings support the notion that megacolon can be associated with impaired differentiation of ENS precursors, or perhaps, with defective enteric gliogenesis. Cumulatively, HSCR substantiates the role of the ENS in survival, as the formation of neural circuits adroit at executing motility need to be prenatally assembled. In the absence of effective interventions, elucidating the subtype composition of enteric ganglia in the ganglionated intestinal segments could help to gain further insight on how to recover aganglionic functionality in HSCR. Additionally, and as the cause of the variable HSCR phenotypes remains elusive, elucidating the contribution of the intestinal milieu for ENS development might indicate that next to genetic and epigenetic factors, environmental cues are also important to determine or attenuate HSCR pathology [56,57,58]. Indeed, vitamin A deficiency, exposure to high concentrations of drugs such as ibuprofen, mycophenolate, statins, and artemisinin during the critical window of ENS development have been reported to induce a HSCR phenotype in animal models [50]. Also, alterations in epithelial integrity, in the mucosal immune system and in gut microbiota have been put forward as plausible candidates contributing to HSCR [59]. Although direct evidence for a role of the immune system in the onset of HSCR is absent, RET and its pathway members are expressed by type 3 innate lymphoid cells (ILC3) and are involved in the organogenesis of intestinal Payer’s patches [60, 61]. Additionally, while it is clear that HSCR patients present with intestinal dysbiosis [62], further investigation would enlighten the paradoxical proposition of whether dysbiosis can contribute to HSCR, or if it is a consequence of impaired elimination mechanisms due to defective intestinal motility [63, 64]. The role of microbes in HSCR pathogenesis is intriguing, yet not entirely surprising, since collective efforts have explored and divulged the important contribution of the gut microbiome in fine-tuning the ENS development and homeostasis [65,66,67,68,69,70,71,72], and vice versa [73].

While significant progress has been made in the genetics and pathophysiology of ENS constituents in HSCR, a detailed understanding of the role of ENS defects in other gastrointestinal disorders sometimes classified as enteric neuropathies remains limited, and their etiology in many cases is defined as idiopathic. For example, a prominent reduction in enteric neuronal cells in the lower esophagus, especially of inhibitory neurons can be a primary cause of achalasia, which is characterized by an absence of esophageal peristalsis and failure of the lower esophageal sphincter to relax upon swallowing [74, 75]. The neuronal loss is believed to be caused by an autoimmune reaction, possibly to a viral infection, in patients with a particular immunogenetic background, and biopsies from patients with achalasia displayed marked immune cell infiltration within the myenteric plexus and increased production of autoantibodies [76,77,78]. Moreover, genetic factors such as polymorphisms of the protein tyrosine phosphatase N22 (PTPN22), interleukin-23 receptor (IL-23R), and interleukin-10 (IL-10) promoters related to immune cell regulation were reported to contribute to the heterogeneity of disease pathogenesis [79]. Another enteric neuropathy, gastroparesis, is characterized by delayed gastric emptying in the absence of mechanical gastric outlet obstruction. Gastroparesis can be caused by complications of diabetes, and less commonly by medications and surgical interventions but the overwhelming majority of cases remains idiopathic [80, 81]. Studies using human gastric biopsies found a pronounced reduction in interstitial cells of Cajal (ICCs) and neuronal fibers in the circular muscle layer associated with increased immune cell infiltration in the myenteric plexus [82]. Whether gastroparesis can, therefore, be considered as a macrophage-driven ‘cajalopathy’, needs further experimental confirmation [83]. The initiation and progression of chronic intestinal pseudo-obstruction (CIPO), which is characterized by inefficient intestinal transit without any physical obstruction, may involve multiple congenital, acquired or idiopathic causes and distinct cell types including smooth muscle and nerve cells [84]. CIPO is associated with the absence of normal migrating motor complexes that, depending on the underlying pathomechanism, follow a particular neuropathic or myopathic pattern. It can result from an underlying non-gastrointestinal disorder or condition, including a wide variety of systemic, metabolic and organic diseases, such collagen vascular diseases, neurological disorders, or as paraneoplastic phenomenon caused by tumor-derived auto-antibodies [85]. Although CIPO mostly occurs in its sporadic form, it may be also associated with familial inheritance in other patients. For instance, CIPO symptoms can manifest as a result of recessive mutations in important mitochondrial genes (TYMP, POLG) [86, 87]; can be caused by mutations in SGOL1 in chronic atrial and intestinal dysrhythmia (CAID) [88]; and mutations in the ACTG2 gene in megacystis-microcolon-intestinal hypoperistalsis syndrome (MMIHS) [89]. Additionally, CIPO has been reported in patients with autosomal dominant mutations in the SOX10 gene in Waardenburg-Shah syndrome [90], and is also related to sex chromosome inheritance in Xq28 with mutations in Filamin A and L1CAM genes [91, 92]. Although abnormalities in gastrointestinal motility in patients with CIPO manifest in the presence of enteric ganglia [85], a neuronal deficit of approximately 50% associated with increased distance between ganglia, neuronal swelling and axonal degeneration and which correlates with the degree of symptom severity has been reported [93]. This is also the case for slow transit constipation, defined by reduced and infrequent intestinal transit, sensation of anorectal obstruction and solid fecal content [94]. Until now, defining the pathophysiology of slow transit constipation has been challenging as subtle alterations in the ENS are not necessarily detected. Nevertheless, a significant reduction in the size and number of enteric ganglia, and in the number of EGCs and neurons, atypical influx of lymphocytes to the ganglia, decreased numbers of ICCs, presence of intestinal neuronal dysplasia type B, and reduced expression of neurochemical markers [mainly vasoactive intestinal polypeptide (VIP) and Substance P (SP)] have been described in slow transit constipation [29, 95].

Several other neuropathies such as diabetic neuropathy, Chagas disease, hypertrophic pyloric stenosis, intestinal neuronal dysplasia, toxic megacolon and internal anal sphincter achalasia affect the ENS [26, 29, 96] but because of their clear secondary origin are not discussed in this review. Also, neurodegenerative diseases including Alzheimer’s and Parkinson’s disease, and neurodevelopmental disorders such as autism spectrum disorder present with gastrointestinal symptoms linked to ENS dysfunction. We guide the interested reader to excellent literature discussing the role of the ENS in these disorders [22].

Regardless of the nature of the disorder, it is indisputable that the ENS performs a major role in the progression of enteric neuropathies, displaying mild to profound alterations in its structure and causing modest to severe symptoms. However, for most enteric neuropathies other than HSCR, it still remains elusive whether the ENS itself has the ability to be the sole participant in disease initiation. Moreover, as preeminent performers of ENS physiology, studies on the influence of EGCs on the etiology of enteric neuropathies remain undervalued and scarce. Advanced high throughput technology has recently provided novel perspectives on possible functions the ENS may portray in disturbing gastrointestinal homeostasis [12, 48], for instance by expressing disease risk genes and through interactions with other cellular systems.

Disorders of gut–brain interaction

Disorders of gut–brain interaction (DGBIs), previously more commonly known as functional gastrointestinal disorders (FGIDs), including functional dyspepsia (FD) and irritable bowel syndrome (IBS), define a spectrum of gastrointestinal disorders associated with chronic or fluctuating gastrointestinal symptoms, such as abdominal pain, diarrhea, constipation, bloating and nausea, without harboring an apparent organic structural or biochemical explanation for these symptoms [97, 98]. With evidence in support of a more organic basis accumulating, the most recent iteration of the Rome diagnostic criteria has facilitated a gradual shift from the ‘functional’ nature of these gastrointestinal disorders. Although the pathogenesis of DGBIs remains ill-defined, dysbiosis, visceral hypersensitivity, intestinal dysmotility, and gut–brain axis dysregulation interplays are likely mechanisms responsible for DGBIs [99,100,101,102,103,104]. Still, their heterogeneous identity makes it unlikely that disease mechanisms can be narrowed down to a single pathophysiological process. Nevertheless, with 40% of persons worldwide anticipated to meet the criteria for DGBIs [105], DGBIs are the most frequent cause of gastroenterological consultations; thus, resulting in major economic effects on global health care systems [106]. Their symptom complex is indicative of altered ENS function. However, only minor evidence of ENS pathology, such as lymphocyte infiltration in myenteric ganglia and increased neurite density in the intestinal mucosa, as well as production of auto-antibodies against neural antigens have been reported, thwarting efforts to identify their intrinsic neurogenic origin [107,108,109,110]. Although a pioneering study by Cirillo and colleagues demonstrated that live cell imaging can be used to probe neuronal function in intestinal biopsies [111], ENS defects are not simple to diagnose if ganglia are present, that is not to say that the presence of ganglia averts gastrointestinal anomalies. Consistent with this idea, the disturbed gastrointestinal function in DGBIs could result from subtle alterations in ENS circuitry that have gone undetected in routine clinical diagnosis [6, 25]. However, to date, only a few studies have focused on mechanistic defects in ENS connectivity that manifest later in ENS ontogeny and are unaccompanied by aganglionosis. Inactivation of Celsr3, a planar cell polarity (PCP) gene critical for guidance and directional growth of neural processes during murine embryogenesis, in enteric neural crest cells, leads to moderate disruptions of axonal tract configuration in the mature ENS but elicits profound uncoordinated motor activity analogous of DGBIs [112]. Similar experimental strategies using novel ENS-specific transgenic models hold great promise to further refine our understanding of gut motility deficits that do not typify the downstream aftermath of malfunctioning gangliogenesis. Interesting candidates for such approaches include NXPH1 (neurexophilin 1), SLC6A4 (serotonin re-uptake transporter, SERT) HTR3E (serotonin receptor type 3E) and HTR4 (serotonin receptor type 4) [113,114,115,116,117,118,119]. Others might be identified from genome-wide association (GWA) studies [120,121,122]. It is plausible that in this way particular gastrointestinal disorders now labeled as ‘functional’, will eventually be identified as enteric neuropathies.

Whilst the cellular outline of the ENS has mostly been configured by birth, the maturation of the neural circuits is finalized amidst > 100 trillion microbes harbored within the postnatal gut [64]. Although the exact means by which the microbiome shapes ENS circuits are unnamed, it is now well-established that the microbial composition impacts the ENS framework, and that the ENS-microbiome work alongside in pathological conditions. Depletion of microbiota beget reductions in myenteric nerve fiber density, alterations in neurochemical coding, decreased excitability in intrinsic primary afferent neurons, defective EGC networks, changes in neurogenic colonic migrating motor complexes and protracted intestinal transit time [69, 72, 123,124,125,126], and these effects can be restored following recolonization of adult germ-free mice with conventional microbiota [65, 69, 72, 126]. The linking of microbiota to ENS regulation of motility is one that appears to be evolutionary conserved and can be traced back to 650 million years ago to cnidarians, such as the Hydra [127, 128]. In germ-free Hydra, there is a reduction in the spontaneous peristaltic movements which are under the regulation of neurons and restoration of these movements are observed following the reconstitution of germ-free Hydra with conventional microbiota [129,130,131]. These findings illustrate that the ENS has evolved to orchestrate responses from microbes and relay them throughout the gastrointestinal tract to influence gut motility. Altered gastrointestinal motility is an important factor in DGBIs [97] and moreover, gut microbiota dysbiosis has been implicated in the onset and progression of DGBIs [132, 133]. Obata et al. [65], identified that enteric neuron-specific deletion of a microbiota-dependent gene, aryl hydrocarbon receptor (Ahr), reduced colonic peristaltic activity akin to microbiota-depleted mice, and supplementation of Ahr ligands rectified intestinal motility; alluding to Ahr as a biosensor in enteric neurons, therein fusing their functional output with the luminal environment. Interestingly, a GWA meta-analysis study identified a role for AHR in the biology of stool frequency, often altered in DGBIs [134], and deficiencies in tryptophan-derived Ahr ligands produced by microbiota may contribute to IBS [135, 136]. Furthermore, together with genetic changes in the serotonergic signaling system (see above), the alterations in enterochromaffin cells observed in IBS patients [137, 138] and the impact of the microbiota on this system [72, 139], highlight the importance of the ENS-microbiome dialogue. A dialogue that if perturbed may contribute to the pathogenesis of DGIBs.

The most well-recognized risk factor for DGBIs is previous acute gastrointestinal infection. Permutations in gut milieu post entero-invasive bacteria evasion are associated with remodeling of the neuronal circuitry [140,141,142], albeit the underlying mechanisms involved in this process are incompletely understood. Recently, Matheis et al. [143] showed that Salmonella enterocolitis induced preferential damage to glutamatergic myenteric neurons expressing specific non-canonical inflammasome machinery and pointed to a neuroprotective role of ENS-associated tissue-resident macrophages mediated through an arginase 1-polyamine axis. As targets of a vagal-cholinergic enteric neuron anti-inflammatory pathway [144], and in keeping with their vital role in ENS survival and function [145, 146], this data touts that muscularis macrophages can temper with the enteric-inflammasome circuit and ergo, prevent remolding of the neurochemical representation of enteric neurons. Defective maturation and altered numbers of macrophages have been documented in IBS [147, 148]. Among other functions, macrophage-colony stimulating factor (M-CSF) governs survival, proliferation, and activation of macrophages, and while interstitial cells of Cajal also produce M-CSF, within the ENS, EGCs generate the vast majority of M-CSF [149,150,151]. Grubišić et al. [150] hindered enteric glial intercellular communication mitigated through connexin-43 hemichannels in a chronic inflammatory mouse model and found that curbing the EGC-regulated activation of macrophages via M-CSF safeguarded against the development of visceral hypersensitivity, another key mechanism in DGBIs. This provides proof of the importance of EGCs in shaping neuronal transmission through the EGC-macrophage axis. Nevertheless, this does not exclude the odds that EGCs interact with other immune entities to tweak ENS circuitry and their functions likely transcend across other gastrointestinal dysfunctions linked to DGBIs.

Although there is a vast amount of literature focusing on visceral hypersensitivity acting through extrinsic afferent pathways [152,153,154,155,156,157,158,159], inflammatory mediators can also modify ENS components, thereby facilitating the development of DGBIs [160, 161]. Increased infiltration and activation of mast cells [162,163,164,165], together with elevated levels of mast cell mediators, including histamine and proteases, but also relevant miRNAs and neurotrophins have been substantiated in DGBIs [166,167,168,169]. A positive correlation between the number of mast cells in close contact with enteric neurons and the degree of abdominal pain was demonstrated in IBS patients [170], albeit, this is not necessarily reflective of a causal relationship. Neurons can trigger mast cell degranulation and mediator release by means of neuropeptides such as vasoactive intestinal peptide (VIP) [171] which, in turn, invoke the egress of histamine and other mediators from nearby mast cells [172], thus, driving neurogenic inflammation [173]. Reciprocally, the vast quantities of mast cell-derived mediators in mucosal supernatants obtained from IBS patients duly triggered enteric neuron excitability with a favored heightened activation of submucosal plexus neurons [174,175,176]. What is more, chronic exposure to mucosal mast cell-derived mediators from IBS patients not only stimulated neuronal activation and sprouting of enteric neurons but also evoked hyperexcitability of visceral and somatic pain pathways [107, 177,178,179,180]; this is suggestive of an association between visceral hypersensitivity and long-lasting enteric neuronal plasticity. Visceral afferent sensitization has been shown to act, for example, through changes in the expression and function of transient receptor potential vanilloid receptor type-1 (TRPV1) and t-type calcium Cav3.2 channels [181,182,183]. However, whether TRPV1, which has been the focus of many studies on visceral hypersensitivity [184], is expressed by enteric neurons is still a matter of debate [185,186,187,188]. Recent single-cell mRNA sequencing data and studies using transgenic Trpv1-Cre reporter mice seem to further support the notion that TRPV1 expression observed in the gut wall is of extrinsic neuronal origin [9, 12, 13, 189, 190]. It, therefore, remains to be determined whether submucosal enteric neurons are involved in the reduced abdominal pain and visceral hypersensitivity observed in IBS patients after histamine receptor H1 (H1R) blocking [191]. Of interest, histamine, via H1R, mediated ATP-induced Ca²⁺ responses in EGCs and a reduction in S100B + EGCs in the colonic mucosa of IBS patients was inversely coupled with abdominal pain [192]. This study infers that restoring EGC network activity could provide an effective strategy to combat pain in IBS patients. Additionally, tachykinins, neuropeptides primarily expressed by neurons in the gut [190], are also tied to distortions in both pain transmission as well as intestinal motility [193,194,195,196], and inhibition of tachykinin pathways offers a potential target for treatment in IBS patients [197]. A tachykinin receptor named neurokinin-2 receptor (NK2R) is chiefly expressed by enteric neurons; and stimulation of NK2R can function as a potent driver of neurogenic inflammation in the ENS through mechanisms that implicate intercellular enteric neuron-glia-nociceptor communication [190]. Mechanistically, antagonizing NK2R signaling restricted the occurrence of a reactive EGC phenotype and increased neurogenic contractions, raising the possibility of a role for EGCs in DGBIs [190, 192].

Going forward, by combining the recent advances that have been made in our understanding of ENS composition and function with the novel findings from DGBIs patient studies [198,199,200,201], it should be possible to further elucidate the ENS culprits in DGBIs and bring about therapies targeting the second brain.

Inflammatory bowel diseases

IBD, a collective term used to describe prolonged inflammation of the gastrointestinal tract, primarily include Crohn’s disease (CD), which can present throughout the entire gastrointestinal tract (the most common localization being the terminal ileum), and ulcerative colitis (UC), presenting in the mucosal layer of the colon. En masse, affecting 2.5–3 million people in Europe [202], IBD is univocally identified as an immune pathology and is believed to develop through interactions between environmental, microbial, and immune-mediated factors in a genetically predisposed host [203, 204]. Nonetheless, it has become apparent that the immune system alone may not account for all aspects of IBD pathology. Neural fiber hypertrophy and hyperplasia in the mucosa, submucosal and myenteric plexus concomitant with infiltration of inflammatory cells, submucosal structural defects, neural fiber retraction and neuromatous lesions have been observed in the inflamed gut of IBD patients. This brought into question whether ENS aberrations are mere epiphenomena related to the inflammatory reaction or that they are involved in IBD pathophysiology, or even, that they can act as predictors of IBD recurrence [205,206,207,208,209]. Echoed by prevailing inflammation-induced derangement of intrinsic neural circuits [210], including neuronal hyperexcitability [211, 212], increased synaptic facilitation [212, 213], reduced descending inhibitory neuromuscular transmission [214,215,216] and even neuronal death [217], long-lasting intestinal dysmotility can persist ensuing inflammation resolution [218]. Moreover, in vivo studies using transgenic mouse models for enteric hyper- and hypoinnervation indicate that the severity of intestinal inflammation is affected by the density of the enteric neurons [219].

The fact that distinct immune cell subsets of the gut are equipped to respond to neuron-derived signals by expressing neurotransmitter and neuropeptide receptors and inversely, that enteric neurons can respond to inflammatory signals via, for instance, expression of cytokine receptors, posits the existence of functional ENS-immune interactions in the modulation of intestinal inflammation [220,221,222,223]. Deciphering these neuro-immune units will likely generate important clues on the role of the ENS in IBD, but will also be relevant for post-infectious DGBIs and those associated with low-grade inflammation. This is elegantly illustrated in a neoteric study from Jarret et al., in which deletion of enteric neuron-derived interleukin-18 (IL-18), a pleiotropic cytokine in intestinal barrier homeostasis and combatting pathogenic infections [224], resulted in increased susceptibility to invasive Salmonella typhimurium [225]. Intriguingly, IL-18 is elevated in patients with IBD and anti-IL-18 therapy has been proven efficient to reverse severe gastrointestinal symptoms [226, 227].

Several neuropeptides and neurotransmitters have been shown to govern immunological functions through discrete subsets of immune cells [228]. SP, a highly conserved neuropeptide that mediates ENS signaling as well as immune cell proliferation and cytokine production, is elevated in IBD and there is evidence of increased numbers of SP-positive neurons in the myenteric plexus of UC patients [229,230,231]. VIP is also involved in a vast number of gastrointestinal functions mediated by the ENS and increased VIP-positive neurons in the submucosal plexus have been documented in CD patients [232, 233]. Notably, mast cells reside in close juxtaposition with intestinal peptide-producing neurons [234] and respond to neuron-derived SP and VIP secretion resulting in mast cell degranulation and cytokine production [235]. Reciprocally, enteric neurons are activated by mast cell-derived mediators inducing neuronal hyperexcitability [236,237,238]. Conforming to this, higher numbers of mast cells or mast cell mediators in immediate proximity to enteric neurons has been revealed in colonic biopsies of IBD patients relative to controls [239,240,241,242], mirroring observations in patients with IBS. Confirmed by live-cell imaging experiments in human intestinal preparations [173], these findings indicate that there is bidirectional communication between mast cells and enteric neurons that might be significant in perpetuating ongoing inflammation.

VIP has also been implicated in the stimulation of innate lymphoid cells (ILC) [243], a rapid-responding, highly abundant group of immune cells found in gut tissues that play a sentinel role in intestinal barrier integrity [244]. Following microbial infection, ILC3, express high levels of the VIP receptor VIPR2, and are the first ILC responders to initiate immune responses in the gastrointestinal tract [245]. Analogously, VIP-producing enteric neurons enhance the release of tissue protective interleukin-22 (IL-22), thereby, serving broadly protective effects on the intestinal epithelium [246]. Genetic deletion of VIP signaling disturbs ILC3 production of IL-22 and leads to increased susceptibility to inflammation in experimental models of colitis [243], and accordingly, dysregulation of ILC3 content and function was also shown in IBD [247]. Enteric cholinergic neurons were also found to promote type 2 inflammation through the release of another highly conserved neuropeptide, neuromedin U (NMU) [248,249,250]. Interestingly, one of its receptors, NMUR1, is selectively enriched in type 2 ILC (ILC2) and appears to be targeted by NMU-producing cholinergic neurons. The NMU-NMUR1 pathway has been shown to activate ILC2 to produce type 2 effector and tissue-repair cytokines in a Myd88-dependent manner.

Comparable to the NMU-mediated signaling from enteric neurons to ILC2, EGCs can respond to microbial-associated molecular patterns in a Myd88-dependent manner by releasing the RET-ligand glial cell line-derived neurotrophic factor (GDNF), which stimulates neighboring ILC3s-expressing the RET receptor to release IL-22, thus, mediating epithelial repair [251]. This work suggests that, in concert with enteric neurons, EGCs also shape the gastrointestinal immunological environment in the course of gut inflammation and aid in facilitating the maintenance of gut homeostasis. Although in experimental models of colitis and IBD-derived human samples the exact source of GDNF in the gut is not unequivocal, increased GDNF expression showed anti-apoptotic epithelial effects [252, 253], and is strongly upregulated in inflamed areas of CD and UC [252, 254] lending further support to the beneficial role of EGCs in modulating epithelial barrier function. However, reminiscent of astrocytes of the CNS, EGCs also undergo phenotypic alterations in response to persistent hyperinflammatory insults, inter alia an upregulation of glial fibrillary acidic protein (GFAP) [255] and may contribute to the influx of microbial infection by reducing their release of beneficial factors that enhance gut barrier function, such as 15-hydroxyeicosatetraenoic (15-HETE) [256] and S-nitrosogluthathione (GSNO) [257]; although, it is unknown whether this defective gliotransmitter release is causal or consequential of disease progression. Aside from a reduction in beneficial gliotransmission, the gut microenvironment can also shift EGCs towards a phenotype that releases specific gliomediators, such as nitric oxide (NO) [258]. IBD is associated with a discernible NO-dependent inflammatory response within the intestinal milieu; mucosal inducible nitric oxide synthase (iNOS) and neuronal nitric oxide synthase (nNOS) in patients with inflamed and non-inflamed UC were upregulated and downregulated, respectively [259]. Pursuant to this, there have been reports of EGCs coordinating the NO-dependent inflammatory response through the circulation of their signaling molecule, S100B, contending that S100B release may antedate the onset of inflammation [260, 261]. Deconstructing the mechanisms by which EGCs and S100B modulate the inflammatory scenario in IBD warrants further research; however, one emerging theme is through the activation of toll-like receptors (TLRs) [262,263,264,265]. Differential expression of TLR subtypes has been reported in IBD patients [266, 267] and strikingly, absence of TLR4, for which polymorphisms have been documented in IBD [268, 269] and which is expressed by enteric neurons and EGCs [270,271,272], affects S100B and GFAP expression and enhances inhibitory neurotransmission and neuronal death through the interaction of NO with purinergic signaling in murine models [273, 274].

From a clinical standpoint, studies highlighting a mechanistic role for the ENS in IBD are rather scarce. Recent profiling of human intestinal tissue at single-cell resolution discerned that IBD risk genes along with multiple genes related to cytokine signaling were enriched in the ENS [12]. Interestingly, a transcriptomic study from the same laboratory identified differential gene expression in a glial subset present in mucosal biopsies obtained from UC patients and suggests glial involvement in both the regulation of T cells and M-like cells together with changes in tumor necrosis factor (TNF) signaling [275].

It is clear that specific neuro-immune units act to integrate immune control in the intestine, a partnership which can be traced back to organisms as primitive as C. elegans [276]. For example, recent work revealed a role for sensory neurons in the regulation of innate immunity during larval development by promoting immune effector transcription, clearance of an intestinal pathogen and resistance to bacterial infection [277]. Increasing evidence also pinpoints the ENS as a crucial player in the interaction between the intestinal microbiome and host defense mechanisms [278], albeit, for now, not in the specific context of IBD. Nonetheless, dysregulation of such interplay likely contributes to intestinal inflammation and further interrogation of the cellular players and mediators perhaps will shed light on the neuro-immune and neuro-microbiome scaffolds conducive to sustaining tissue homeostasis.

Colorectal cancer

Therapeutic developments and screening measures for early detection have improved survival rates for CRC patients. However, the fact that CRC remains the third most common malignancy and the second leading cause of cancer mortality [279] has prompted for a deeper understanding of cancer progression; shifting the focus from genetic and epigenetic aberrations to the contribution of the tumor microenvironment [280,281,282,283]. The composition of the tumor microenvironment contains an assortment of cellular players such as endothelial cells, pericytes, fibroblasts, myofibroblasts, immune cells and neural cells [284, 285]. Although research on the role of nerves in cancer is a growing topic in the oncology field [286] and increased neural markers in colorectal tumors converge with poorer prognosis [287], enteric neurons and glial cells have been largely neglected. The relevance of neural infiltration for tumor progression was first believed to be a passive process in which nerves act as roots for cancer cell migration. This has mainly been studied in the context of perineural invasion (PNI), which in CRC, is associated with a reduced disease-free survival and can serve as an independent prognostic marker for patient outcome [288, 289]. However, active crosstalk contributing to cancer growth, mostly involving paracrine signaling between neurons and cancer cells has been described for several cancer types including pancreatic and gastric cancer [286]. This involves the expression of neurotrophins and axon guidance molecules by cancer cells to induce their own innervation, and active communication from neurons to cancer cells [290,291,292,293,294].

To date, several lines of evidence hint towards a role of the ENS in colorectal carcinogenesis [295, 296]. For instance, the expression of netrin-1, a tumor suppressor protein synthesized by enteric neurons during gastrointestinal organogenesis [297], is found to be reduced in CRC [298, 299]. Also, although the increased cholinergic innervation observed in later phases of CRC indicating poorer prognosis was attributed to parasympathetic extrinsic nerves [300], the extensive population of intrinsic cholinergic neurons could be involved in exacerbating CRC. Interestingly, reduced risk of CRC is observed in patients with megacolon presenting with intrinsic colonic denervation of the intestine [301], suggestive of a protective role of diminished intestinal innervation. This is also in line with a study where chemical ablation of neurons in rats is correlated with a lower incidence of CRC [302]. Proof of tumor epithelial cells docking and migrating along enteric neuronal fibers was reported by Duchalais et al. [303], highlighting the potential of the ENS in steering the migration of CRC. Accounting for the vastness of ENS networks and its extensive interactions with the extrinsic nervous system, it is tentative to speculate that the extra physical support from the ENS could enable tumor cells to travel large ranges and penetrate neighboring tissues, which is consistent with the prognostic value of perineural invasion in CRC [288, 289]. Recent findings by Valès and colleagues indicate that, next to enteric neurons, EGCs could also play a role in the CRC tumor microenvironment [304]. In response to tumor-derived ligands, EGCs converted towards a pro-tumorigenic phenotype, stimulated the activation of cancer stem cells which present tumor-initiating capacities, and in turn promoted tumorigenesis. Additionally, elevated levels of S100B observed in CRC patients might foster a pro-malignant micro-environment, regulating various pro-inflammatory, angiogenic and anti-apoptotic factors [305]. Remodeling of EGCs in the earlier stages of colon carcinogenesis is not unforeseen, weighing the extensive phenotypic plasticity of EGCs in the face of a dynamic environment [8]. Despite preliminary, the above findings herald the ENS as an active player in CRC. From a therapeutic perspective, understanding the molecular mechanisms that allow functional adaptations of ENS signaling during tumorigenesis represents an exciting challenge.

Conclusions and future perspectives

The precise etiology of enteric neuropathies, IBD, DGBIs and CRC is unknown, and ascertaining if a malfunctioning cell type is the trigger in these disorders or instead a consequence, is not always straightforward. Their heterogeneous presentation and multifactorial nature also withhold us from taking an ENS-centric view. However, alterations in the number of enteric neurons and neuronal subtype composition are implicated in enteric neuropathies, IBD, and DGBIs, though to a lesser extent, alluding that abnormalities in genes or signaling pathways governing ENS circuits are universal across these disorders and may contribute to their pathogenesis. This is corroborated by the concurrence of IBD and IBS, the presence of IBS-like symptoms in IBD patients in apparent remission, the mutual mucosal immune activation and neurochemical coding alterations. Some of which is ultimately reflected in the interchangeable application of in vivo models by researchers investigating both DGBIs and IBD. What is more, while it is well-established that the majority of mutations identified in HSCR are attributed to the RET signaling pathway, antagonizing RET signaling has now also been shown to attenuate post-inflammatory and visceral hypersensitivity in pre-clinical IBS studies [306]. Thus, by impacting motor, sensory and immune functions of the gut, albeit to a variable degree depending on the clinical phenotype, the absence or altered ratio of discrete ENS constituents seems to be a common denominator underpinning the characteristics of these disorders. This leads us to theorize that the ENS may be accountable for a gamut of disorders extending from hypoganglionosis to conditions with an increased enteric neuron density (Fig. 1). Situated within the confines of this spectrum, subtle changes in the connectivity of ENS circuits likely contribute to the pathophysiology of gastrointestinal disorders. However, together with the arcane mechanisms of enteric neuron wiring, the ramifications of faulty ENS connections remain an understudied area. Carving a distinction between the involvement of the ENS in these disorders is quite arduous as the functional-organic dichotomy between some enteric neuropathies, DGBIs and IBD is possibly archaic. Nevertheless, a more comprehensive understanding of ENS development, function and its modes of failure could be crucial in our pursuit of establishing the foundations of some of these diseases and may point towards the endowment of ENS factors responsible for their pathogenesis. Undeniably, gastrointestinal physiology is not solely reliant on a normal ENS but is also influenced by multilateral interactions between the ENS and other systems such as the intestinal epithelia, immune system and microbiome. Evidence gathered in the past few years demonstrates that the neural-immune-microbiome entente, for which the assembly can be retraced to primitive species, is pivotal in a disease context as aberrations in their crosstalk result in variable degrees of abnormal gut function, exemplified in the described disorders.

Fig. 1

Graphical representation of gastrointestinal disorders from an enteric nervous system perspective. At homeostasis, enteric nervous system (ENS) components (enteric neurons and enteric glial cells) are exposed to and work in concert with the outer and inner microenvironment of the gut to regulate bowel motility, transmucosal movement of fluids, immune responses and local blood flow. Dysfunctions in the ENS may be accountable for a gamut of disorders extending from hypoganglionosis (as observed in some enteric neuropathies) to conditions with a conceivable increased enteric innervation (colorectal cancer). Within this spectrum, subtle changes in ENS circuits as well as alterations in the ENS-immune, ENS-epithelium/enteroendocrine axis and ENS-microbiota axis likely contribute to the pathophysiology of disorders of gut–brain interaction (DGBIs) and inflammatory bowel disorder (IBD). ENS-associated tissue resident macrophages are crucial for normal ENS function and play key roles across the disease spectrum. Defective maturation and altered numbers of macrophages have been documented in DGBIs and IBD. Also, higher mast cell numbers and mediators are present in DGBIs and IBD, while dysregulation of innate lymphoid cells-ENS circuits has been observed in IBD. Disturbances of the physiological microbiome composition can be both a cause or consequence of these disorders. Created with Biorender.com

Unified interdisciplinary efforts will be needed to further disentangle the Daedalian biology of the ENS and its cross-talk with other cellular systems [21]. Coupling such insights with in vivo gene editing strategies of candidate genes emanating from recent high-resolution transcriptomic mapping and advances in human intestinal organoids [307, 308] together with novel live-cell imaging modalities [309] will help consolidate the current platforms to model these disorders. In addition, the development of novel ENS-based therapeutic strategies warrants human studies to confirm the clinical relevance of experimental findings.