Article Text

Download PDFPDF

Role of nitric oxide in intestinal water and electrolyte transport
  1. Department of Medicine,
  2. American University of Beirut,
  3. Beirut, The Lebanon
  4. Digestive Diseases Research Centre,
  5. St Bartholomew’s and The Royal London
  6. School of Medicine and Dentistry,
  7. Turner Street, London E1 2AD, UK
  1. Dr Turvill.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Since Palmer et al and Ignarroet al showed that vascular endothelial cells could synthesise nitric oxide (NO), this soluble gas has emerged as an important mediator, messenger and regulator of cell function in a number of physiological systems and pathophysiological states.1-4 The effect of NO on the intestinal epithelium, the local microcirculation, the enteric nervous system, and inflammatory cascades has implicated it as a potential mediator of intestinal water and electrolyte transport.5 Data produced by different groups over the past few years have been contradictory, some showing NO as an absorbagogue and others as a secretagogue.

Biology of nitric oxide and its source in the intestine

In biological systems, NO has a half life of less than 5 seconds, rapidly degrading to nitrite and nitrate in the presence of oxygen and water.1 Being soluble in both water and lipid, it freely traverses cell membranes and passes into adjacent target cells.6 The potential sources of NO in the gut are the endothelial cells, the intrinsic intestinal tissue (mast cells, epithelium, smooth muscle, neurones), residing and infiltrating leucocytes (neutrophils and monocytes), reduction of luminal gastric nitrates, and to a lesser extent denitrification by commensal intestinal bacteria.5 Nitric oxide is formed froml-arginine by the action of a stereospecific group of enzymes called nitric oxide synthases (NOS).7 In the gut, NOS have been localised in the myenteric and submucosal neurones in the subepithelial compartment and lamina propria including submucosal arterioles and venules, and in the apical epithelial cells.8-12 Nitric oxide can be produced by enterocytes through both the constitutive and the inducible NOS.13 ,14

Nitric oxide has been considered as a regulator of basal intestinal water transport, as a mediator of pathological conditions where disturbance in water transport plays a role, and as an effector substance in both laxatives and antidiarrhoeal agents.

Role of nitric oxide in basal water transport

Our understanding of the physiological action of NO comes from studies on the effects on water and electrolyte movement of substances that inhibit NOS and thus prevent NO release. Serosal addition of the NOS inhibitorsN G-methyl-l-arginine (l-NMA) orN G-nitro-l-arginine (l-NNA) (0.01–0.3 mM) to unstripped mouse ileum mounted in Ussing chambers resulted in an increase in short circuit current (Isc) which was reversed byl-arginine (0.1–10 mM) implying a net proabsorptive effect for NO on ion transport.15 However, Rolfeet al and Eutamene et al found no effect of NOS inhibition on water transport in rat ileum in vitro and in rat colon in vivo, respectively.16 ,17 Others found that NOS inhibition decreased intestinal fluid absorption in canine Thiry-Vella fistula and rat ileum.18 ,19 Furthermore,N G-nitro-l-arginine methyl ester (l-NAME) induced net water secretion in rabbit ileum when given intra-arterially (0.07 × 10−3mmol/kg/min), in rat jejunum when given intravenously (0.02–4 × 10−3 mmol/kg/min), and in guinea pig ileum when given orally for seven days.20-23 Intraperitoneal and intraluminal administration of a high dose of l-NAME (0.37 mmol/kg and 1–20 mM, respectively) resulted in net water secretion in rat jejunum which was associated with intestinal ischaemia, both effects being prevented by l-arginine.24 It seems from these studies that NO has a physiological proabsorptive effect in the small intestine but whether this is a direct effect on enterocytes, or indirect through the enteric nervous system or the regulation of intestinal blood flow is still not completely clear (fig1, 1–4).

Figure 1

Putative mechanisms of action of nitric oxide (NO) as a mediator of intestinal water and electrolyte transport under resting conditions. The release of background levels of NO modulates absorption by maintaining microvascular villus perfusion (1); by a direct action on epithelial cell function (2) and tight junction integrity (3); and by the activation of neurotransmission within the enteric nervous system (4). Raised NO concentrations mediate secretion by a direct action on epithelial cell function (or by cyclooygenase- and 5-hydroxytryptamine dependent pathways) (A); and by neurotransmission within the enteric nervous system (B) and the autonomic nervous system (C).

Effect of nitric oxide donors on intestinal water and electrolyte transport

The effect of NO donors on small and large intestinal water and electrolyte transport depends on the route of administration and the method used to study the effect. In vitro studies showed that addition of sodium nitrite to the serosal side of mouse ileum resulted in a decrease in Isc, implying that NO has a proabsorptive effect.15 Li et al found that a saturated NO solution had no effect onIsc in rat ileum whereas other investigators demonstrated an increase in Isc after serosal addition of the NO donors sodium nitroprusside (SNP), isosorbide dinitrate (ISDN), S-nitroso acetyl penicillamine (SNAP), or saturated NO solutions to guinea pig small intestine, rat ileum, rat and human colon, suggesting that at high doses NO has a secretory effect (fig 1, A–C).16 ,25-29

In vivo studies are more controversial. Even before the discovery of NO, Hegarty et al and Hellieret al showed that l-arginine, unlike other amino acids, induced water secretion when perfused in human jejunum.30 ,31 In addition, Thomaset al showed, 15 years ago, that SNP given subcutaneously induced fluid accumulation in mouse intestine.32 Similarly, we found that intraluminal infusion of l-arginine (20 mM) into rat jejunum induced water and electrolyte secretion which could be inhibited by a low concentration of l-NAME (0.1 mM); however, subcutaneous administration of l-arginine (0.47 and 2.37 mmol/kg) had no effect.26 Furthermore, Shirgi-Degen et al found that intravenousl-arginine (0.04 mmol/kg/min) or SNP (0.07 × 10−3 mmol/kg/min) had no effect on basal water transport in rat jejunum, whereas other investigators showed thatl-arginine and other NO donors promoted water absorption in rabbit ileum and canine jejunum.20 ,21 ,24 In addition, Wapnir et al have shown recently thatl-arginine added to oral rehydration solutions could improve water absorption when perfused in rat jejunum.33These studies suggest that the predominant role for NO is as a secretagogue in both the small intestine and colon, an effect clearly demonstrated in vitro but with contradictory results in vivo.

Role of nitric oxide in hypersecretory states

The role of NO in intestinal secretory states has been studied in relation to enterotoxins, serotonin, some inflammatory mediators, bile acids, and laxatives. Rolfe et al and Haydenet al found that NOS inhibition had no effect on changes in Isc induced byEscherichia coli heat stable toxin (STa) in stripped rat ileum and pig jejunum and colon.16 ,34However, l-NAME notably decreasedIsc in unstripped ileum and decreased fluid secretion caused by STa in vivo implying a significant role for NO in STa induced secretion.16 By contrast, Shirgi-Degenet al found that STa induced secretion in ligated rat jejunal loops was increased by intravenousl-NAME and inhibited by intravenous l-arginine and SNP and therefore concluded that NO has a proabsorptive tone in the intestine.21 Similar controversy is observed with cholera toxin (CT). Beubler et al recently found that intravenous l-NAME enhanced CT induced fluid secretion in ligated rat jejunal loops whereas intravenous l-arginine inhibited it.35 Qiu et al used a similar model of ligated rat jejunal loops but found no change in CT induced fluid secretion after administration of NOS inhibitors or NO donors.36 Using rat jejunum in situ, we found that the NOS inhibitors l-NAME and l-NMA, and the NO precursor l-arginine caused a reduction in CT induced secretion, implying a dual role for NO as a secretagogue and absorbagogue.37 Finally, Reddix et al demonstrated that CT increased basal nitrite level in guinea pig ileum mounted in Ussing chambers and that l-NAME inhibited CT induced secretion during the first 30 minutes of exposure (fig 2).38

Figure 2

Putative mechanisms of action of nitric oxide (NO) as a mediator of intestinal water and electrolyte transport in hypersecretory states. The inhibition of NO release inhibits secretion by: modulating microvascular villus perfusion (1); a direct action on epithelial cell function (2); the activation of neurotransmission within the enteric nervous system (myenteric plexus) (3) and the autonomic nervous system (4); and by the inhibition of free radical formation (5). Increased levels of NO inhibit secretion by a direct action on epithelial cell function (A) by: the stabilisation of enterochromaffin cells (B) and mast cells (C); inhibiting 5-hydroytryptamine (5-HT) induced secretion (D); and by inhibiting intestinal motility (E).

5-hydroxytryptamine (5-HT) is a neurotransmitter and a potent intestinal secretagogue released from enterochromaffin cells by CT and plays an important role in the pathogenesis of CT induced secretion.39-41 Kadowaki et aldemonstrated that inhibition of NOS ameliorated 5-HT induced chloride secretion in guinea pig distal colon in vitro and noticeably decreased 5-HT induced diarrhoea in mice; both effects were reversed by the NO precursor l-arginine.42 These investigators and others concluded that NO may play an important role in the secretory response to 5-HT which could be partly due to the activation of neurones that generate NO.43 By contrast, Beubleret al found that intra-arterial 5-HT induced secretion was inhibited by l-arginine.35

Nitric oxide production is increased in inflammatory bowel disease (IBD)43 ,45 but whether this plays a role in the pathogenesis of IBD or in the associated diarrhoea has not been elucidated. Recently, an increase in the production of interleukin-1 (IL-1) in the mucosa of patients with IBD has been demonstrated.46 IL-1 causes colonic water secretion in rats, and its effects on endothelial cells are mediated by local synthesis and release of NO.47 ,48 In this context, Eutamene et al showed that IL-1β induced fluid secretion in rat colon could be inhibited by intraperitoneal administration of l-NMA, an effect reversed byl-arginine.17 These results imply that increased IL-1 production in IBD could cause colonic secretion by activating NOS and releasing NO, playing a final role in IL-1 induced hypersecretion.

Bile acids stimulate fluid and electrolyte secretion in jejunum, ileum and colon in both animals and humans.49 Mascoloet al showed in a rat model that bile salt induced diarrhoea and intestinal fluid secretion could be inhibited by intraperitoneal l-NAME (9–90 × 10−3mmol/kg) and dexamethasone.49 This inhibition was reversed by l-arginine and isosorbide-5-mononitrate (ISMN). Induction of NO formation by bile salts has also been observed in human colon.50

The exact mechanism of intestinal secretion of many laxatives is not understood. Mascolo et al and Capassoet al demonstrated that castor oil induced diarrhoea and intestinal secretion could be notably decreased by NOS inhibition in rats, an effect that could be reversed byl-arginine, implying that NO is involved in the laxative action of castor oil.51 ,52 Similar results were obtained with other laxatives including magnesium sulphate, bisacodyl, phenolphthalein, senna, and cascara.53-55 By contrast, Beubler et al came to the conclusion that NO is important in the antidiarrhoeic effect of loperamide and showed that loperamide could reverse the secretory effect ofl-NAME.21 They also showed that loperamide induced NO formation in freshly prepared jejunal epithelial cells and therefore could exert its proabsorptive effect via this route.

Possible mechanisms of action of nitric oxide

Nitric oxide could be involved in intestinal water transport either by acting directly on the epithelium or indirectly by stimulating neuronal reflexes, or by stimulating the release of other agents from the epithelium or the enteric nervous system that can modify water transport, or by affecting mucosal blood flow.

Nitric oxide can activate soluble guanylate cyclase resulting in an increase in cGMP, a potent activator of intestinal secretion.56-60 SNP and NO gas evoke electrolyte secretion in rat small bowel and colon associated with increased cGMP levels; this phenomenon is inhibited by the guanylate cyclase inhibitor, methylene blue, implying a significant stimulation of soluble guanylate cyclase in the intestine by NO (fig3).12 ,16 ,26-28

Figure 3

Putative actions of nitric oxide (NO) on the enterocyte. Actions of NO on the enterocyte include the maintenance of tight junctional integrity (1); the modulation of cell respiration (2); the activation of soluble guanylate cyclase (GC) (3); the opening of apical chloride channels (4); the activation of basolateral potassium channels (5); the stimulation of cyclooxygenase activity to generate prostaglandins (PG) (6); and the modulation of receptor activity (7).

Nitric oxide stimulates cyclooxygenase activity directly, independently of cGMP and Wilson et al reported a significant increase in cGMP and prostaglandin (PG) E2production in colonic mucosal strips stimulated with SNP.12 ,61 Cyclooxygenase inhibitors notably attenuate changes in Isc induced by NO donors in animal small intestine and colon and human colon.26-29

Nitric oxide could also produce effects on water transport by its action on enteric neurones. Neuronal inhibition by tetrodotoxin inhibits the secretory effect of NO donors, but whether NO has a direct stimulatory effect on the enteric nervous system or through activation of cGMP as a second neuronal messenger needs to be investigated further (fig 1B).16 ,27-29 ,62 Rolfe et al concluded that STa increases electrogenic Clsecretion across intact rat ileum in vitro by activating a capsaicin sensitive (afferent) and a NO dependent (efferent) myenteric plexus secretory reflex, proposing NO as a neuronal secretagogue (fig2).16

Vasoactive intestinal polypeptide (VIP) is present in enteric neurones and has been proposed as a stimulatory transmitter of secretory processes in the submucous plexus and the mucosa.63 ,64 It has been demonstrated that VIP release from rat enteric synaptosomes can be stimulated by the NO donors SNP and 3-(morpholino) sydnonimine (SIN-1), as well as by l-arginine.65Similarly, in isolated perfused canine ileum, VIP output was reduced byl-NNA and increased by NO donors, so NO can exert a secretory effect by releasing VIP from nerve terminals.66Furthermore, VIP releases NO, establishing the scenario that these two secretagogues could act synergistically, with NO potentially amplifying VIP’s biological effects.67

Nitric oxide is also known to modulate free radical generation and is capable of combining with free radicals such as peroxide to form the highly toxic peroxynitrite which may have an effect on epithelial cell membrane lipid peroxidation.68 This could alter normal physiological regulation of electrolyte transport in the small intestine and colon. Oxyradicals have been shown to stimulate intestinal electrolyte transport in rabbit and rat intestine and thus NO and other nitrogen oxides could stimulate intestinal secretion by the nature of their free radical structure.69-71 Tamaiet al demonstrated that the chloride transport blocker, bumetanide, inhibited the NO induced increase inIsc in rat colon, implying that NO exerts its secretory effect by opening chloride channels (fig3).28

Nitric oxide is a vasodilator and its continuous endogenous production is important in maintaining mesenteric microcirculation and mucosal integrity. NOS inhibition leads to a significant decrease in mesenteric blood flow.72 ,73 It is well known that vasoconstrictor agents, like vasopressin, cause water and electrolyte secretion in the small intestine.74 In addition, Robinsonet al showed that intestinal ischaemia causes net water and electrolyte secretion secondary to a profound inhibition of water and sodium absorption by villus whereas the intact crypt cells continue to secrete sodium, chloride and water.75 ,76 This effect of NO on blood flow could play a role in modulating intestinal water transport (figs 1 and2).24

Nitric oxide is known to inhibit mast cell degranulation, and as mast cells and enterochromaffin cells degranulate by a similar calcium dependent mechanism, it is postulated that NO would also inhibit enterochromaffin cell degranulation.77 CT induced depletion of tissue 5-HT concentrations was prevented byl-arginine implying that the elevated levels of NO inhibit enterochromaffin cell degranulation and may actually, under these circumstances, result in the failure of activation of 5-HT dependent secretory pathways (fig 2B) (personal communication).

Finally, Shirgi-Degen et al showed that NO could activate basolateral K+ channels, an effect which may mediate its proabsorptive properties (fig 3).78


Nitric oxide acting at many different sites in the intestine with some opposing effects could explain the controversy in the literature on the final effect of NO on water transport. It seems that NO could act as both an absorbagogue and a secretagogue depending on the circumstances and on the site of delivery. Studies suggest that physiologically NO promotes fluid absorption, but in pathophysiological states it may be produced in high concentrations capable of evoking net secretion. The development of selective NOS inhibitors will assist in dissecting out these different aspects of NO function. Ultimately, the inhibition of the pathological rather than protective effects may yield therapeutic benefits. Until that time, however, it remains to be seen whether NO is a mucosal “friend” or “foe”.


nitric oxide
nitric oxide synthase
NG-nitro-l-arginine methyl ester
sodium nitroprusside
isosorbide dinitrate
S-nitroso acetyl penicillamine
Escherichia coli heat stable toxin
cholera toxin
inflammatory bowel disease
vasoactive intestinal polypeptide
3-(morpholino) sydnonimine
guanylate cyclase



  • Leading articles express the views of the author and not those of the editor and editorial board.