Volume 98, Issue 1 p. 25-37
Free Access

The dorsal motor nucleus of the vagus and regulation of pancreatic secretory function

Bashair M. Mussa

Bashair M. Mussa

University of Melbourne, Department of Medicine, Clinical Pharmacology and Therapeutics Unit, Austin Health, Heidelberg, Victoria 3084, Australia

Search for more papers by this author
Anthony J. M. Verberne

Anthony J. M. Verberne

University of Melbourne, Department of Medicine, Clinical Pharmacology and Therapeutics Unit, Austin Health, Heidelberg, Victoria 3084, Australia

Search for more papers by this author
First published: 01 June 2012
Citations: 55
A. J. M. Verberne: University of Melbourne, Clinical Pharmacology & Therapeutics Unit, Austin Health, Heidelberg, Victoria 3084 Australia. Email: [email protected]

New Findings

  • What is the Topic of this review?

    This review describes new evidence for the control of pancreatic exocrine and endocrine secretion by the dorsal motor nucleus of the vagus. It describes the effects of several gastrointestinal peptides on pancreatic secretion and considers their sites of action in the vago-vagal reflex pathway.

  • What advances does it highlight?

    This review draws together recent electrophysiological, anatomical and pharmacological studies to provide a state-of-art view of the parasympathetic control of the exocrine and endocrine pancreas.

Recent investigation of the factors and pathways that are involved in regulation of pancreatic secretory function (PSF) has led to development of a pancreatic vagovagal reflex model. This model consists of three elements, including pancreatic vagal afferents, the dorsal motor nucleus of the vagus (DMV) and pancreatic vagal efferents. The DMV has been recognized as a major component of this model and so this review focuses on the role of this nucleus in regulation of PSF. Classically, the control of the PSF has been viewed as being dependent on gastrointestinal hormones and vagovagal reflex pathways. However, recent studies have suggested that these two mechanisms act synergistically to mediate pancreatic secretion. The DMV is the major source of vagal motor output to the pancreas, and this output is modulated by various neurotransmitters and synaptic inputs from other central autonomic regulatory circuits, including the nucleus of the solitary tract. Endogenously occurring excitatory (glutamate) and inhibitory amino acids (GABA) have a marked influence on DMV vagal output to the pancreas. In addition, a variety of neurotransmitters and receptors for gastrointestinal peptides and hormones have been localized in the DMV, emphasizing the direct and indirect involvement of this nucleus in control of PSF.

The dorsal motor nucleus of the vagus (DMV) is considered to be the main, if not the sole, source of the vagal innervation of various organs within the gastrointestinal (GI) tract, including the stomach and the pancreas (Kalia, 1981; Loewy & Spyer, 1990; Hornby & Wade, 2011). The secretomotor function of the DMV has been recognized for more than 30 years, and so its role in regulation of GI secretion, motility and pancreatic secretory function (PSF) has been well documented (Kerr & Preshaw, 1969; Loewy & Spyer, 1990; Hornby & Wade, 2011). Recent developments in neural–pancreatic research have revealed that, in addition to the pancreatic hormones and peptides, the DMV is a major brain region involved in mediating pancreatic secretion (PS). It is evident that the DMV is the site of origin of vagal efferent neurons that innervate both the endocrine and the exocrine pancreas (Rinaman & Miselis, 1987; Berthoud & Powley, 1991; Jansen et al. 1997).

Several neurophysiological studies have shown that electrical and chemical stimulation of the DMV activates both endocrine and exocrine secretion via a cholinergic pathway. An early study by Ionescu and colleagues showed that unilateral electrical stimulation of the DMV produced a 100–200% increase in plasma insulin levels. These increases were completely blocked by atropine methonitrate and vagotomy, suggesting that the influence of the DMV on the endocrine pancreas is vagal in nature (Ionescu et al. 1983). In addition, another study demonstrated that bilateral electrical stimulation of the medial and the lateral DMV elicited significant increases in gastric acid secretion and in insulin and glucagon secretion, respectively (Laughton & Powley, 1987). Recently, we have found that bilateral chemical activation of the DMV using bicuculline methiodide (GABAA receptor antagonist) increased the exocrine and endocrine PS significantly. Treatment with atropine methonitrate (a muscarinic receptor antagonist that does not cross the blood–brain barrier) completely abolished these excitatory effects, emphasizing the involvement of a cholinergic mechanism (Mussa & Verberne, 2008; Mussa et al. 2011). The present review aims to summarize the key findings from the recent literature on the physiological roles of the DMV in the control of PS.

Functional anatomy of the DMV

The DMV was first described by Stilling in 1843, and since then, researchers have investigated different anatomical, histological, physiological and functional aspects of this structure. The DMV has been acknowledged as the largest source of parasympathetic preganglionic neurons within the lower brainstem (Stilling & Wallach, 1843; Huang et al. 1993). It is located in the dorsomedial caudal medulla oblongata, close to the floor of the fourth ventricle (Fig. 1). In the horizontal plane, the DMV replicates the ‘Y’ appearance of the nucleus of the solitary tract (NTS) and lies medial and ventral to this nucleus throughout the rostrocaudal extent of the medulla oblongata. More specifically, in the caudal part of the medulla oblongata, the DMV fuses close to the mid-line dorsolateral to the central canal. In the rostral medulla oblongata, both sides of the DMV lie ventromedial to the NTS as the fourth ventricle opens. Similar to the NTS, the DMV is a paired bilateral and symmetrical nucleus, which sends projections to the viscera of the abdomen (stomach, liver, pancreas, small intestine and proximal large intestine) in a bilateral fashion (Bystrzycka & Nail, 1985; Loewy & Spyer, 1990; Powley et al. 1992; Blessing, 1997; Jordan, 2011). Electrophysiological and histological studies have shown that neurons in the DMV are heterogeneous with respect to size, physiological type and morphology. However, two major types have been classified, namely multipolar neurons in the rostral and caudal DMV and bipolar neurons in the intermediate DMV with transverse diameter ranges between 12 and 30 μm (Bystrzycka & Nail, 1985; Nosaka, 1986). The afferent inputs and efferent outputs of the DMV will be discussed in more detail in the following two subsections.

Details are in the caption following the image

Schematic diagram depicting the relationship between the area postrema (AP), the dorsal motor nucleus of the vagus (DMV), the nucleus of the solitary tract (NTS) in the dorsal medulla oblongata and the branches of the thoracic and subdiaphragmatic abdominal vagal nerve trunks

Afferent projections to the DMV The DMV receives myelinated and unmyelinated vagal afferent input from the periphery, as well as afferent input from CNS structures. In vitro studies have demonstrated the presence of visceral vagal afferent projections within the DMV. However, these projections seem to be thinly distributed along the dorsal and the lateral edges of the nucleus (Shapiro & Miselis, 1985; Champagnat et al. 1986; Neuhuber & Sandoz, 1986). A review by Loewy and Spyer across various species has revealed that the DMV also receives a range of projections from other areas within the CNS, including the insular cortex, the central nucleus of the amygdala, the paraventricular hypothalamic nucleus, the lateral hypothalamic area, the dorsomedial hypothalamic nucleus, the posterior hypothalamus, the mesencephalic central grey matter, the parabrachial nucleus, the A5 catecholamine cell group, the NTS, the medullary reticular formation and the raphé obscurus nucleus (ter Horst et al. 1984; Loewy & Spyer, 1990). Injection of the retrograde neuronal tracer horseradish peroxidase into the DMV revealed a wide distribution of labelled neurons within the lower brainstem (ter Horst et al. 1984; Loewy & Spyer, 1990).

Efferent projections of the DMV Eighty per cent of the neurons in the DMV give rise to the parasympathetic preganglionic fibres that innervate GI organs, including the pancreas. Although the majority of these neurons are cholinergic and non-cholinergic, centrally projecting neurons were also identified within the DMV. It has been found that the centrally projecting DMV neurons project to the parabrachial nucleus, the cerebellar cortex and the cerebellar nuclei (Loewy & Spyer, 1990). The efferent component of the DMV that innervates the abdominal viscera is organized into a series of longitudinally arrayed columnar subnuclei that correspond to specific branches of the abdominal vagus (Fox & Powley, 1985; Norgren & Smith, 1988; Powley et al. 1992). Four of these columnar subnuclei form bilateral and symmetrical pairs of preganglionic neurons projecting from both sides of the medulla oblongata, whereas preganglionic neurons within the fifth column are distributed through the longitudinal extent of the left DMV. The medial columnar subnucleus, which occupies the medial two-thirds of the DMV on each side of the medulla oblongata, contains preganglionic neurons corresponding to the gastric vagal branches. The lateral pole of each side of the DMV contains preganglionic neurons that give rise to the two coeliac branches (left DMV, accessory coeliac branch; right DMV, coeliac branch). In contrast, the fifth columnar subnucleus contains preganglionic neurons corresponding to the hepatic branch. Compared with the gastric branch, this branch contains 10% fewer efferent fibres (Fox & Powley, 1985; Norgren & Smith, 1988; Powley et al. 1992).

The DMV and GI hormones

It is interesting to note that most GI peptides and hormones that are involved in regulation of PS are secreted locally in the GI tract and have receptors within the DMV, NTS or area postrema (AP; Fig. 2; Travagli & Browning, 2011). This supports the notion that DMV has both a direct and an indirect influence on PSF.

Details are in the caption following the image

Schematic diagram showing the release sites of several gastrointestinal hormones and the location of their receptors within the dorsal vagal complex
Vagal afferents project to neurons in the NTS which, in turn, influence vagal motor neurons of the DMV. Preganglionic vagal motor neurons are excitatory (cholinergic), while postganglionic neurons are either inhibitory (nitrergic?) or excitatory. Abbreviations: AP, area postrema; DMV, dorsal motor nucleus of the vagus; GI, gastrointestinal; GLP-1, glucagon-like peptide-1; NO, nitric oxide; NTS, nucleus of the solitary tract; PP, pancreatic polypeptide; and TRH, thyrotrophin-releasing hormone.

Pancreatic polypeptide (PP) Pancreatic polypeptide is secreted from the endocrine pancreas postprandially to inhibit PS in vivo in a variety of species (Lin et al. 1977; Taylor et al. 1978, 1979; Adrian et al. 1979; Schwartz, 1983). It has been demonstrated that PP inhibits both endocrine and exocrine PS (Lee et al. 1988). Interestingly, these effects of PP on PS seem to be dependent on a cholinergic mechanism (Jung et al. 1987). However, in vitro studies on isolated pancreatic acini, which were prepared by enzymatic digestion of rat pancreas, failed to reproduce the inhibitory effect of PP on PS, suggesting that PP acts indirectly and probably centrally to mediate this effect (Louie et al. 1985; Jung et al. 1987). In addition, several studies have shown that GI tract tissues, including pancreatic acini and ductal cells, lack PP receptors, with the exception of the basolateral membranes of the canine small intestine (Louie et al. 1985; Gilbert et al. 1988; Sheikh et al. 1991). Using in vitro receptor autoradiography, Whitcomb et al. (1990) have localized high-affinity PP receptors in high concentration within the DMV. In addition, it was found that microinjection of PP into the DMV produced significant inhibition of PS, and this effect was site specific, because microinjection of PP outside the DMV did not affect PS (Okumura et al. 1995a). Furthermore, an electrophysiological study in rats has shown that PP influences vagal outflow to the pancreas directly via modulation of DMV pancreas-projecting neurons or indirectly via modulation of a synaptic connection between these neurons and the NTS and the AP (Browning et al. 2005b). It has also been found that lesions of the AP, a structure which lies outside the blood–brain barrier, alters the effects of peripheral PP on PSF (Deng et al. 2001). Together, these findings led to the suggestion that PP, which is released from the pancreas into the circulation after meal ingestion, gains access to the DMV via the AP. There, PP suppresses vagal excitatory inputs to the pancreas via direct modulation of the DMV pancreatic vagal outflow (for review see Travagli et al. 2006; Fig. 2).

Similar to PP, neuropeptide Y (NPY) and peptide YY (PYY) exert their inhibitory effects on PS via activation of Y1 and Y2 receptors within the DMV, producing inhibition of pancreatic cholinergic vagal outflow (Leslie et al. 1988; Lynch et al. 1989; Dumont et al. 1990; Masuda et al. 1994; Naruse et al. 2002; Browning & Travagli, 2003).

Ghrelin Ghrelin is known to be involved in stimulation of PS. This peptide acts as an endogenous ligand for the growth hormone secretagogue receptor and is produced by endocrine cells in the oxyntic glands of the stomach and intestine, as well as by neurons of the hypothalamic arcuate nucleus (Kojima et al. 1999; Date et al. 2000; Cummings et al. 2001). Intravenous injection of ghrelin increases PS in a dose-dependent manner, and this effect was abolished in the presence of atropine and hexamethonium, suggesting an involvement of a cholinergic mechanism (Li et al. 2006). It has been shown that PS was significantly increased in response to intracerebroventricular infusion of ghrelin. Vagotomy completely blocked this effect, demonstrating that ghrelin acts centrally via vagal pathways to stimulate PS (Li et al. 2006). In addition, experiments in conscious and anaesthetized rats have shown that both intracerebroventricular and intravenous administration of ghrelin induced c-fos expression in the DMV, the NTS and the AP (Date et al. 2001; Li et al. 2006). Interestingly, c-fos expression in the DMV and the NTS after intravenous infusion of ghrelin was significantly reduced after lesions of the AP (Li et al. 2006). Moreover, a study by Zigman et al. (2006) has demonstrated that the highest expression of ghrelin receptor mRNA was observed in the AP, supporting the hypothesis that ghrelin gains access to the brain primarily via the AP and influences the neuronal activity of the NTS and the DMV. On the contrary, ghrelin receptor immunoreactivity was found in the NTS and DMV but not the AP (Lin et al. 2004). Indeed, ghrelin has also been shown to inhibit the activation of NTS catecholamine neurons by visceral inputs, although this response is probably associated with the effects of ghrelin on feeding (Cui et al. 2011). These findings suggest that ghrelin stimulates PS via activation of a vagovagal reflex which, in turn, increases the vagal outflow from the DMV to the pancreas.

Glucagon-like peptide-1 (GLP-1) Glucagon-like peptide-1, 7–36/7–37 amide is a signal peptide which is involved in stimulation of insulin release via an interaction with specific receptors. Glucagon-like peptide-1 is an incretin hormone and is released from the enteroendocrine L-cells of the intestinal mucosa into the circulation postprandially in response to the presence of luminal nutrients, including fat and carbohydrates (Fehmann et al. 1992; Schirra & Göke, 2005). Glucagon-like peptide-1 stimulates insulin secretion in a glucose-dependent manner via activation of its receptors expressed on islet β-cells. Glucagon-like peptide-1 is considered to be a unique pancreatic secretagogue because it also replenishes insulin stores via activation of proinsulin gene transcription. It also suppresses glucagon secretion from islet α-cells (Drucker et al. 1987; Fehmann et al. 1992, 1995; Drucker, 1998). Several studies have used a variety of techniques, including immunohistochemistry, receptor binding, radioimmunoassay and chromatography, to demonstrate the presence of GLP-1 in the CNS, and the highest density of GLP-1 binding sites and mRNA was detected in the NTS and DMV (Shimizu et al. 1987; Jin et al. 1988; Kanse et al. 1988; Kreymann et al. 1989; Uttenthal et al. 1992; Göke et al. 1995; Larsen et al. 1997; Fig. 2). This suggests that besides the direct action of GLP-1 on the endocrine pancreas, GLP-1 may also act indirectly via central vagal pathways to stimulate insulin release. A study by Kastin et al. (2002) has supported this suggestion by demonstrating that GLP-1 gains access to the brain by simple passive diffusion. Furthermore, an electrophysiological study has shown that superfusion of GLP-1 onto DMV neurons excites a subpopulation of pancreas-projecting preganglionic vagal motorneurons (Wan et al. 2007a). The selectivity of this effect was confirmed using a specific GLP-1 receptor agonist (exendin-4) and antagonist [exendin-(9–39)], which mimicked and prevented this excitatory effect, respectively. This finding emphasized the hypothesis that GLP-1 also influences endocrine secretion via a direct activation of its receptors on the DMV or on the NTS, which modulates neural input to the DMV. The latter, in turn, excites the efferent output to the endocrine pancreas (Wan et al. 2007a).

Thyrotrophin-releasing hormone (TRH) Thyrotrophin-releasing hormone is another hormone that stimulates PS via an action in the DMV. Thyrotrophin-releasing hormone was originally described as a hypothalamic hormone that is synthesized and released from hypothalamic neurons to stimulate the release of thyrotrophin secretion from the anterior pituitary gland (Taylor et al. 1990; Rang et al. 1999). Thyrotrophin-releasing hormone was also detected in different GI organs, including the pancreas (Morley et al. 1977; Dutour et al. 1987). Thyrotrophin-releasing hormone is co-localized with insulin within the islet cells and is secreted in response to a high level of glucose, whereas TRH release is inhibited by insulin in isolated pancreatic islets (Leduque et al. 1987; Benicky & Strbak, 2000). It also modulates the basal secretion of glucagon in vitro, and this response was inhibited by immunoneutralization of TRH with specific antibodies (Ebiou et al. 1992). Several studies have demonstrated the presence of TRH receptors and immunoreactive fibres in various medullary areas, including the DMV and the NTS (Manaker & Rizio, 1989; Fig. 2). This suggests that TRH may act centrally to regulate pancreatic function (Kubek et al. 1983; Palkovits et al. 1986; Manaker & Rizio, 1989; Rinaman et al. 1989; Rinaman & Miselis, 1990). This suggestion is supported by the finding that application of TRH or a TRH analogue into the cerebrospinal fluid stimulates PS in rats (Kato & Kanno, 1983). More specifically, other studies have shown that microinjection of a TRH analogue into the dorsal vagal complex (DVC), a term used to refer to the two closely interacting groups of neurons in the DMV and the NTS, stimulates PS in a dose-dependent manner. Microinjection of the same analogue into the brainstem but outside the DVC did not produce any changes in PS (Okumura et al. 1995b). Vagotomy and atropine completely blocked this excitatory effect, confirming the involvement of vagal cholinergic pathways (Okumura et al. 1995b). Likewise, a study by Yoneda et al. (2005) has shown that pancreatic blood flow (PBF) was increased in a dose-dependent manner in response to microinjection of a TRH analogue into the DVC. In contrast, microinjection of vehicle into the DVC or microinjection of the TRH analogue outside the DVC produced no change in the PBF. Cervical vagotomy and atropine administration blocked the excitatory effects of TRH microinjection into the DVC, suggesting that the peptide increased PBF via vagal cholinergic pathways. Moreover, involvement of nitric oxide pathways was also suggested, because intravenous administration of l-NAME (a nitric oxide synthase inhibitor) inhibited TRH-induced stimulation of PBF, and this inhibition was abolished in the presence of l-arginine (Yoneda et al. 2005).

Nitric oxide Recent interest in NO research has led to the establishment of a very clear relationship between this molecule and pancreatic endocrine and exocrine secretion. It is well documented that NO acts as a widespread neurotransmitter or neuromodulator in the brain (Knowles et al. 1989; Bredt et al. 1990; Moncada et al. 1991; Bredt & Snyder, 1992). Nitric oxide synthase (NOS) activity has been detected in various areas within the medulla oblongata, including the NTS (Forstermann et al. 1990; Vincent & Kimura, 1992). A study by Lu et al. has shown that NADPH-diaphorase, which can be used to identify NOS in the CNS, was detected in different regions of the NTS where the majority of the visceral vagal afferents terminate. The NADPH-diaphorase activity was eliminated upon removal of the nodose ganglion, which contains the cell bodies of visceral vagal afferents, confirming the role of NO as the neuromodulator in these afferents (Lu et al. 1994; Lin et al. 1998). Moreover, evidence obtained in vivo and in vitro has demonstrated excitatory effects of NO on NTS neurons, because intravenous and direct administration of l-NAME significantly reduces the discharge of NTS neurons (Ma et al. 1995). Interestingly, several immunohistochemical studies in rats and cats have shown that NOS-containing neurons are localized in the DMV (Mizukawa et al. 1989; Krowicki et al. 1997). In fact, NADPH-diaphorase activity was detected in two populations of preganglionic neurons in the DMV, suggesting that NO may play a role in the control of some GI functions (Travagli & Gillis, 1994; Krowicki et al. 1997). In addition, it has been found that the firing rate of DMV neurons increased in response to the presence of NO or its donors. Several lines of evidence support this finding by showing that NMDA-induced increases in DMV neuronal firing rate were blocked by NO inhibitors and scavengers (Travagli & Gillis, 1994). Taking into account all these findings, along with the fact that NO within the DMV is involved in regulation of gastric secretion (Beltrán et al. 1999), we have proposed that DMV NO may also be involved in mediation of PS. This hypothesis was supported by our recent findings that the excitatory effects of chemical stimulation of the DMV on insulin secretion were significantly enhanced after blockade of peripheral NOS, suggesting that NO exerts an inhibitory influence on insulin secretion (Mussa et al. 2011).

The DMV and vagovagal reflexes

Vagovagal reflexes constitute major extrinsic neural pathways that are involved in control of many GI and pancreatic functions (for review see Chang et al. 2003; Travagli & Browning, 2011). The vagovagal reflex pathway consists of first-order vagal sensory afferents whose cell bodies are located within the nodose ganglion, neurons of the NTS and efferent vagal motor neurons of the DMV. Although an extensive amount of work has been done to identify the components of the vagovagal reflexes that regulate gastric motility and secretion, less is known about the vagal pathways that are involved in control of PSF. Our hypothesis was that GI hormones, such as cholecystokinin (CCK) and 5-HT, activate receptors on pancreatic vagal afferents that terminate onto and activate neurons within the NTS via glutamate receptors (Liao et al. 2005). These neurons, in turn, influence pancreatic preganglionic neurons (PPNs) projecting from the DMV and modulate the vagal output to the pancreas via glutamatergic and GABAergic inputs (Travagli et al. 1991, 2006; Buijs et al. 2001; Love et al. 2007; Mussa & Verberne, 2008; Fig. 3). We have previously shown that the vagal output to the pancreas is mainly under GABAergic control (Mussa & Verberne, 2008). Although a variety of neurotransmitters have been localized in the DMV, it is evident that endogenous glutamate and GABA have a major impact on DMV vagal output (Travagli et al. 1991). Excitatory and inhibitory postsynaptic currents recorded from DMV neurons that were elicited by electrical stimulation of the NTS were mediated by glutamate and GABA, respectively (Love et al. 2007). Furthermore, a recent study by Babic et al. (2011) has demonstrated that a majority of DMV neurons are modulated by tonic glutamatergic and GABAergic inputs.

Details are in the caption following the image

Schematic diagram showing the basic components of the pancreatic vagovagal reflex
Vagal afferents arising from the pancreas, as well as from the stomach and other parts of the gastrointestinal tract, have their somata in the nodose ganglion and project into the NTS. The NTS neurons, in turn, project to the preganglionic vagal motor neurons of the DMV. Abbreviations: DMV, dorsal motor nucleus of the vagus; and NTS, nucleus of the solitary tract.

Several lines of evidence strongly support the existence of a pancreatic vagal sensory innervation. Using retrograde tracing techniques, Sharkey & Williams (1983) described the morphology of pancreatic vagal afferents. Injection of the fluorescent retrograde tracer True Blue into multiple parts of the pancreas, including the duodenal and splenic lobes, resulted in extensive labelling of cells in the nodose ganglion. The number of labelled cells was significantly reduced by treatment with capsaicin, which is known to destroy unmyelinated and small-diameter myelinated afferent fibres selectively. This suggests that the majority of pancreatic vagal afferents are also unmyelinated or small-diameter myelinated fibres (Sharkey & Williams, 1983). Similar findings were reported in a study in which the pancreatic vagal innervation was labelled using horseradish peroxidase as a retrograde axonal tracer (Carobi, 1987). Following injection of horseradish peroxidase into the pancreas via the common bile duct, many labelled neurons were detected in the right and left nodose ganglion. Capsaicin treatment led to a complete loss of the small labelled neurons of the nodose ganglion and a reduction in the number of the labelled neurons with a larger diameter, indicating that most of the pancreatic vagal afferents are capsaicin-sensitive fibres with small-diameter unmyelinated axons (Carobi, 1987).

In view of these findings, we investigated the effects of systemic administration of CCK and the 5-HT3 receptor agonist phenylbiguanide (PBG) on pancreatic vagal afferent discharge (PVAD). Furthermore, we wished to identify the receptors that were involved in these effects. In good agreement with previous reports, the results of this study have demonstrated that pancreatic vagal afferents are also highly sensitive to CCK and 5-HT (Mussa et al. 2008). Almost 30 years ago, Niijima studied the effects of glucose, 2-deoxyglucose, CCK and 5-HT on pancreatic vagal afferents (Niijima, 1981). However, compared with Niijima's study, we found that CCK produced more rapid and larger increases in PVAD. This may have been due to several differences between the two studies including the following: (i) CCK dose range; (ii) route of CCK administration; and (iii) the branch of the pancreatic vagus nerve that was used to record the afferent discharge. Intravenous administration of CCK and PBG doses in random order increased PVAD significantly in a dose-related manner. The responses to these two substances did not change after cervical vagotomy, indicating that the pancreatic vagal nerve discharge was afferent in nature (Mussa et al. 2008). Moreover, blockade of CCK1 and 5-HT3 receptors using specific antagonists produced significant reductions in the PVAD response to CCK and PBG, respectively. This finding suggests that CCK1 and 5-HT3 receptors mediate CCK- and PBG-induced activation of pancreatic vagal afferents. Pancreatic vagal efferents also play a key role in regulation of PSF, and the following section reviews this issue in more detail.

The DMV and pancreatic secretagogues: a focus on CCK and 5-HT

Cell bodies of preganglionic parasympathetic neurons that innervate the pancreas arise mainly from the DMV (Kalia, 1981; Loewy & Spyer, 1990). Several GI hormones are known to be involved in regulation of PSF, including the inhibitory peptide hormones PP, NPY, PYY and somatostatin and the excitatory hormones secretin, CCK, 5-HT, gastrin, secretin-releasing peptide and CCK-releasing peptide, ghrelin, GLP-1 and TRH. However, CCK and 5-HT are the principal mediators of PS in various species (Chey & Chang, 2001; Konturek et al. 2003).

Early studies of the effects of CCK in various parts of the brain have shown that ionophoretic application of CCK excites neurons within the substantia nigra and the dorsal horn and inhibits neurons within the NTS (Jeftinija et al. 1981; Morin et al. 1983; Hommer et al. 1985). In addition, immunohistochemical and tracing studies have demonstrated the presence of CCK-responsive neurons in a variety of regions within the CNS (Fallon & Seroogy, 1985; Herbert & Saper, 1990). Furthermore, several reports have demonstrated that peripheral administration of CCK induces Fos expression in restricted regions of the DMV (Rinaman et al. 1993, 1994). It has been shown previously that local application of CCK-8 to identified DMV neurons in vitro has led to depolarization and hyperpolarization of 43 and 11% of these neurons, respectively.

Furthermore, a non-active form of CCK-8, non-sulphated CCK-8, had no effect on the electrophysiological properties of these neurons, suggesting the presence of CCK-8 receptors within the DMV and the selectivity of the DMV neuronal responses to CCK-8 (Plata-Salamán et al. 1988). A recent whole-cell patch-clamp electrophysiological study conducted using brainstem slices has examined the effect of CCK on pancreas-projecting DMV neurons identified by retrograde tracing from the pancreas (Wan et al. 2007b). In that study, perfusion of the pancreas-projecting DMV neurons with CCK-8 activated about 60% of the neurons tested. This response was mediated by CCK1 receptors, because pretreatment with lorglumide, a CCK1 receptor antagonist, significantly attenuated the excitatory effect of CCK-8. Agonists of CCK2 receptors, even at a high concentration, did not affect the neural discharge of pancreas-projecting DMV neurons. In addition to the direct action of CCK-8 on pancreas-projecting DMV neurons, indirect actions of CCK-8 on the synaptic inputs to the DMV have also been proposed. This was supported by the finding that tetrodotoxin was able to reduce the excitatory effects of CCK-8 on pancreas-projecting DMV neurons significantly by blockade of action potential-dependent synaptic inputs to the DMV (Wan et al. 2007b).

Although it is well documented that CCK-8 acts in a paracrine fashion to activate a vagovagal reflex to influence PS, a direct mechanism of action of CCK on DVC has also been suggested. Recently, it was found that microinjection of CCK-8 in the DVC produced excitatory effects on PS in control and vagal deafferented rats (Viard et al. 2007). In line with previous findings, CCK1 receptors are likely to mediate the central effects of CCK-8 on PS. This suggestion is based on the fact that perfusion of the fourth ventricle with lorglumide significantly attenuated the increases in PS produced by release of endogenous CCK (Viard et al. 2007).

Recent compelling evidence suggests that 5-HT may play a fine-tuning role in regulation of DMV neural activity, because several subtypes of 5-HT receptors, including 5H-T1A, 5H-T1B, 5H-T2 and 5H-T3, are found within the DMV (Twarog & Page, 1953; Manaker & Verderame, 1990; Thor et al. 1992a,b; Steward et al. 1993; Wright et al. 1995). In addition, 5-HT-containing neurons were immunohistochemically localized and widely distributed within the DMV (Steinbusch, 1981; Sykes et al. 1994). In vivo and in vitro studies have shown that activation of 5-HT receptors produced excitatory effects on DMV neurons. It was found that 95% of DMV neurons were strongly excited in response to 5-HT in vitro. It is more likely that this effect was mediated via activation of the 5-HT2 receptors, because application of a 5-HT2 receptor antagonist blocked this excitatory effect. 5-Hydroxytryptamine receptor antagonists at other subtypes were only partly effective or not effective at all (Albert et al. 1996). In agreement with this finding, an electrophysiological study by Browning & Travagli (1999) has demonstrated that exogenous application of 5-HT had excitatory effects on different subgroups of identified DMV neurons. They noted that 69 and 47% of DMV intestinal and gastric neurons, respectively, were excited in response to application of 5-HT into the DMV, and these responses were completely abolished in the presence of ketanserin, a 5-HT2 receptor antagonist. In contrast, another study by Wang et al. (1996) has investigated the effects of ionophoretic application of PBG on identified neurons of the DMV. That study demonstrated that the majority of the DMV preganglionic neurons were excited by PBG. Intravenous and ionophoretic administration of the 5-HT3 receptor antagonists, granisetron and tropisetron, either abolished or attenuated the excitatory effects of PBG, suggesting that these effects were mainly mediated via 5-HT3 receptor activation. Interestingly, 5-HT2 and 5-HT1A receptor antagonists did not alter the excitatory effects of PBG, excluding the possibility that these receptor subtypes participate in the PBG-induced excitation of DMV neurons (Wang et al. 1996). A follow-up study demonstrated that PBG-induced excitation of DMV neurons by PBG was mediated by facilitation of glutamatergic inputs to these neurons. This conclusion was drawn by taking into account several findings, including the finding that ionophoretic application of Mg2+ and Cd2+ onto the identified neurons abolished the PBG-induced excitation of the DMV neurons (Wang et al. 1998). In contrast, an in vivo study has shown that ionophoretic application of 5-HT into the DMV had both excitatory and inhibitory effects on vagal preganglionic neurons. The excitatory effects of 5-HT seem to be mediated via 5-HT1A receptors, because (±)-pindolol and WAY-100635, 5-HT1A receptor antagonists, attenuated these effects (Wang et al. 1995). A plausible explanation for these different findings is that the effects of 5-HT are largely dependent on the 5-HT receptor subtypes present in the DMV.

Most of these previous studies were performed in vitro and used tracing techniques to localize the DMV neurons, and these techniques have several limitations, including potential leakage of tracers to other organs outside the pancreas. In addition, in vitro studies are unable to examine the effects of modulators that act upon vagal afferent neurons that, in turn, influence DMV neurons via a relay in the NTS. Therefore, it was important to localize DMV PPNs in vivo to allow a more detailed investigation of the acute effects of CCK and PBG on the activity of PPNs. To achieve this aim, we made extracellular single-unit recordings from DMV PPNs identified by antidromic activation from the pancreatic branch of the vagus (Mussa et al. 2010). Although three populations of dorsal vagal preganglionic neurons were identified based on their responses to CCK and PBG, some properties were common to all neurons studied, including uniform axonal conduction velocities in the C fibre range and their basal firing rates. Pancreatic preganglionic neurons within the caudal DMV were activated by CCK and PBG, but their number was very small compared with PPNs within the intermediate and rostral DMV that were inhibited by or were insensitive to CCK and PBG, respectively (Mussa et al. 2010). In agreement with previous reports, this study demonstrated that the DMV is a heterogeneous nucleus. Three unique subpopulations of PPNs have not been described previously and may represent different functional roles. This also suggests that the influence of the pancreatic secretagogues, CCK and PBG, on DMV PPNs is more complex than previously thought.

Conclusion

The DMV is a crucial component of the pancreatic vagovagal reflex pathway, and although it was described two centuries ago, interest in studying its physiological roles has only increased recently. Functional studies in vivo have emphasized the relationship between the DMV and the pancreas by showing that electrical and chemical stimulation of this nucleus activates both endocrine and exocrine pancreatic secretion (Ionescu et al. 1983; Berthoud et al. 1990; Mussa & Verberne, 2008; Mussa et al. 2011). In addition, several in vitro studies have documented that all vagal efferents that project to the pancreas originate from the DMV (Fox & Powley, 1986; Rinaman & Miselis, 1987; Loewy & Haxhiu, 1993). However, major questions regarding the exact details of these pathways are yet to be elucidated. This could be due to technical limitations and the lack of neuroanatomical maps that describe the specific motor output from the DMV to the pancreas (Loewy & Haxhiu, 1993). Nevertheless, using whole-cell patch-clamp recording techniques, the electrophysiological and morphological characteristics of DMV PPNs have been described (Browning et al. 2005a). Although there were identifiable differences between the gastric and intestinal preganglionic neurons, it was difficult to find a single specific characteristic to distinguish them (Browning et al. 2005a). Thus, heterogeneity of the PPNs was confirmed by this and other studies. This supports previous findings that PSF is regulated by heterogeneous vagal efferent output from the DMV (Berthoud & Powley, 1987; Berthoud et al. 1990; Wang et al. 1999; Browning et al. 2005a). Studies using retrograde tracers have also enriched our knowledge of DMV PPNs. However, it is clear that both in vivo and in vitro studies are required to advance our knowledge of vagal control of the pancreas.

Evidence has been presented for existence of an inhibitory pathway that controls insulin secretion (Mussa et al. 2011). Previous reports have shown that some gastric functions are under the control of both excitatory and inhibitory motor inputs from the DMV (Travagli et al. 2006). The gastric inhibitory pathway consists of cholinergic and nitrergic preganglionic neurons and non-cholinergic and non-adrenergic postganglionic neurons (Travagli et al. 2006). The involvement of a nitrergic pathway is strongly supported by the finding that peripheral inhibition of NO enhanced the excitatory effects of chemical stimulation of the DMV on insulin secretion (Mussa et al. 2011). Given that postganglionic nitrergic neurons also innervate the pancreas, it is possible that these neurons inhibit the release of ACh and subsequently PS. This does not exclude the possibility that pancreatic nitrergic nerves are tonically involved in inhibition of insulin secretion.

In conclusion, this review reveals that the relationship between the DMV and pancreatic secretion is modulated by several factors, including vagal pathways, central autonomic regulatory circuits, pancreatic secretagogues and GI hormones.

Future directions

Exocrine and endocrine secretions are involved in digestion and regulation of various metabolic processes, respectively. Thus, it is not surprising that malfunction of PSF leads to life-threatening conditions, including pancreatitis, diabetes and pancreatic cancer. For identification of any defects associated with PSF, it is essential to have a complete understanding of the mechanisms that control this process. This will contribute significantly to identification of new therapeutic targets and future potential treatments.

There have been several attempts to identify and clarify the precise relationship between the DMV and pancreatic function, but there is much more to be done. It will be important to identify the role of the DMV, if any, in regulation of other components of PS, such as secretion of water, bicarbonate and glucagon. In addition, further investigations of DMV PPNs that respond to circulating agents that act via vagal afferents, as well as to neuroglucoprivation produced by insulin-induced hypoglycaemia or 2-deoxyglucose, are required. Furthermore, verification of morphological, electrophysiological and functional differences between DMV pancreatic preganglionic neurons that control exocrine and endocrine pancreatic secretion will also be an important goal.

Call for comments

Readers are invited to give their opinion on this article. To submit a comment, go to: http://ep.physoc.org/letters/submit/expphysiol;98/1/25

Appendix

Acknowledgements

This work was supported by an Austin Hospital Medical Research Foundation Grant to A.J.M.V. and a postgraduate scholarship to B.M.M. from the University of Melbourne.