Sensing the Gut: Enteroendocrine Cells and Nutrient Signals

When light silences one kind of cell in the lining of a mouse’s duodenum, the animal no longer reliably prefers sugar to a calorie-free sweetener. The cell is not a neuron, yet it reaches the nervous system. This lesson starts in the epithelium, where vagal sensory neurons leave off, and asks what the cell detects, how it signals, and what that does to preference. Everything is mouse work unless a sentence says otherwise. Smell and taste owns the tongue, so taste here is only the control.

The sensory cell that is not a neuron

Enterochromaffin (EC) cells make more than 90 percent of the body’s serotonin yet are under 1 percent of intestinal epithelial cells. Both figures are background statements in Bellono and colleagues (Cell, 2017), a study co-led by the laboratories of David Julius and Holly Ingraham. The rarity had blocked single-cell work. The authors used mouse intestinal organoids (cultured mini-intestines) in which EC cells carried a genetic tag, and checked the findings in native tissue.

The EC cells expressed chemosensory receptors: TRPA1 acted as an irritant receptor, and the odorant receptor Olfr558 as a sensor of a microbial metabolite (Olfr558 was also tested in HEK293 cells). The cells were electrically excitable. Fibres that carry 5-HT3 receptors made what the authors called synaptic-like contacts with EC cells. In an ex vivo colon preparation, epithelial application of norepinephrine or isovalerate evoked firing in single mucosal afferent fibres, and the 5-HT3 antagonist alosetron blocked it. A trap sits here. The recorded fibres were pelvic (spinal) afferents, and the paper did not show any fibre to be vagal, so “EC cells synapse onto the vagus” goes beyond it. The in vivo test came later: Bayrer and colleagues (Nature, 2023) reported in mice that enhancing EC cell function was sufficient to cause hypersensitivity to gut distension, that EC cell activity was necessary for the sensitising action of isovalerate (a bacterial short-chain fatty acid), and that the EC-to-mucosal-afferent circuit appeared tonically engaged in females.

The point is the cell type: a non-neuronal receptor cell with chemosensory receptors, excitability and an output onto a nerve, as taste cells have. Whether that output is read as a rate, a pattern or a labelled line is the question of sensory coding.

Neuropods: a synapse in the epithelium, and how it was shown

Before 2018 the prevailing view, as Kaelberer and colleagues put it, was that the brain senses gut stimuli only through passive hormone release. Bohorquez and colleagues (J. Clin. Invest., 2015) showed that a cytoplasmic process on enteroendocrine cells, earlier called a neuropod, puts the cells in direct contact with mucosal nerves. Kaelberer and colleagues (Science, 2018) asked whether the contact was a synapse, and called synaptically connected cells neuropod cells. The chain of evidence has five steps, and each step answers a different doubt.

1. Connectivity. Modified rabies virus given by enema into the colon of PYY-Cre pups, or by oral gavage into CCK-Cre pups, infected enteroendocrine cells and labelled nerves. In the colon experiment some labelling reached vagal nodose neurons (2 of 5 mice with infected colon nerves, with the pseudotyped virus).
2. Transmitter machinery. Enteroendocrine cells expressed VGLUT1 transcripts, which fits glutamate as the transmitter.
3. Speed. In cocultures of enteroendocrine cells and nodose neurons (n = 9 pairs), light-activated cells evoked excitatory postsynaptic currents within 60 to 800 ms.
4. Pharmacology. Kynurenic acid (3 mM), an ionotropic glutamate-receptor blocker, abolished the currents, and they recovered on washout (n = 4 neurons). Recovery matters because it separates blocked receptors from damaged cells.
5. Necessity and sufficiency in vivo. In anaesthetised mice with a recording from the cervical vagus, perfusing sucrose into the intestine raised vagal firing within seconds (peaking about a minute or more later), and phloridzin (an SGLT1 blocker) abolished that response. Halorhodopsin silencing of CCK-labelled cells abolished the sucrose response (necessity), and channelrhodopsin activation of the same cells raised vagal firing (sufficiency), so the authors concluded the cells are necessary and sufficient to transduce a glucose stimulus.

Three limits belong with that list. Only 18.9 percent of CCK cells contacted Pgp9.5-positive sensory fibres in fixed tissue (n = 3 mice), so it is wrong to say that enteroendocrine cells in general synapse onto the vagus. The three timescales are different measurements: coculture latency in milliseconds, in vivo firing “within seconds”, and learned preference over 24 to 48 hours. And optogenetic silencing of a Cre-labelled population is acute but not pure by type, because a Cre line marks cells that expressed the driver at some time. This does not make hormones irrelevant. It means that a synaptic route exists beside them.

Sugar without taste: SGLT1 and learned preference

Tan and colleagues (Nature, 2020), from Charles Zuker’s laboratory, offered mice sugar against acesulfame K, a non-nutritive sweetener at a concentration equally attractive at first. Naive mice drank from both bottles at a similar rate, preference shifted within 24 hours, and by 48 hours they drank almost exclusively sugar. Mice lacking T1R2/T1R3 (n = 5) or TRPM5 (n = 7), which cannot taste sweetness, still developed the preference. A taste explanation is therefore not needed for learned preference in this assay. For how the tongue codes sweet, see smell and taste.

Vagal ganglion imaging found neurons that responded to sugar delivered to the gut but not to sweeteners, and intestinal phlorizin caused a dramatic loss of the glucose responses. Silencing transmitter release in sugar-activated neurons of the caudal nucleus of the solitary tract (labelled with the activity-dependent driver TRAP2; n = 6) abolished the development of sugar preference without removing innate attraction to sugar or sweeteners. Chemogenetic activation of Penk-Cre neurons in the same region created a preference for a previously less-preferred stimulus. That separates what the animal likes at first sip from what its gut teaches it. Tan and colleagues left the intestinal cell type open.

Buchanan and colleagues (Nature Neuroscience, 2022) then tied the CCK-labelled duodenal neuropod cells to the choice, and reported that sugar stimulated glutamatergic and sweetener purinergic transmission to the vagus. Of 132 CCK-GFP small-intestinal cells (N = 3 mice), 60.1 percent expressed only Slc5a1 (SGLT1), 19.6 percent both SGLT1 and Tas1r3 (T1R3), 1.2 percent only T1R3, and 19.1 percent neither. The transducer for sugar is epithelial SGLT1, not a sweet receptor. The only human data are glutamate release from organoids of one donor, which the authors say suggests conservation: that is not human behaviour.

A reward result is easy to misuse. Han and colleagues (Cell, 2018) activated gut-innervating vagal sensory neurons with light in adult male mice. Activation on the right side, but not the left, sustained self-stimulation, flavour and place preference and dopamine release, relayed through glutamatergic neurons of the dorsolateral parabrachial region to dopamine cells in the substantia nigra. The stimulus was light, not a nutrient. The correct reading is that this route can sustain reward. For central reward circuits see motivation and reward.

Fat has its own receptors and its own vagal line

Li and colleagues (Nature, 2022) tracked preference for 1.5 percent Intralipid against 3 mM acesulfame K with lick counters over 48 hours. TRPM5-mutant mice, blind to the taste of fat, still developed a strong preference. Vagal imaging revealed two parallel systems: a general nutrient sensor responding to sugar, fat and amino acids, and one activated only by fat, a TRPA1-expressing cluster. Silencing VIP vagal neurons markedly impaired nutrient preference and left immediate attraction to sugar and fat unaffected.

The knockout table reads cleanly. In SGLT1-knockout mice no vagal neurons responded to intestinal sugar, and fat responses were unaffected. Fat responses fell substantially only in GPR40/GPR120 double and CD36/GPR40/GPR120 triple knockouts. A germline deletion like these is neither acute nor restricted to one cell type, so it shows the receptor is needed in the animal, not in which cell.

Worked problem: sugar versus sweetener

Buchanan and colleagues used mice with a light-driven silencer (halorhodopsin) in CCK-labelled neuropod cells, an intraluminal fibre and a one-hour choice between 300 mM sucrose and 15 mM sucralose. Controls under the same light chose sucrose 90.8 percent of the time (N = 5), silenced mice 58.9 percent (N = 8), and mice with duodenal glutamate receptors blocked locally 44.0 percent (N = 4).

fraction lost = (control - silenced) / (control - 50)

(a) Against 50 percent chance, the preference advantage is 90.8 - 50 = 40.8 points in controls and 58.9 - 50 = 8.9 points silenced. (b) Lost: 90.8 - 58.9 = 31.9 points, and 31.9 / 40.8 = 0.78, about 78 percent. (c) The glutamate-block value, 44.0, is 6.0 points below chance. Do not round it to indifference, and do not compare it with 90.8: its matching vehicle control was 82.4 percent in a separate cohort. (d) Guard: locomotion and 24-hour food and water intake were unchanged, so this is not malaise. With N of 4 to 8, the percentages carry wide uncertainty that the arithmetic hides.

Light, controls (N = 5) Neuropod cells silenced (N = 8) Duodenal glutamate receptors blocked (N = 4)

Control preference, C 90.8 % sucrose Manipulated preference, M 58.9 % sucrose

Control advantage, C − 50

Manipulated advantage, M − 50

Points lost, C − M

Fraction lost, (C − M) / (C − 50)

In this sketch: mice, a one-hour choice between 300 mM sucrose and 15 mM sucralose, preference scored as the percentage of choices for sucrose with 50 percent as chance, and the lesson’s rule applied to whatever values are set. Any value other than the lesson’s is labelled as set in this sketch.

Rule to keep: measure preference as the advantage over 50 percent chance, then divide the points lost by the control advantage, not by the raw preference; and compare each group only with its own matching control. Small groups (N of 4 to 8) leave wide uncertainty.

A second reading uses Li and colleagues’ VIP neurons (9 ganglia): 60 of 203 responded to nutrients. (e) 60 / 203 = 29.6 percent, matching “approximately 30 percent”. (f) About 4 percent responded to fat, so, if that 4 percent is of all 203 neurons, 0.04 x 203 is about 8 neurons, an approximate count because the paper gives “about 4 percent”. (g) 8 / 60 is about 13 percent: most VIP nutrient responders are not fat responders.

GLP-1: three routes often collapsed into one

L cells release GLP-1 after nutrients (background in a rat study, Krieger and colleagues, Diabetes, 2016). The peripheral system is gut enteroendocrine cells, and the central system is brain preproglucagon (PPG) neurons; Brierley and colleagues (Nature Metabolism, 2021) concluded that the two suppress eating through independent gut-brain circuits. Brainstem PPG neurons receive vagal input predominantly from oxytocin-receptor neurons rather than GLP-1-receptor neurons, and are not necessary for eating suppression by GLP-1 receptor agonists in mice. So “gut GLP-1 to vagus to brain GLP-1 neurons” is not supported.

Vagal GLP1R neurons are a third item. Williams and colleagues (Cell, 2016), covered in vagal sensory neurons, reported that they have intraganglionic laminar endings and act as gastrointestinal mechanoreceptors; GPR65 neurons, which supply most villi, were insensitive to GLP-1 or cholecystokinin. Those authors also reported that over 95 percent of GLP1R neurons and over 97 percent of GPR65 neurons are slow C fibres. A 2025 erratum corrects only an EKG trace in Fig. 3A; the C-fibre percentages are unaffected.

For drugs, mechanism only. In rats, vagal deafferentation did not block the weight effect of a long-acting GLP-1 receptor agonist given for 14 days (about 10 percent less weight gain than controls in both groups; Secher and colleagues, J. Clin. Invest., 2014). In mice, fluorescently labelled liraglutide bound arcuate neurons after peripheral injection (Secher and colleagues, 2014). Deleting the receptor from glutamatergic (vGlut2) neurons eliminated the weight effect of liraglutide, while deleting it from vGAT neurons did not (Adams and colleagues, Diabetes, 2018). None of this is a statement about people. The one human measurement here is pharmacokinetic: after intravenous infusion of GLP-1(7-36) amide, intact peptide was 19.9 percent of the rise in immunoreactivity measured by a COOH-terminal assay (Deacon and colleagues, Diabetes, 1995; n = 8 normal subjects). Vagal GLP-1 receptors still matter physiologically: knockdown in rat nodose neurons enlarged meals, sped gastric emptying and raised postmeal glucose, with normal long-term energy balance (Krieger and colleagues, 2016).

Design problem

A reader is told a gut hormone “tells the brain through the vagus.” Design experiments that decide, for one nutrient, whether the message is hormonal or synaptic, and where the receptor that matters sits.

One way to do it

Separate transmitter from hormone first. Look for a vesicular transporter transcript in the cell, record a fast current with a measured latency in a coculture, block it with an ionotropic glutamate-receptor antagonist and show recovery on washout, and test whether silencing the cells removes the vagal response. Get both directions: silence the cells for necessity and drive them for sufficiency, and state that optogenetic silencing of a Cre-labelled population is acute but not pure by type, while a germline receptor deletion is neither acute nor cell-specific. Dissociate routes next: compare a receptor deletion restricted to one neuronal class with interruption of afferent supply to the gut, and measure the same behaviour in each. Use a malaise control, and test innate attraction separately from learned preference. Close with species: all route evidence here is rodent, and the human data are one organoid donor and one pharmacokinetic measurement, so the design must be repeated before any statement about people.