Article Text


Vagotomy inhibits the jejunal fluid secretion activated by luminal ileal Escherichia coli STa in the rat in vivo
  1. V E Rolfe,
  2. R J Levin
  1. Department of Biomedical Science, University of Sheffield, UK
  1. Dr V E Rolfe, Waltham Centre for Pet Nutrition, Freeby Lane, Waltham-on-the-Wolds, Melton Mowbray, Leicestershire LE14 4RG, UK.


BACKGROUND Escherichia coli heat stable enterotoxin (STa) is a major cause of secretory diarrhoea in humans.

AIMS To assess the effects of instilling STa into the ileum on remote fluid secretion in the jejunum and colon in rats in vivo by a gravimetric technique.

METHODS AND RESULTS Ileal STa (55 ng/ml) stimulated fluid secretion in both ileal and jejunal loops but not in the colon. The fluid secretion induced by ileal STa was inhibited by l-NAME (Nω -nitro-l-arginine methyl ester, 40 mg/kg intraperitoneally) but not by d-NAME (Nω -nitro-d-arginine methyl ester). Ileal carbachol (183 mg/ml) instilled into the lumen stimulated ileal secretion but not jejunal secretion, and was unaffected by l-NAME. Capsaicin (10 μM), instilled luminally with STa in the ileum, blocked both the ileal and jejunal fluid secretion. Acute bilateral vagotomy prevented luminal ileal STa from inducing jejunal fluid secretion but not from activating ileal fluid secretion.

CONCLUSION IlealE coli STa stimulates remote secretion in the rat jejunum but not in the colon, probably by a nitrinergic, vagal reflex mediated by C fibres. This neural pathway will amplify the action of the toxin in its generation of secretory diarrhoea.

  • intestinal fluid secretion
  • Escherichia coli STa
  • vagus
  • nitric oxide
  • C fibres

Statistics from

Escherichia coli heat stable enterotoxin (STa) is a major cause of human secretory diarrhoea. It induces electrogenic Cl secretion and inhibits NaCl absorption in the small intestine,1 acting by binding to, and activating guanylate cyclase at the luminal membrane which elevates enterocyte cyclic GMP.2 The enteric nervous system also participates in its action by creating a fluid secretion when both STa and STb are instilled into the jejunum of rats and cats.3Studies in vitro confirmed that STa could act indirectly to cause electrogenic secretion through a neural mechanism which was shown to involve a nitrinergic, myenteric, secretory reflex mediated by C afferent fibres, and that the activation of fluid secretion by STa in vivo could be blocked by pretreatment with l-NAME (Nω -nitro-l-arginine methyl ester).4 The mechanism of activation of the C fibres by STa is unknown, but it may involve the release of trigger substance(s) from the enterocytes or other cells of the intestinal epithelium. Nzegwu and Levin have recently reported that the NK2 tachykinin receptor antagonist MEN 10 627 specifically inhibited fluid secretion activated by E coli STa in rat jejunum in vivo.5

Numerous C fibres originate from the intestine with their cell bodies in the dorsal root ganglion and the fibres enter the spinal cord between segments T1 and L2.6 In addition, ileal sensory afferent fibres enter the nucleus of the tractus solitarius in the brainstem via the vagus in the rat.7 The pattern of innervation suggests that E coli STa placed in the lumen of the ileum may induce secretion reflexively not only in the ileum but also in a remote part of the intestine by mediation of C fibres.

We thus investigated whether the stimulation of afferents by luminal STa in the rat ileum in vivo influenced secretion in a jejunal segment with no luminal STa. Preliminary data have been reported to the Physiological Society.8

Materials and methods


Fed, male albino rats (body weight 230–250 g, Sheffield University Strain) were anaesthetised with Na+pentobarbitone (60 mg/kg body weight intraperitoneally). Fluid transport (g/g gut dry weight/h) was measured gravimetrically for 30 minutes as described previously4 in 10 cm tied off loops of mid-ileum, mid-jejunum, and a 5 cm loop of mid-colon. The “stimulating” loop was at least 10 cm arboral to the jejunal loop and contained 0.8 ml of 0.9% NaCl with either STa (55 ng/ml), carbachol (183 mg/ml), d- or l-arginine (20 mg/ml), or together with capsaicin (10 μM). The “indicator” jejunal or mid-colonic loops contained either 0.8 ml or 0.4 ml of 0.9% NaCl respectively. Some rats were pretreated with either l- or d-NAME (40 mg/kg body weight intraperitoneally) 15 minutes before luminal instillation of the fluids. A number of rats were acutely, surgically vagotomised in the neck under the pentobarbitone anaesthesia and a cannula placed in the trachea; they were then used immediately for the secretion experiments. Sham operated controls had identical surgery except that the carotid/vagal complex was manipulated with blunt scissors. The rats were killed after use by anaesthetic overdose, thoracotomy, and heart incision.


Results are given as the mean (SEM) fluid transport with the numbers of rats used in brackets. Absorption was regarded as a negative fluid transport and secretion as a positive fluid transfer. Statistical significance of the results was tested using Student’s unpairedt test and multiple comparisons were made by ANOVA with significance accepted at p<0.05.


All chemicals were obtained from Sigma (Poole, Dorset, UK).



When saline was placed in both ileal and jejunal loops in control, untreated rats the fluid transport observed was that of net absorption which was not significantly different (ileum −8.4 (1.5) (n=4) versus jejunum −5.5 (1.5) (n=4) g fluid/g gut dry weight/h, p>0.05). Instillation of l-arginine (20 mg/ml), a substrate for nitric oxide synthesis, into the ileum, induced fluid secretion not only in the ileum but also in the jejunum which had only 0.9% NaCl instilled (ileum 3.4 (2.5) (n=4) versus jejunum 2.2 (2.4) (n=4) g fluid/g gut dry weight/h). Ileal instillation of d-arginine (20 mg/ml), however, had no effect on the usual absorption for fluid in the ileum and jejunum (ileum −6.1 (1.8) (n=6) versus jejunum −3.4 (0.9) (n=6) g fluid/g gut dry weight/h).


Table 1 shows the effects of d- and l-NAME treatment on the fluid secretion induced by ileal STa and carbachol. Instillation of STa into the ileum of d-NAME pretreated rats induced secretion in both the ileal and jejunal loops.l-NAME, however, inhibited the secretory action of STa in the ileum and jejunum and net absorption resulted which was significantly different from d-NAME controls (p<0.001). Ileal luminal carbachol induced secretion only in the ileum, having no significant effect on the absorption of the jejunum. This secretory action of luminal carbachol in the ileum was unaffected byl-NAME treatment compared with d-NAME.

Table 1

Effects of ileal STa (55 ng/ml) and carbachol (183 mg/ml), and pretreatment of rats with l- ord-NAME (40 mg/kg) on fluid transport in the ileum and a saline filled jejunal loop


Table 2 shows the effect of vagotomy on the ability of ileal STa to induce secretion in the jejunum. In the control sham operated group, ileal STa not only induced a significant fluid secretion in the ileal loop but also induced secretion in the jejunum. However, in the vagotomised group, although ileal STa induced a fluid secretion in the ileum it had no effect on the jejunal loop which displayed absorption, significantly different from sham operated controls (p<0.02).

Table 2

Effects of bilateral vagotomy on ileal and jejunal fluid movement


Table 3 shows the effect of luminal capsaicin on the ability of ileal STa to induce jejunal and colonic fluid secretion. Control luminal STa in the ileum induced significant secretion in both the ileal loop and the jejunal loop but not in the colon which displayed absorption. When capsaicin (10 μM) was added to the STa, a significant reduction in fluid secretion was observed in the stimulating ileal loop (p<0.05) and in the indicator jejunal loop (p<0.01) where net absorption occurred. Colonic absorption was unchanged by the ileal capsaicin.

Table 3

Effects of luminal STa and STa plus capsaicin (10 μM) in the stimulatory ileal loop on fluid movements in the ileum, jejunum, and colon   


Neither bilateral vagotomy nor capsaicin altered basal fluid transport compared with control ileum and jejunum (control ileum −4.3 (1.4) (n=4) versus vagotomised −2.6 (0.6) (n=4) versus capsaicin −2.5 (0.7) (n=4) g fluid/g gut dry weight/h, p>0.05, ANOVA; and control jejunum −1.8 (1.1) (n=4) versus vagotomised −1.8 (1.5) (n=4) versus capsaicin −4.0 (1.2) g fluid/g gut dry weight/h, p>0.05, ANOVA).


E coli STa stimulates a myenteric secretory reflex in the rat ileum mediated by capsaicin sensitive C fibres.4 The gastrointestinal tract is extensively innervated by vagal and spinal afferent fibres,6 a pattern of afferent innervation that suggests that E coli STa placed in the lumen may activate secretion in remote intestinal regions not in contact with the toxin. Our results confirmed this, clearly showing that instillation of the enterotoxin STa into the rat ileum activated both ileal and jejunal fluid secretion and both could be inhibited by the nitric oxide synthase inhibitorl-NAME but not the inactive isomer d-NAME. Instillation of the cholinergic agonist carbachol into the ileum lumen activated ileal secretion but had no significant effect on jejunal fluid transport, and furthermore, was unaffected by l-NAME. Thus, non-specific systemic effects of l-NAME could not be the cause of the inhibition of the STa induced jejunal fluid secretion. Instillation of the nitric oxide synthase substratel-arginine into the ileum also induced both ileal and jejunal secretion, but equivalent d-arginine did not alter fluid transport, indicating that the fluid secretion induced was not an osmotic effect. Recently, the nitric oxide (NO) synthase cofactor tetrahydrobiopterin was implicated in the regulation of electrogenic secretion stimulated by STa and l-arginine in the mouse ileum in vitro.9 This strongly suggests that NO is involved in intestinal secretion, and our present results show that NO is implicated in the activation of remote jejunal secretion.

It should be mentioned that the experimental evidence for the regulating role of nitric oxide in small intestinal fluid secretion (prosecretory) or absorption (proabsorptive) in vivo is conflicting. Practically all the evidence relies on the use of eitherl-NAME as an inhibitor of NO generation and on NO donors and interpretation of their actions on basal and secretagogue induced fluid transport. Rolfe and Levin4 were the first to report that in the rat in vivo STa induced fluid secretion was inhibited by the intraperitoneal injection of l-NAME. Gaginella and colleagues10 and Izzo and colleagues11 also found that in the rat intraperitoneal or subcutaneous injectedl-NAME inhibited the fluid secretion induced by bisacodyl and phenolphthalein in the former, and by senna and cascara in the latter. All these studies thus indicated a prosecretory role for NO. In later in vivo rat experiments,12 however,l-NAME enhanced the fluid secretory effect of luminal STa interpreted as a proabsorptive action of NO. In a study by Qui and colleagues13 NOS inhibitors augmented the ileal secretion induced by Clostridium difficile toxin A in the rat. We cannot explain the clearly contradictory findings of these authors but it may be simply due to differences in the method of administering the inhibitor l-NAME. In these studies the antagonist was given by continuous venous infusion. This type of administration may cause l-NAME to have different actions than when it is injected subcutaneously or intraperitoneally in slug doses. However, despite the contradictory findings about the possible role of NO in fluid transport, they do not influence the basic finding of the present experiments that luminal STa in the ileum activates remote secretion in the jejunum through a neural mechanism.

Previous in vitro studies4 14 showed that mucosal STa induced ileal electrogenic secretion that was blocked in the presence of capsaicin only in ileal preparations where the myenteric plexus was intact and in communication with the submucosal plexus. This indicated that the action of STa was mediated by afferent fibres which connected with the myenteric plexus. Furthermore, the action of the toxin was also blocked by l-NAME only when the myenteric plexus was present, suggesting that the secretory efferent pathway may have had a nitrinergic component, although we cannot exclude a nitrinergic afferent component. Fluid secretion activated by STa in vivo was also inhibited by l-NAME in the rat ileum4 and by luminal capsaicin in the jejunum, ileum, and colon.15 It is interesting in this context to find that the myenteric plexus is also implicated in the secretion of fluid induced by cholera toxin in the rat jejunum.16

In the present experiments in vivo, placing capsaicin with the STa in the ileal loop prevented the fluid secretion both in the ileal and jejunal loop, while capsaicin alone did not influence basal absorption. Capsaicin selectively targets sensory afferent fibres17; thus it appears that mediation of C fibres is essential for STa ileal activated jejunal secretion to be induced. TheClostridium difficile toxin A also causes fluid secretion inhibited by capsaicin treatment whereas cholera toxin appears unaffected.18 As mentioned previously, Nzegwu and Levin5 have recently shown the involvement of the tachykinin receptor NK2 in the action of STa; it is therefore possible that the activation is via a released tachykinin(s). Thus, a toxin specific trigger mechanism activates mucosal afferent fibres but the exact mechanical or chemical nature of these substances has still to be established.

Our results show that bilateral vagotomy prevented ileal STa from inducing jejunal secretion, but had no effect on the activation of ileal secretion by STa. Furthermore, bilateral vagotomy did not affect the basal absorption in either the ileum or jejunum. The STa toxin in the ileal loop must activate jejunal secretion either by neural mediation via the cord and vagal efferents, or by vagal afferents and efferents. Further investigations are necessary to characterise fully the afferent neural pathways involved and to determine whether there is bipolarity of activation—that is, whether the STa ileal activation of jejunal secretion can be mirrored by a similar STa jejunal activation of ileal secretion. Anterograde labelling techniques have allowed the organisation of intestinal reflexes to be viewed, and ileal afferent fibres have been shown to converge on the dorsal motor nucleus of the vagus in the rat.7 Here, efferent vagal neurones project to the intrinsic intestinal ganglia and structures in the brain, thus coordinating intestinal functions. The routes may provide a central pathway for the activation of remote jejunal secretion by ileal STa.

Although the ileal STa clearly induces secretion in both the ileum and jejunum, we did not observe any activation of fluid secretion in the mid-colon which contrasts with a study using cholera toxin.19 The group reported that instillation of cholera toxin into the rat jejunum evoked not only jejunal fluid secretion but colonic secretion as well. In this study transection of the intestine to isolate the colon actually caused an inhibition of cholera toxin stimulated colonic secretion, suggesting that the normal stimulatory mechanism was via the enteric nervous system. However studies in the guinea pig have suggested that individual neurones only have a short (approximately 12 mm) projection.20 The jejunum in the rat could be some 30–40 cm away, a remarkably long projection. At present it is not possible to characterise the method of neural transmission that STa uses to activate remote secretion. Another study did not show a remote effect to mucosal STa in rabbit ileum in vitro. Hubel and colleagues21 were unable to show any activation of electrogenic secretion in a segment of small intestine located about 0.4 cm distal to a stimulating segment in contact with STa, but their in vitro experiments were conducted with the intestine stripped of its outer muscle which removes the myenteric plexus, thus removing neural connections which in the rat are known to mediate the secretory reflex. In our experiments, while severing the vagus reduced the remote fluid secretion jejunally, we cannot completely rule out the enteric nervous system. However this has to be mediated by C fibre activity.

One intriguing feature of the activation of fluid secretion in vivo byE coli STa is that it can be either totally or very severely reduced by treatment with either luminal capsaicin14 or by pretreatment of the rat withl-NAME8 (confirmed by present results). This suggests strongly that in vivo the mediation of electrogenic Cl secretion by STa activating guanylate cyclase and generating cyclic GMP in the enterocytes is not an important ionic driving force for fluid secretion. Early studies3 first drew attention to the fact that part of the fluid secretion activated by STa in vivo was through the nervous system. Jodal22suggested that 60–70% of the fluid secretion was so mediated, although in some cases nearly total inhibition was observed using nerve blocking agents. In our present rat studies and in those previously mentioned, capsaicin and l-NAME often show such total inhibition of STa activated fluid secretion. It thus seems likely that the usual mechanism of activation of net fluid secretion by luminal STa in vivo is by the neural path and not by the direct action of STa on the enterocyte. A recent review23 may explain this paradox. While STa causes electrogenic Cl secretion by a direct action on villus enterocytes, according to Jodal and Lundgren the upper third of the villus interstitium exhibits such a high osmotic pressure, both in the absorptive and secretory state, that this hyperosmolarity prevents net fluid secretion across the villus. However, the luminal STa, by activating the C fibre reflex mediated through the myenteric plexus and submucosal ganglion, induces the crypts to secrete Cl and fluid in the latter moving into the lumen because the interstitial osmolarity around the crypts is far less than in the upper villus. Using the intestine in vitro, especially with muscle stripped preparations devoid of the myenteric plexus (thus C fibre mediated secretion) and only measuring the short circuit current as an index of secretion creates conditions that emphasise the direct action of STa on the enterocytes to create the electrogenic secretion. It is only when the myenteric plexus is present (intact intestine in vitro and in vivo) that STa can activate electrogenic ion and fluid secretion from the crypts. Interestingly, in the guinea pig small intestine, the excitatory postsynaptic potential input (the classical nicotinic receptor response) to the submucosal neurones is largely derived from the myenteric plexus24 and vagal efferent fibres only reach this plexus.25

The effect of this STa activated neural path may also be of importance in relation to the action(s) of the recently discovered endogenous intestinal secretagogue guanylin. While guanylin is a weaker endogenous ligand for the E coli STa receptors one of the highest concentrations of intestinal guanylin is found in the ileum.26 Although there has been some contention as to the exact localisation of guanylin in the small intestine, the initial localisation in Paneth cells27 appears to have been an artefact; more recent studies have now localised it by immunocytochemistry in Goblet cells.28-30 The presence of guanylin in the lumen of the intestine has been reported.31 Presumably it is released when the Goblet cells discharge. It has been suggested that by activating fluid secretion it could make the discharged mucus more fluid and lubricative. Because guanylin is a weaker agonist than STa in inducing fluid secretion its putative physiological role in the ileal lumen activating remote secretion in the jejunum clearly needs investigation.

In conclusion, luminal STa in the ileum in vivo can activate fluid secretion in the jejunum but not in the colon, by a neural mechanism(s) that is inhibited by vagotomy, l-NAME, and capsaicin. This neural pathway(s) will clearly amplify the action of the toxin in its generation of secretory diarrhoea.


Part of this study was undertaken during the tenure of an MRC scholarship by V Rolfe in cooperation with Reckitt and Coleman, Kingston-upon-Hull, UK.


View Abstract


  • Abbreviations:
    N ω-nitro-d-arginine methyl ester
    N ω-nitro-l-arginine methyl ester
    Escherichia coli heat stable enterotoxin

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.