Gut Motility, Secretion and Gut Hormones

Mice born without one chloride channel, ANO1, have normal-looking pacemaker cells in the gut but no electrical slow waves in the gut muscles tested. Your gut contracts on a clock it keeps itself, under nerves in its own wall, and your stomach makes acid near pH 1 beside cells at pH 7. The first look at the gut, Pavlov’s sham-fed dogs and the first hormone table are in digestive systems; here they get cells, receptors and numbers.

Cajal’s cells keep the beat; nerves decide how hard

Record from a gut smooth muscle cell and you see slow waves: rhythmic depolarizations lasting seconds. In the human stomach, pacemaker ICC generate slow waves at about 3 per minute; the waves start high on the corpus and spread toward the pylorus. Rates differ by region and species; stretching mouse antral muscle in vitro has been reported to raise the rate from about 2.5 to 3.6 per minute.

A slow wave is not an action potential. It sets when the muscle can contract; a contraction follows only when a wave carries the muscle past threshold for Ca2+ entry, often with spikes on its plateau. Think of a metronome and a pianist: the metronome fixes the beat, the pianist decides how loud each note is. The rhythm is intrinsic to the muscle tissue, but how strongly each wave couples to contraction is set by excitatory and inhibitory neurons and by hormones. Force generation itself belongs to smooth muscle.

The beat comes from interstitial cells of Cajal (ICC), branched cells in networks between and within the muscle layers. Isolated smooth muscle cells do not typically show this rhythm; isolated ICC do. In paired recordings from mouse gut muscle, slow waves were larger in myenteric ICC than in neighbouring smooth muscle, and they reached the muscle passively, shrinking as they spread, so smooth muscle cannot regenerate them.

The modern experiment. Two groups showed that ICC are necessary: small-intestinal muscle from Kit-mutant W/Wv mice, with few myenteric ICC, lacked the slow waves always present in wild-type siblings (Ward and colleagues, 1994), and Huizinga and colleagues (1995) reported that W mutants lacked both the myenteric ICC network and pacemaker activity. Neither said what makes an ICC oscillate. Hwang and colleagues in the Sanders and Ward laboratory (2009) tested ANO1 (TMEM16A), a Ca2+-activated Cl- channel. Mice homozygous for a null Tmem16a allele failed to develop slow waves by birth, and gut muscle from them was studied in organ culture, yet their ICC networks looked structurally normal with Kit antibodies. The authors also reported ANO1 in ICC of macaque and human gut muscle. This rules in a Ca2+-activated Cl- current in ICC as necessary for the rhythm, separately from building the network. It does not show that ANO1 is the whole pacemaker (periodic Ca2+ release from stores is the proposed upstream trigger), and it describes neonatal mice, not adults or any human disorder.

The gut wall has its own nervous system

The enteric nervous system is the network of neurons in the gut wall, in two plexuses. The myenteric (Auerbach’s) plexus, between the circular and longitudinal muscle layers, mainly controls movement; the submucosal (Meissner’s) plexus mainly controls secretion and local blood flow. Each holds sensory neurons, interneurons and motor neurons, enough to run reflexes without the brain; extrinsic nerves adjust these circuits, and some functions, such as normal defecation, also need spinal pathways (the hypothalamus and autonomic system; vagal sensory neurons).

The classic experiment. In 1899 William Bayliss and Ernest Starling, three years before their secretin work (signals and distance), recorded the movements of small-intestinal loops in anesthetized dogs. If coordinated movement were organized by the brain and spinal cord, cutting the extrinsic nerves, the vagi and splanchnics, should have abolished it. Instead, a local stimulus such as pinching the wall or a bolus distending it produced contraction on the oral side and relaxation on the anal side, so contents were driven anally. The response persisted after the extrinsic nerves were cut and was lost when nervous conduction within the wall was blocked. This ruled in a polarized reflex carried by nerves within the gut wall and ruled out a need for the central nervous system. It did not show which neurons or transmitters carry it, or where the rhythm comes from. Their summary became the “law of the intestine,” and a later test bounds it: in isolated guinea-pig ileum, stroking the mucosa evoked a peristaltic wave that travelled 13 cm with no fluid in the lumen, yet those authors suggested the preparation does not conform to the 1899 law (Spencer, Walsh and Smith, 1999). The law describes some preparations, not every gut.

The same intrinsic wiring ends the tract. The rectum stores feces. When it is distended, an enteric reflex relaxes the internal anal sphincter, which is smooth muscle and not under voluntary control. The external anal sphincter is skeletal muscle supplied by the pudendal nerve; it keeps up a resting contraction and can be squeezed voluntarily to hold defecation off; sacral parasympathetic outflow through the pelvic nerves strengthens colonic and rectal contraction.

Mouth to stomach: wet, swallow, store, grind, meter

An adult makes roughly 1 to 1.5 L of saliva a day, in two steps. Acinar cells secrete a fluid close to plasma in composition; the ducts, poorly permeable to water, then reabsorb Na+ and Cl- and secrete K+ and HCO3-, so the saliva ends hypotonic. At high flow the ducts have less time per volume and the saliva comes out closer to plasma tonicity. Both autonomic divisions stimulate secretion, with parasympathetic input dominant.

Swallowing starts voluntarily; the pharyngeal phase is a reflex patterned in the brainstem, and esophageal peristalsis carries the bolus down while vagal pathways activate inhibitory enteric neurons that relax the lower esophageal sphincter. The proximal stomach then relaxes as food arrives (receptive relaxation) and stays relaxed as it fills (accommodation), so volume rises with little rise in pressure. The antrum grinds: contractions push chyme against a pylorus that is closing, and most of it is thrown back, until particles are small enough to pass.

In one review’s example for healthy humans, water is half gone in about 10 minutes and a high-calorie liquid only at about 2 hours, with real times depending on the meal’s volume and calories, and the stomach delivers roughly 1 to 4 kcal per minute to the duodenum. Digestible solids leave only after a lag. Emptying a solid meal takes hours: in one multicentre scintigraphy study of a standard meal, a median of 24 percent remained at 2 hours and about 1 percent at 4 hours. After a meal, CCK and GLP-1 slow emptying, partly through vagal reflexes; in fasting, motilin speeds it, and ghrelin given as a drug does too, although whether the body’s own ghrelin does is less clear.

Fasting brings the migrating motor complex (MMC), a cycle interrupted by feeding. Phase I lasts about 45 to 60 minutes, with slow waves but few or no contractions; phase II has irregular contractions; phase III is a 5 to 15 minute front of strong contractions that sweeps toward the pylorus and down the small bowel. Each phase III contraction still rides on a slow wave, so the front contracts at the local slow-wave rate (about 3 per minute in the antrum); enteric neurons, with vagal and hormonal input, decide when the front starts and how it moves. Isolated mouse small intestine can produce migrating complexes without slow waves, but that has not been shown for the human stomach. In ten fasting volunteers, plasma motilin peaked about every 110 minutes, and in humans intravenous motilin or ghrelin can start antral phase III.

Acid from a pump, protection from a gel

At rest the parietal cell keeps its H,K-ATPase in cytoplasmic tubulovesicles. Stimulation moves the pumps into the membrane of the secretory canaliculus, a deep fold of the apical surface. Each cycle exchanges cytoplasmic H+ for luminal K+ using ATP, electroneutrally; K+ returns to the lumen through apical channels and Cl- exits with it, so HCl accumulates. The H+ comes from carbonic anhydrase, and the matching HCO3- leaves across the basolateral membrane in exchange for Cl-, the alkaline tide of acid-base balance. The canaliculus of a stimulated cell reaches about pH 1.

Acetylcholine acts on Gq-coupled M3 receptors and gastrin on Gq-coupled CCK2 receptors, both raising Ca2+; histamine acts on Gs-coupled H2 receptors and raises cAMP (GPCRs and G proteins; cAMP and PKA). The H2 receptor is James Black’s story (drugs against a part). In humans, gastrin probably acts mainly by releasing histamine from enterochromaffin-like (ECL) cells beside the parietal cells. In 14 healthy fasted volunteers (11 men, 3 women) given synthetic gastrin, H2 block completely blocked the acid response, and muscarinic block cut the response to the lower gastrin dose by about 73 percent; the authors concluded that histamine and cholinergic mechanisms are both essential and argued against a significant direct effect of gastrin alone (Zimmerhackl and colleagues, 1993).

The phases follow the meal. The cephalic phase, from sight, smell and chewing, is vagal (digestive systems). In the gastric phase, digested peptides and amino acids act directly on antral G cells, which sense them through receptors such as the calcium-sensing receptor, and distension excites vagal and local reflexes that release gastrin through GRP-containing neurons. Acid in the antrum releases somatostatin from D cells, which restrains G cells, a feedback inside the gastric phase. The intestinal phase is mostly a brake: in the duodenum secretin, and CCK acting at CCK1 receptors (probably via somatostatin), inhibit acid. The lining survives because a mucus gel holds secreted bicarbonate against the surface, keeping it far less acidic than the lumen; prostaglandins support that secretion, and tight junctions limit H+ back-diffusion (epithelia).

Worked problem: what does it cost to make pH 1?

In this illustrative example, stimulated parietal cells secrete 1.0 L of juice at pH 1.0. Assume cytosolic pH 7.0, body temperature 310 K, R = 8.314 J per mol per K, and electroneutral exchange, so the membrane potential drops out.

1. Luminal [H+] = 10^-1.0 = 0.10 mol per L, so 1.0 L holds 0.10 mol H+. At 1 H+ per ATP, the most the energy budget allows at pH 1, that takes at least 0.10 mol ATP.
2. The concentration ratio is 10^-1 / 10^-7 = 10^6.
3. Free energy per mol H+:

delta-G = 2.303 x R x T x (pH cytosol - pH lumen)

2.303 x 8.314 x 310 = 5,936 J per mol per pH unit; x 6 = 35,600 J, about 35.6 kJ per mol H+. Ignoring the uphill K+ step makes this a minimum. 4. ATP hydrolysis in cells releases roughly 50 to 60 kJ per mol (take 55). Two H+ at pH 1 would cost 71.2 kJ, too much; at luminal pH 6.1 they cost 2 x 5.94 x 0.9 = 10.7 kJ. Two per ATP stays affordable while 2 x 5.94 x delta-pH is at most 55, so delta-pH is at most 4.6, a luminal pH of about 2.4.

Whether the pump ever moves 2 H+ per ATP is debated: one vesicle study reported a fall from 2 to 1 as luminal pH dropped, another found 1 per ATP at pH 6.1 to 6.9, and a 2019 crystal structure found a single bound K+, which favours 1 H+ per ATP. The budget shows why 2 per ATP cannot work at pH 1, not which ratio the pump uses at milder pH.

Luminal pH (canaliculus) 1.0 (the lesson’s stimulated cell: about 1) Cytosolic pH 7.0 (the lesson’s example: 7.0)

Protons moved per ATP

1 H⁺ per ATP 2 H⁺ per ATP

Energy released by one ATP 55 kJ per mol (illustrative; the lesson takes 55)

pH difference, cytosol − lumen

Free energy per mol H⁺ = 5.936 × pH difference

Work per cycle = n × free energy per H⁺

Energy released by one ATP

Luminal pH at which 2 H⁺ per ATP stops being affordable = pH cytosol − E ÷ (2 × 5.936)

Illustrative model, not a measurement. In this sketch (made up for the sketch unless the lesson gives it): each cycle spends one ATP; the cost is 2.303 × 8.314 × 310 × pH difference, at 310 K with electroneutral exchange so the membrane potential drops out, which is the lesson’s rule; the K+ step is ignored, so every cost shown is a minimum; the starting values, cytosolic pH 7.0, luminal pH 1.0 and 55 kJ per mol per ATP, are the lesson’s illustrative worked-problem values (ATP hydrolysis in cells releases roughly 50 to 60 kJ per mol); the slider ranges for pH are made up (the energy range is the lesson’s roughly 50 to 60); the sketch lets you pick 1 or 2 H+ per ATP at any luminal pH, although the lesson’s sources measured 2 at luminal pH 6.1 and 1 below pH 3 and record the ratio, not this reason.

Rule to keep: each pH unit of uphill gradient costs about 5.94 kJ per mol of H+ at 310 K, and a cycle pays n times that out of one ATP, so the more protons per cycle, the smaller the pH difference one ATP can pay for; at the lesson’s 55 kJ per mol and cytosol pH 7.0, two protons per ATP stop being affordable below a luminal pH of about 2.4.

Pathophysiology: reading gastrin against acid. In CCK-B receptor knockout mice, basal gastric pH rose from 3.2 to 5.2 and plasma gastrin was about 10-fold higher: the mutants had fewer parietal and ECL cells, fewer antral somatostatin cells and more gastrin cells: less acid, less somatostatin brake, more gastrin. Atrophic gastritis, with parietal cells lost, gives the same pattern in humans. A gastrinoma, a neuroendocrine tumour releasing gastrin that acid cannot suppress, reverses one arrow: gastrin high, basal acid output and parietal cell mass raised, and gastric and duodenal pH low enough to inactivate pancreatic lipase. Gastrin is high in both while acid moves in opposite directions, so reading the two together points to feedback or to an autonomous source, although diagnosing a gastrinoma needs further tests.

Neutralize, emulsify, and the hormone ledger

Acid entering the duodenum releases secretin, which drives pancreatic duct cells to secrete bicarbonate-rich fluid: in humans the juice reaches 140 to 150 mM HCO3- under secretin, about 6- to 7-fold over plasma, by a route that depends on CFTR. Mouse and rat juice reach only about 50 and 70 mM, so their numbers do not transfer directly to humans, although guinea pigs, also rodents, reach human-like values and mice remain useful for molecular work. CCK and acetylcholine drive acinar enzyme secretion, CCK contracts the gallbladder, and secretin also drives bicarbonate from bile-duct cells; bile and absorption are in absorption and the liver.

Hormone	Cell and site	Stimulus	One action here
Gastrin	G cells, antrum	Peptides and amino acids, vagal input, distension	Acid, mainly through ECL histamine
Secretin	S cells, proximal small intestine	Acid in the duodenum	Duct fluid and bicarbonate
CCK	I cells, duodenum	Fatty acids and digested protein	Gallbladder contraction
GIP	K cells, upper small intestine	Glucose and fat	Glucose-dependent insulin release
GLP-1	L cells, mostly ileum and colon	Nutrients in the gut	Slows gastric emptying
Motilin	Mo cells, upper small intestine	Cyclic release in fasting	Can induce antral MMC phase III
Ghrelin	P/D1 cells in humans (X/A-like in rodents), corpus	Fasting; suppressed by a meal	Speeds emptying when given as a drug

The incretin actions of GIP and GLP-1 belong to insulin, glucagon and blood glucose, and these cells as sensors to gut-brain sensing.

Design problem

Gastrin-deficient mice made by gene targeting (Friis-Hansen and colleagues, 1998) have no basal acid secretion, and acid cannot be induced by histamine, carbachol or gastrin; six days of gastrin partly restores it. A classmate says the histamine result is impossible, because histamine acts directly on parietal H2 receptors. What explains the data, and what experiments would test your explanation?

One way to do it

Propose that gastrin is trophic as well as acute: without it from conception the acid-secreting unit is not built or maintained, leaving fewer or immature parietal cells, scarce pumps or H2 receptors, and little for histamine to act on.

Tests, against wild-type littermates: count parietal cells and measure H,K-ATPase, H2 receptor and ECL histidine decarboxylase expression. Track recovery over days of gastrin replacement against a vehicle control. Separate development from adult maintenance with an inducible deletion in adults. Converge with the CCK-B receptor knockout, which should share the phenotype if it runs through that receptor.

Limits: these are mice with gene loss from conception, and in humans the acute gastrin route is mostly ECL histamine, so the trophic result extends to people only by inference.