Enteric Inhibitory Neurotransmission, Starting Down Under

Introduction and Historical Note

Gastrointestinal motility requires neural regulation to coordinate contractions of smooth muscle cells in the walls of the organs. Neurons providing motor regulation have cell bodies in the myenteric plexus and can provide both excitatory and inhibitory input to the gastrointestinal (GI) tract. These neurons are activated through sensory inputs and their input generates patterns of contractile activity, such as tonic contractions, peristalsis, and segmentation.

The notion of an inhibitory innervation of the gut arose when Geoff Burnstock’s group at the University of Melbourne measured inhibitory responses in the taenia coli in the early 1960s (Burnstock et al. 1963; Bennett et al. 1966). This type of neural regulation is not common to the autonomic nervous system and therefore must be thought of as a very special capability of the enteric nervous system and in regulation of gut motor patterns. Enteric inhibitory neurotransmission became known as non-cholinergic, non-adrenergic (or NANC) neurotransmission because of its resistance to antagonists of norepinephrine and acetylcholine. In 1970 Professor Burnstock published evidence that ATP (or a closely related nucleotide) fulfilled the criteria as an inhibitory neurotransmitter (Burnstock et al. 1970). This highly cited paper demonstrated release of p urines during transmural nerve stimulation and similarities between the responses to exogenous ATP and the enteric inhibitory neurotransmitter in several GI muscle preparations. Later apamin was found to block purinergic responses, providing evidence that small-conductance Ca-²⁺ activated K⁺ (SK) channels are responsible for inhibitory junction potentials (IJPs) (Banks et al. 1979). A great deal of study was applied to the pharmacology of nucleotide receptors, and the molecular age allowed identification of several families of genes encoding a large number of p urine and pyrimidine receptors. Like cholinergic receptors, one family of purine receptors mediated ionotropic effects (P2X) and another family mediated metabotropic effects (P2Y).

Enteric Inhibitory Motor Neurotransmitters

At the current time, the identity of ATP as the enteric inhibitory neurotransmitter has come into doubt, as the Eccles criteria for a neurotransmitter are better met by β-nicotinamide adenine dinucleotide (β-NAD) (Mutafova-Yambolieva et al. 2007; Hwang et al. 2011) or a metabolite ADP-ribose (ADPR) (Durnin et al. 2012). Stimulation of motor neurons in mouse or primate colon by agonists that activate ganglionic receptors causes release of ATP and β-NAD/ADPR, but only the release of β-NAD/ADPR was blocked by tetrodotoxin, suggesting that most of the ATP released comes from ganglionic sources, not from varicosities along projections of motor neurons that innervate the smooth muscle layers (Durnin et al. 2013).

There was a flurry of activity regarding the role of peptides as neurotransmitters in the GI tract during the 1980s. Several studies linked peptides, such as neurokinins and vasoactive intestinal polypeptide (VIP), to motor neurotransmission, and VIP (or related peptides) were initially considered the long sought-after non-cholinergic, non-adrenergic (NANC) inhibitory transmitter (Goyal et al. 1980). However, responses to nerve stimulation were considerably different than responses to VIP (Mackenzie and Burnstock 1980; Hills et al. 1983), and it is now difficult to picture how responses that are so clearly dependent upon NO, could have been reduced or blocked by systemic immunoneutralization of VIP. There were also studies using VIP fragments (e.g. VIP_10–28) to antagonize VIP receptors and peptidases (α-chymotrypsin) to break down peptides released into the interstitium and claims that antagonists of this sort could block NANC neurotransmission (Grider and Rivier 1990), but other investigators failed to show antagonism of NANC neurotransmission with VIP analogues or block of NANC responses by peptidases. Recent studies using VIP^−/− mice have clearly shown a VIP-dependent component of motor neurotransmission, but VIP-dependent responses are generally induced by sustained, multi-sec stimulation and/or at higher stimulus frequencies than 5 Hz (Keef et al. 2013).

A new, non-conventional neurotransmitter emerged on the scene as an important mediator of NANC inhibitory neurotransmission in 1989–1990 (Gillespie et al. 1989; Bult et al. 1990). The evidence for nitric oxide (NO) as a NANC neurotransmitter was rather convincing from the first papers on this topic: a component of inhibitory neurotransmission was blocked by inhibitors of NO synthesis and by hemoglobin, an isoform of nitric oxide synthase (nNOS) was localized in nerve varicosities (where it co-localized with VIP), stimulation of intrinsic neurons caused release of NO, and NO elicited hyperpolarization and relaxation of GI muscles (Sanders and Ward 1992). The concept of NO as an enteric inhibitory neurotransmitter was not without controversy, and proponents of VIP (and related peptides) continued to argue aggressively that the authentic neurotransmitter was VIP and NO was a secondary neuromodulator produced in post-junctional cells (Murthy et al. 1996) or as a secondary mediator released from neurons, but only after release of VIP (Mashimo and Goyal 1999). Eventually, the arguments against NO or involving the need for a peptide middleman to induce NO-dependent relaxation failed to explain the data, and studies with specific knockouts of NOS demonstrated that nNOS, expressed by motor neurons, was the source of this important component of enteric inhibitory responses (Dick et al. 2002).

Purines, Nitric Oxide and Peptides Elicit Different Parts of the Enteric Inhibitory Response

Enteric inhibitory responses are multi-facetted. Stimulation of motor neurons releases at least three types of neurotransmitter substances, and recordings of post-junctional changes in membrane potential and suppression of the spontaneous electrical activity of GI muscles, best, but not fully, enunciate these responses. The relative contribution of each class of enteric neurotransmitter to post-junctional responses, differs in various tissues of the GI tract and in different species. Responses are dependent upon stimulus frequency: single or a few pulses of transmural nerve stimulation elicit post-junctional responses that are mediated by purines and NO and are blocked by apamin, P2Y1 receptor antagonists and N ^G -nitro-L-Arginine; stimulus frequencies of 5 Hz or more are typically required to elicit responses due to release of peptides. The kinetics of purinergic and nitrergic responses are markedly different. The purine response is a fast, large amplitude hyperpolarization response that can approach the K⁺ equilibrium potential, but the response wanes within several hundred milliseconds. The nitrergic response develops more slowly, and responses to even a single stimulus can persist for a few seconds. The nitrergic hyperpolarization is smaller in amplitude than the purinergic response and may be due either to activation of K⁺ channels (Koh et al. 2001) or suppression of an inward current, most likely suppression of a non-selective cation conductance or Ca²⁺-activated Cl⁻ channels (Zhang and Paterson 2002; Hirst et al. 2004). The integrated response to purinergic and nitrergic neurotransmission is a two phase inhibitory junction potential, consisting of a fast IJP and a slow IJP. VIP and PACAP elicit responses that are much slower to develop and are partially mediated by activation of K_ATP channels (Keef et al. 2013).

Responses to specific components of the IJP can be dissected pharmacologically because relatively selective reagents are available to block the nitrergic and purinergic components (via block of NO synthesis, NO receptors and purine receptors). Responses due to release of peptides are more difficult to delineate pharmacologically, but as above, VIP^−/− mice are being used to clarify the nature of peptidergic responses (Keef et al. 2013). Present data suggest that the mouse is a relatively good experimental model for studies of neural regulation in the GI tract, however it is important to continue to check for species differences.

Obviously, motility is dependent upon contractile activity, so it is important to comment on the differences in contractile responses elicited by different classes of neurotransmitters. Purinergic motor responses appear to be mediated by an electrical mechanism (i.e. hyperpolarization and consequent reduction in smooth muscle excitability), but both nitrergic and petidergic responses may have, in addition to electrical mechanisms, important non-electrical mechanisms of action (e.g. Ca²⁺-desensitization of the smooth muscle contractile apparatus). Ca²⁺ sensitization mechanisms in GI muscles have been studied largely by addition of excitatory or inhibitory neurotransmitter substances to organ baths and measuring the extent of protein phosphorylation in Ca²⁺ sensitization pathways or by testing the effects of Rho kinase antagonists. However, study of responses to exogenous compounds is not an accurate means of investigating the role of Ca²⁺ desensitization mechanisms in neurotransmitter responses, because neurotransmitters may not bind the same population of receptors as exogenous transmitter substances added to the organ bath (Bhetwal et al. 2013).

Post-junctional Cells that Mediate Enteric Neurotransmission

The classic concept of motor neurotransmission in the gut is that neurotransmitters released into the interstitium diffuse to receptors expressed by smooth muscle cells to elicit responses. In fact the receptive fi eld for motor neurotransmitters is more complicated and composed of multiple types of cells that express receptors and transduction mechanisms for enteric motor neurotransmitters. At least two classes of interstitial cells (interstitial cells of Cajal (ICC) and PDGFRα⁺ cells) have been shown to contribute to neural responses: ICC mediate at least a portion of the responses to NO and acetylcholine (ACh) and PDGFRα⁺ cells mediate purinergic neurotransmission (Burns et al. 1996; Ward et al. 2000; Kurahashi et al. 2011). Interstitial cells convey electrical conductance changes to the smooth muscle cells via gap junctions. Electrical connectivity between these cells means that conductance changes in any of the coupled cells affects the excitability of the smooth muscle component. Together smooth muscle cells, ICC and PDGFRα⁺ cells have been referred to as the SIP syncytium (Sanders et al. 2012), and this integrated syncytium of cells constitutes the receptive fi eld for enteric neurotransmission.

Receptors for Enteric Inhibitory Neurotransmitters

We now know that enteric inhibitory neurotransmission is mediated by P2Y1 receptors (Gallego et al. 2006), and the fast IJPs associated with purinergic neuro-transmission are absent in P2Y1^−/− mice (Gallego et al. 2012; Hwang et al. 2012). PDGFRα⁺ cells express P2Y1 to a far greater extent than other cells of the SIP syncytium (Peri et al. 2013). PDGFRα⁺ cells respond to P2Y1 agonists with activation of SK currents, possibly due to openings of SK3 channels that are also highly expressed in these cells (Klemm and Lang 2002; Iino and Nojyo 2009; Kurahashi et al. 2011; Peri et al. 2013). Ca²⁺ transients are elicited in PDGFRα⁺ cells in response to purines, and this signaling is likely to couple P2Y1 receptors to activation of SK3 channels (Baker et al. 2013). Current data suggests that smooth muscle cells are unlikely to mediate purinergic hyperpolarization responses, because at physiological transmembrane potentials, smooth muscle cells generate small amplitude inward currents and depolarization in response to purines (Hwang et al. 2011; Kurahashi et al. 2011).

The receptor for NO is soluble guanylate cyclase (sGC), composed of α and β subunits (Groneberg et al. 2011). Several studies have shown that ODQ, an inhibitor of sGC, blocks the nitrergic component of enteric inhibitory neurotransmission (Franck et al. 1997; Hirst et al. 2004). sGC has been shown to be expressed robustly in ICC of the GI tract, and sGC expression has also been resolved in some types of PDGFRα⁺ cells (Iino et al. 2009). Expression of sGC has not been resolved in SMCs by immunohistochemistry, suggesting that levels of sGC are comparatively low in these cells. Recent studies have employed cell-specific iCre expressing mice to knock down expression of sGC in ICC, SMC or in both cell types (Groneberg et al. 2013). These studies showed that knock-down of sGC in ICC or smooth muscle cells did not block nitrergic response, but when knock-down occurred in both types of cells, nitrergic responses were depressed. Thus, nitrergic responses may require activation of signaling molecules and effectors in multiple types of cells in the SIP syncytium.

Cellular Effectors that Mediate Enteric Inhibitory Responses

Many studies have identified apamin-sensitive K⁺ channels as the main mediators of purinergic inhibitory input to GI muscles (Banks et al. 1979). As above, SK3 channels are highly expressed in PDGFRα⁺ cells (Klemm and Lang 2002; Iino and Nojyo 2009; Kurahashi et al. 2011; Peri et al. 2013), however SMC also express SK channels and can respond to purines by activation of SK currents (Koh et al. 1997; Vogalis and Goyal 1997; Ro et al. 2001; Klemm and Lang 2002). It should be noted that the current density due to SK channels is far lower in SMC (Kurahashi et al. 2011), and as mentioned above, the net response of SMC at physiological membrane potentials is generation of inward, not outward currents.

There is controversy about the mechanism of NO action GI muscles. Initially, the hyperpolarization response to nitrergic stimulation suggested activation of a K⁺ conductance (Dalziel et al. 1991; Thornbury et al. 1991), however others have suggested that suppression of an inward current may be the mechanism for hyperpolarization responses (Zhang and Paterson 2002; Hirst et al. 2004). The nature of the conductance responsible for the inward current may vary by anatomical location and in different species. For example, responses in the guinea pig colon were unaffected by 9-AC, a chloride channel blocking drug, but in lower esophageal sphincter, responses were inhibited by another chloride channel blocking drug, niflumic acid. Contractile responses to nitrergic stimulation may also be influenced by Ca²⁺ uptake into stores and by Ca²⁺ desensitization mechanisms. To affect contractility, these responses would likely occur in SMC, however if nitrergic responses are linked to Ca²⁺-activated Cl⁻ channels (CaCC), then responses involving Ca²⁺ sequestration and release from intracellular stores would be mainly focused upon ICC, because it is ICC that express CaCC in the SIP syncytium (Hwang et al. 2009).

Summary and Major Questions Remaining for Enteric Inhibitory Regulation of the GI Tract

There are still significant questions to be answered regarding enteric inhibitory neurotransmission. Progress has been made on the relative contributions of specific cells and pathways that are activated by enteric neurotransmitters from studies of clearly identified isolated cells of the SIP syncytium. Delineation of the cells, receptors and ion channels mediating enteric inhibitory responses will likely result from genetic studies in which genes can be knocked-down selectively in specific classes of cells. It will be important to assess the degree of knock-down in specific cellular populations to verify cell-specific loss-of-function, and as seen with the sub-resolution levels of sGC in SMC, protein levels undetected by immunohistochemical techniques may be effective in mediating neurotransmission. Which SIP syncytium cells mediate responses to neuropeptides; additional studies on isolated cells are needed to investigate this question. Study of how SIP cells develop and why they are lost or become dysfunctional in disease conditions might explain defective neural regulation in a variety of GI motor disorders. Of particular interest is how inflammatory factors affect the responsiveness of the SIP syncytium to neurotransmission. It is possible, for example, that inflammatory mediators might remodel the responses of the SIP syncytium, such that aberrant motor responses might be elicited by normal neural inputs. Very little is known about prejunctional regulation of neurotransmitter release from enteric motor neurons. Knowledge of prejunctional regulatory pathways might provide ideas about enhancing or depressing the level of regulation by enteric motor neurons and could develop as useful therapeutic tools. More studies to characterize enteric neural responses in human muscles are needed to incorporate and validate what is known from animal studies into our knowledge of the physiology and pathophysiology of human neurogastroenterology.