432 Gut, 1992,33,432-434 LeadingArticle Gut: first published as 10.1136/gut.33.4.432 on 1 April 1992. Downloaded from The diarrhoea of famine and severe malnutrition - is glucagon the major culprit? Intractable diarrhoea is usually the terminal condition of cant change in the enteric neural innervation ofthe smallintes- victims of famine and severe malnutrition. Despite many tine as it could be obtained even in the presence of the neural studies on the effects of starvation and malnutrition on the toxin tetrodotoxin,'4'5 noris there any significantchange in the absorptive functions ofthe small intestine in animals,'"4 and a enterocyte's receptor sites for the neural transmitters as the few in man,5'7 no satisfactory explanation exists for this hypersecretion could be obtained with agents that bypassed diarrhoea. While it has often been attributed to bacterial these sites.'4"' However, cyclic AMP concentrations were infection, extensive bacteriological studies failed to confirm higher in enterocytes from starved intestine, both in the basal this and it is estimated that only approximately 15% of cases state and when stimulated by secretagogues.2' The reversal of have a diarrhoea of infectious origin.8-" In the most detailed the basal absorptive tone in the fed ileum to a secretory one analysis of the effects of extreme malnutrition on Danish correlated well with the increase in cyclic AMP found in this subjects (who had been incarcerated in a concentration area on starvation.2' Confirmation of the hypersecretory camp), Helweg-Larsen, Hoffmeyer, Kieler et all identified action of starvation on small intestinal secretory function in three types ofdiarrhoea. The most prevalent was an endemic, vitro has been reported in the jejunum of starved piglets.22 progressive, afebrile 'hunger diarrhoea' while 'infectious The obvious question that arises from the rat studies is, diarrhoea' was infrequent. Sudden realimentation caused an what causes the hypersecretory behaviour of the enterocytes 'alimentary diarrhoea'. All were potentially lethal in the and colonocytes? Before we can answer this, it is necessary to severely weakened state of the subjects. According to describe briefly the current model23 for the secretory Roediger,'2 the terminal diarrhoea of kwashiorkor and behaviour of enterocytes (and colonocytes). Chloride is marasmus (corresponding to hunger diarrhoea) is the result transported into the cells against its electrochemical gradient of lack of luminal nutrition of enterocytes, and especially of by the Na+-K+-2Cl- contransporter in the basolateral mem- colonocytes, which severely reduces their absorptive brane. The intracellular Na+ ion concentration is kept low by function. In his explanation, the diarrhoea is mainly caused the action of the Na+-K+-adenosine triphosphatase pump, http://gut.bmj.com/ by the loss of the reabsorptive (salvage?) capacity of the also in the basolateral membrane, pumping out 3 Na+ ion in colon. exchange for 2 K+ ions. The membrane is also equipped with Ethical problems prevent controlled studies on the K+ channels that when opened allow the K+ to leave the cell secretory behaviour of the intestine in starved and severely under the influence of the K+ gradient. At the luminal brush malnourished individuals but their lack in an experimental border ofthe cell, Cl- channels which are normally closed can animal model is surprising. In order to investigate the effects be opened by intracellular second messengers (possible cyclic of starvation and undernutrition on the secretory function of AMP or Ca++) allowing the intracellular Cl- ions to be driven the small and large intestine, the rat has been used as an out of the cell by their concentration difference and the on September 28, 2021 by guest. Protected copyright. animal model. Starvation for 24 hours did not significantly transluminal membrane potential difference (pd, cell interior alter small intestinal secretory function but between 24 and negative to the lumen) as electrogenic Cl- secretion. This 72 hours of starvation the duodenum,'3 jejunum,'4 and outward movement ofCl- depolarises the luminal membrane ileum'5 all showed hypersecretory activity compaired with potential, which reduces the driving force and attenuates Cl- fed intestine when activated to secrete by neurotransmitters secretion. Secretagogues, however, activate the basolateral and secretagogues. Moreover, in the case of the ileum, even K+ channels (probably by the raised intracellular Ca++), and the fed basal absorptive tone was converted to a secretory one the passive outflow of intracellular K+ hyperpolarises the after 24 hours offood withdrawal and its magnitude increased membrane. Electrical coupling via the intercellular pathway on continued food deprivation. 15 Further studies showed that maintains the luminal membrane pd which preserves the the proximal colon,'6 and even the rectum,'7 from rats starved electrical potential drive for the Cl- movement. The activation for three days responded to cholinergic agonists by electro- ofthe enterocyte's luminal Cl- and basolateral K+ channels is genic Cl- hypersecretion. Analysis of the types of secretory normally via the cholinergic and vipergic neurotransmitters responses elicited by various neurotransmitters, bacterial acetylcholine and vasoactive intestinal polypeptide (mediated toxins, and secretagogues showed clearly that in the case of through intracellular second messengers) but other agencies the small intestine, all of the agents caused a hypersecretion such as prostaglandins, 5HT, bacterial toxins, and various in the starved rat.'14' In the proximal colon, however, only hormones can also directly or indirectly activate secretion. those agents that activate secretion by increasing intracellular Starvation in the rat reduces the mitotic activity in the Ca++ (cholinergic agents, 5-hydroxytryptamine (5HT), A duodenal/jejunal crypts, they decrease in size and the supply 23187) elicited the hypersecretory response,'6 those acting of new enterocytes onto the villus is slowed.3 As old through increases in intracellular cyclic adenosine monophos- enterocytes are still sloughed off at the villi tip exfoliation phate (AMP) (theophylline, dibutyryl cyclic adenosine mono- zones, villus height, and cell population decrease. At first it is phosphate, forskolin, and prostaglandin2) or in mid colon tempting to explain the increased secretion of the starved through cyclic guanosine monophosphate'8 (Escherichia coli small intestine as being the result of this decreased cell STa) did not cause electrogenic Cl- hypersecretion. In fact, turnover and villus size on the basis of the simplistic model they often elicited less secretion compared with the fed state. '9 that villi absorb and crypts secrete. Ifthe decreased enterocyte This lack of effect of cyclic AMP mediated secretagogues to population on the villi had a reduced absorption while the induce hypersecretion has been confirmed in the starved crypts secreted even normally, then the imbalance could mouse colon.20Thehypersecretion is not caused byany signifi- cause an apparent hypersecretion. We have previously The diarrhoea offamine and severe malnutrition - isglucagon the major culprit? 433 discussed this as being unlikely for a number of reasons.'4 and stimulates the uptake of Na+-dependent glucose and Firstly, in the starved state the enterocytes are on the villus amino acids. It is highly likely that the chronic increase in longer and this ageing seems to increase their absorptive plasma glucagon brought about by starvation induces changes functions3 14; secondly, the crypt cell population also decreases in the enterocytes (perhaps by raising cyclic AMP levels?) Gut: first published as 10.1136/gut.33.4.432 on 1 April 1992. Downloaded from yet secretion increases; thirdly, the enterocytes on the villus creating the hyperpolarisation of the luminal membrane are now known to be able to secrete Cl- electrogenically24; which can then give rise to the various enhanced transport and finally, there is controversy as to whether human villi do properties of the starved intestine. Interestingly, even when decrease in size in conditions known to reduce luminal added directly to small intestine in vitro, glucagon has been nutrition.25 Thus explanations other than the changed balance shown to cause Cl- secretion across mouse jejunum,33 and between crypt and villous enterocytes must be sought. when acutely injected causes jejunal secretion in human The pd across the luminal membranes of the starved subjects in vivo." enterocytes in both the jejunum and the ileum becomes At present all the experimental evidence linking glucagon hyperpolarised afteronly 24hours offasting and is maintained with the hypersecretion of starvation is only attributable to for at least 72 hours of starvation.26 This membrane hyper- the small intestine. It is not known whether glucagon is also polarisation creates an increased driving force on the involved in the hypersecretion induced in the colonocytes. intracellular Cl- ions when the Cl- channels of the luminal Even if further experimental studies show that glucagon is membrane are opened byneurotransmitters or secretagogues. implicated in colonic hypersecretion an explanation has to be Thus the hyperpolarisation of the starved enterocyte's sought for the finding that starved colonocytes respond to Ca- luminal membrane can account for the hypersecretory mediated secretagogues by hypersecretion but not to those activity shown by the small intestine in the food deprived