Fig. 4: PYY–vagal circuits use TLR5 and Y2R to drive changes in food intake.

Mice were fasted overnight and given an enema of either 1 μg ml⁻¹ flagellin or PBS before receiving access to standard chow ad libitum. a,b, Tlr5^fl/fl littermate control mice ate significantly less food post-flagellin enema at 20, 40 and 60 min (a), whereas Pyy^cre;Tlr5^fl/fl mice did not (b; n = 5 mice per genotype; *P < 0.05 genotype–enema–time interaction by repeated-measures ANOVA with post hoc two-tailed Tukey HSD). c,d, In wild-type mice, PBS and flagellin enemas were performed with the addition of either 10 μM TH1020 in the enema solution (n = 8 mice; c); or intraperitoneal injection of 1 μM BIIE-0246 before enema (n = 10 mice; d). TLR5 and Y2R inhibition attenuated the reduction of food intake post-flagellin enema (P < 0.05 drug–enema–time interaction by repeated-measures ANOVA with post hoc Tukey HSD; see Extended Data Fig. 9c,d for vehicle controls). i.p., intraperitoneal. e, Crunch Master (top schematic) tracks feeding microstructure by means of audio and video recording. Black ticks (bottom) indicate biting patterns of n = 7 mice post-PBS or flagellin enema. f, Left, flagellin enema significantly delayed onset of the first three bites (black: bite 1; blue: bite 2; green: bite 3; *P < 0.0001 main effect of treatment by repeated-measures ANOVA). Right, intake (spillage-corrected) was significantly decreased over 1-h test session after flagellin enema (two-tailed unpaired t-test, *P < 0.01). Violin plots show median indicated by thick line and mean indicated by thin line. g, In wild-type germ-free mice, flagellin enema decreased food intake compared to PBS enema (n = 12 mice; *P < 0.05, enema–time interaction by repeated-measures ANOVA with post hoc two-tailed Tukey HSD). Error bars represent s.e.m. h, Model for microbial pattern sensing by neuroepithelial circuits to drive behavioural change. Bacterial flagellin is detected by TLR5 in colonic PYY-labelled cells, which in turn release PYY to activate vagal neurons through Y2R. Activation of this circuit contributes to overall food intake. Graphics in h adapted with permission from ref. ³, AAAS.