MRs are expressed by a wide variety of tissues and cell types and control key digestive and metabolic functions. Salivary gland secretion, gastric, and intestinal fluid transport, cell proliferation, mucus production, motility, and mesenteric vascular constriction and dilation are all responsive to MR signaling (Tobin et al., 2009; Muise et al., 2017). In the stomach, M₃R, M₄R, and M₅R activation modulates hydrochloric acid secretion from parietal cells (Aihara et al., 2005), and M₁R and M₃R activation stimulates pepsinogen secretion from chief cells (Xie et al., 2005). GI motility, through intestinal smooth muscle cell action, involves communication between the central and enteric nervous systems. These effects are partially mediated by M₁R through M₃R (Moro et al., 2005), with M₂R and M₃R playing a role in regulating longitudinal muscle contraction, and all three MR subtypes involved in circular muscle function (Harrington et al., 2010; Tanahashi et al., 2021). MR-mediated regulation of smooth muscle function extends throughout the entire GI tract. Nonetheless, it is likely that MRs play additional roles in regulating small intestinal function; for example, M₂R is expressed in the stem cell compartment and may be involved in enterocyte turnover (Muise et al., 2017). ACh has both pro- (Koyama et al., 1992; Brunn et al., 1995) and anti-inflammatory (Pavlov and Tracey, 2006) effects, the latter mediated in part by reducing systemic levels of tumor necrosis factor. While the ubiquity of MRs within the digestive tract makes them attractive therapeutic targets to modulate health and disease, this same ubiquity complicates efforts to design selective agents while minimizing off-target adverse effects.

Responses of MRs to ligand binding are subtype specific. Activation of odd-numbered MRs (M₁R, M₃R, and M₅R) stimulates phospholipid turnover and increases intracellular calcium levels while activation of even-numbered MRs (M₂R, M₄R) inhibits adenylyl cyclase activity, thereby reducing levels of intracellular cAMP. M₁R, M₃R, and M₅R (MR_odd) canonically couple to G_q/11 which induces the phospholipase C-mediated hydrolysis of phosphatidylinositol (4.5)-bisphosphate into diacylglycerol and inositol (1,4,5)-trisphosphate. The latter binds an endoplasmic reticulum receptor stimulating intracellular calcium release. However, these may represent oversimplifications; experimental findings suggest differential interactions of individual MR_odd and MR_even with their downstream targets. For example, although both M₁R and M₃R signal through phospholipase C, CHO cells expressing M₁R exhibited four-fold greater cAMP production in response to carbachol compared to cells expressing M₃R (Burford et al., 1995). Likewise, although M₂R and M₄R (MR_even) act primarily by binding G_i/o family proteins to alter adenylyl cyclase activity, their actions can also prolong potassium channel opening, thereby causing cellular hyperpolarization (Bubser et al., 2012).

These general principles do not tell the whole story–despite substantial sequence homology among MR subtypes they demonstrate surprising individuality in their responses to stimuli, even within the same cell and when responding to the same ligand. Pancreatic acinar cells provide a useful model to study muscarinic control of exocrine digestive function. Using acinar cells prepared from M₁R- and M₃R-deficient mice as well as M₁/M₃ chimeric receptors, Nakamura et al., demonstrated greater ACh-induced IP₃ release in cells expressing only M₁R compared to those expressing uniquely M₃R (Nakamura et al., 2013). Moreover, in M₃R-compared to M₁R-expressing cells, these differences were associated respectively with oscillatory versus monotonic patterns of cytosolic calcium release. Oscillatory calcium release was a function of a C-terminal region of M₃R with considerable variability among MR subtypes (Nakamura et al., 2013). In murine gastric chief cells, both M₁R and M₃R mediate pepsinogen secretion–deletion of either MR subtype reduces and combined M₁R and M₃R deficiency ablates cholinergic agonist-induced proenzyme secretion (Xie et al., 2005). Thus, in some cell types, MR_odd have overlapping functions whereas in other cell types, MR subtype signaling appears divergent. In addition to the influence of their cell and tissue localization, other mechanistic differences between MR subtypes result in sometimes-opposing effects. GPCRs, including MRs, can also undergo “pre-coupling”, wherein a stable multimeric complex is present before ligand binding. Unlike other MR subtypes, M₁R and M₃R pre-couple with G_i/o G-proteins, their non-preferential G protein, thereby potentially altering downstream effects (Jakubík et al., 2011). Crystal structures of inactive M_1-4R subtypes provide some insight into different allosteric and orthosteric binding sites (Kruse et al., 2012; Thal et al., 2016), but our understanding of the resulting functional differences between MR subtypes continues to evolve.

As a result of their central role in maintaining homeostasis in the GI tract, dysregulated MR signaling can be an important modifier of intestinal disease. In Hirschsprung disease, lack of mucosal cholinergic innervation in aganglionic colon segments increases the risk of postoperative enterocolitis (Keck et al., 2021). In diarrhea-predominant irritable bowel syndrome (IBS-D) without a concomitant psychiatric disorder, pyridostigmine (an acetylcholinesterase inhibitor) induces a stronger IL-6 response that is highly correlated with symptoms (Dinan et al., 2008). Given the pharmacotherapies targeting MRs already approved or being explored to treat IBS (Tables 1, 2), achieving a more precise mechanistic understanding of the role MR dysregulation plays in IBS is important.

Diseases associated with MR dysregulation are not restricted to the lower GI tract. In the stomach, cholinergic signaling is balanced with histamine and gastrin release to regulate gastric acid levels; peptic ulcer disease is associated with greater MR expression in the gastric body, whereas progressive MR loss in that region is associated with chronic gastritis (Pfeiffer et al., 1995). In progressive systemic sclerosis and Sjogren’s syndrome, an autoimmune condition which impairs lacrimal and salivary function, esophageal dysmotility may be associated with anti-M₃R antibodies (Goldblatt et al., 2002; Kawaguchi et al., 2009; Gyger and Baron, 2012); anti-M₃R antibodies are also reported in progressive systemic sclerosis with anal dysmotility (Singh et al., 2009; Gyger and Baron, 2012). Intravenous immunoglobulin to neutralize anti-M₃R antibodies may be beneficial (Smith et al., 2005).

Compared to the normal liver, individuals with primary biliary cholangitis (PBC) are more likely to have a CHRM3 single nucleotide polymorphism (rs4620530) of uncertain significance; this is not associated with baseline disease characteristics or treatment responses (Greverath et al., 2020). PBC is more commonly associated with anti-M₃R antibodies than other liver diseases (Tsuboi et al., 2014); those with anti-M₃R antibodies are more likely to have a benign disease course. Nonetheless, M₃R antibody levels do not correlate with treatment responses or serological markers either at baseline or during the disease course (Mayer et al., 2020). A subset of patients with PBC develop Sjogren’s syndrome; the shared increase in anti-M₃R antibody levels in both conditions suggests overlapping features could form the basis for a mutual treatment.

MRs play key roles in normal cell proliferation and turnover. As reviewed by Campoy et al. (2016), presumably to benefit tumor progression, neoplastic cells hijack MR-dependent proliferative signal transduction pathways. Treating neoplastic cells with exogenous ACh and inhibiting ACh hydrolysis promotes their proliferation and, conversely, reducing M₃R expression and activation is anti-proliferative. Moreover, because neoplastic cells tend to lose cellular polarity, receptors normally expressed on the basolateral membrane may be expressed more diffusely around the cell membrane, thereby facilitating their access to orthosteric and allosteric ligands in the tumor microenvironment and GI lumen (Cheng et al., 2002). For example, bile acids, at concentrations achieved in stool, promote atropine-inhibitable colon cancer cell proliferation (Cheng and Raufman, 2005).

Abundant data support the conclusion that M₃R plays an important role in colon cancer progression. In mouse models of sporadic and genetic colon cancer, using azoxymethane (AOM)-treated and Apc ^Min/+ mice, respectively, Chrm3 ablation with resulting M₃R deficiency substantially reduces the intestinal tumor burden (Raufman et al., 2008; Raufman et al., 2011). As M₃R deficiency primarily reduces the number of adenocarcinomas rather than adenomas, the major impact of blocking M₃R activation appears to be on promotion, rather than initiation, of neoplasia. M₃R activation has similar pro-proliferative effects on gastric cancer (Hayakawa et al., 2017; Wang et al., 2018). M₃R expression is enhanced in cholangiocarcinoma and associated with reduced cell differentiation, perineural invasion, and metastasis (Feng et al., 2012; Feng et al., 2018).

In contrast to the impact of M₃R deficiency, M₁R deficiency in mice does not attenuate, and may modestly enhance, AOM-induced colon carcinogenesis. Strikingly, mice with combined M₁R and M₃R deficiency develop as many colon tumors as control mice (Cheng et al., 2014); that is, M₁R deficiency negates the anti-neoplastic effects of M₃R deficiency. Likewise, M₁R agonism appears protective against pancreatic ductal adenocarcinoma (PDAC) and counteracts enhanced carcinogenesis following vagotomy (Renz et al., 2018), suggesting a potential therapeutic opportunity. In contrast, in hepatocellular and prostate carcinomas, M₁R activation promotes cellular migration and invasiveness (Yin et al., 2018; Zhang et al., 2020). Notably, many of these studies are limited by using global rather than conditional knockout mouse models. Hence, it remains uncertain whether the respective MR deficiencies are due to effects on neoplastic cells versus other cellular elements in the tumor microenvironment (e.g., immunocytes). Nonetheless, these observations argue strongly for the importance of MR subtype selectivity in designing and developing therapeutics.

Branches of the vagus nerve, a major source of ACh signaling within the GI tract, innervate the liver and modulate hepatocyte regeneration by progenitor cells and fibrosis by stellate cells (Cassiman et al., 2002). The current lack of effective anti-fibrotic therapies highlights the potential of leveraging these muscarinic actions to prevent or reverse fibrosis in advanced liver disease and stimulate hepatocyte regeneration. For example, in rodents, carbon tetrachloride (CCl₄)-induced hepatic fibrosis can be attenuated by vagotomy and treatment with atropine (Lam et al., 2008). M₃R expression and activation protects against AOM-induced liver fibrosis (Khurana et al., 2010; Khurana et al., 2013; Rachakonda et al., 2015). Surprisingly, M₁R expression and activation appears to have opposite effects, worsening AOM-induced hepatic fibrosis (Rachakonda et al., 2015). Thus, in the absence of effective anti-fibrotic therapy, manipulation of MR subtype activity to limit or reverse fibrosis may have therapeutic potential although, again, divergent effects in different tissues warrants caution.

MR subtype, tissue distribution, and off-target side effects have hindered efforts to manipulate MR activity precisely and effectively with drugs. MR antagonists are most effective in treating chronic obstructive pulmonary disease and overactive bladder (Table 1) (Eglen et al., 1999; Athanasopoulos and Giannitsas, 2011)—their utility for GI and hepatic disorders is currently limited. Cholinesterase inhibitors that increase ACh levels, also used clinically for digestive tract disorders, have similar limitations as their actions are largely non-selective. Adverse effects with these classes of drugs are attributed primarily to off-target effects on the CNS (e.g., convulsions, confusion) and other peripheral MR subtypes (e.g., sialorrhea, rhinitis, diaphoresis, diarrhea, nausea, vomiting, and bronchospasm). Novel MR agonists and antagonists are currently under investigation primarily for diseases of the central nervous system such as Alzheimer’s disease and schizophrenia (Table 2) (Davie et al., 2013).

Several MR agonists and cholinesterase inhibitors are in clinical use. Oral and topical pilocarpine. and cevimeline (Evoxac), an M₃R-selective activator, augment salivary gland secretions in xerostomia due to radiation therapy and Sjogren’s syndrome (Iga et al., 1998; Fife et al., 2002; Petrone et al., 2002; Weber and Keating, 2008a; Berk, 2008; Mitchelson, 2012a; Davies and Thompson, 2015; Panarese and Moshirfar, 2021). Bethanechol, a structural analogue of ACh that resists hydrolysis by cholinesterases, has potential to treat esophageal dysmotility. Currently approved to treat urinary retention and neurogenic bladder (Gaitonde et al., 2019), bethanechol strengthens esophageal contractions in subjects with ineffective esophageal motility (Agrawal et al., 2007) and augments lower esophageal sphincter pressure in gastroesophageal reflux disease (Farrell et al., 1973). Nonetheless, in a pilot study, topical bethanechol did not significantly improve esophageal motility (O’Rourke et al., 2013). Edrophonium, a cholinesterase inhibitor used to diagnose myasthenia gravis, was used to provoke esophageal spasm in the investigation of non-cardiac chest pain, but the lack of correlation between symptoms and objective changes in esophageal manometry limited its utility (Botoman, 2002).

Gastric acid secretion is controlled by a mix of cholinergic muscarinic stimulation and hormonal signaling by gastrin and histamine; thus, only partial inhibition of acid release is achieved with anti-muscarinic agents. Consequently, histamine-2 receptor and H⁺-K⁺ATPase (proton pump) inhibitors are highly successful and MR antagonists rarely prescribed. Pirenzepine (Gastrozepin), an M₁R antagonist that is not FDA approved, has limited use to treat acid-related disorders in the EU (Tryba and Cook, 1997). Scopolamine, a non-selective MR antagonist, is commonly used as a transdermal patch for nausea associated with anesthesia or motion sickness (Riad and Hithe, 2021).

Dicyclomine (Bentyl), an M₁R- and M₃R-selective antagonist that inhibits small and large intestinal motility, is used as an anti-spasmodic agent to treat IBS (Giachetti et al., 1986; Doods et al., 1987). Neostigmine, a cholinesterase inhibitor, is used to treat acute intestinal pseudo-obstruction associated with critical illness or opioid use, another condition of impaired smooth muscle motility. Colonic decompression may be achieved with intravenous neostigmine (De Giorgio et al., 2001), although cardiac monitoring is important and rapid administration of atropine may be required for resulting bradycardia.

Although a potential role for modulating MR activity to treat cancer was demonstrated in a variety of cell types (Shah et al., 2009), except for an ongoing trial to investigate the utility of bethanechol before surgery for resectable PDAC (U.S. National Library of Medicine, 2021), the efficacy of modulators of MR activity in digestive tract cancers has not been tested in the clinic. Moreover, anti-tumor efficacy may be limited by the inability to achieve adequate concentrations in target tissues while, at the same time, preventing off-target adverse effects. An ideal agent would exhibit target organ and MR-subtype specificity, goals hampered by the extensive similarity between orthosteric and allosteric ligand binding sites among the five MR subtypes (Liu et al., 2018). Studies of naturally occurring ligands, such as muscarinic toxins in snake venom, have provided insight into how subtype-selective agents may be formulated (Maeda et al., 2020). Such agents with potential for oncotherapy continue to be developed. For example, the M₃R-specific antagonist darifenacin which is approved to treat bladder dysfunction (Yamada et al., 2006) reportedly inhibits tumor progression and invasiveness in human-derived cell lines, most recently in colorectal cancer cell lines (Hering et al., 2021). As darifenacin is in clinical use with a known safety profile, it is an attractive candidate for adjunctive therapy, especially for cancers already shown to overexpress M₃R, like colon cancer cells (Frucht et al., 1999; Cheng et al., 2014), PDAC (Zhang et al., 2016), and non-small cell lung cancer (Lin et al., 2014).

Some therapeutic approaches may circumvent the need for MR subtype and tissue specificity. For example, treating colorectal cancers with poorly absorbed oral agents or drugs with extensive first-pass metabolism may target GI mucosal lesions with limited systemic side effects. However, even within a limited area of distribution, MRs are not constrained to only one downstream signaling pathway; the same receptor may have contradictory effects on neighboring cell types. Even when occupying the same binding pocket, ligands can influence the activation of pathways on other cell membrane surfaces via signaling bias and functional selectivity (Randáková and Jakubík, 2021). A ligand may bind several MR subtypes, but only activate one or a few, thereby compensating for binding pocket homogeneity. Furthermore, through selective interactions with residues in the binding pocket of a single subtype, ligands can encourage activated receptor configurations that favor interaction with certain G proteins. As an example, the MR agonist cevimeline increased intracellular calcium levels in Chinese hamster ovary (CHO) cells transfected with rat M₁R but did not increase cAMP levels. In contrast, carbachol, a non-selective MR agonist, elevated both calcium and cAMP levels (Gurwitz et al., 1994). Even more intriguing, cevimeline did not activate signaling in M₃R-transfected cells, contrary to its clinical use in Sjogren’s syndrome which is thought to be mediated by M₃R activation. This complexity makes it difficult to predict the clinical effects of new MR agonists and antagonists but suggests highly selective agents can be developed.