Adenosine triphosphate is co-secreted with glucagon-like peptide-1 to modulate intestinal enterocytes and afferent neurons

Abstract

Enteroendocrine cells are specialised sensory cells located in the intestinal epithelium and generate signals in response to food ingestion. Whilst traditionally considered hormone-producing cells, there is evidence that they also initiate activity in the afferent vagus nerve and thereby signal directly to the brainstem. We investigate whether enteroendocrine L-cells, well known for their production of the incretin hormone glucagon-like peptide-1 (GLP-1), also release other neuro-transmitters/modulators. We demonstrate regulated ATP release by ATP measurements in cell supernatants and by using sniffer patches that generate electrical currents upon ATP exposure. Employing purinergic receptor antagonists, we demonstrate that evoked ATP release from L-cells triggers electrical responses in neighbouring enterocytes through P2Y₂ and nodose ganglion neurones in co-cultures through P2X_2/3-receptors. We conclude that L-cells co-secrete ATP together with GLP-1 and PYY, and that ATP acts as an additional signal triggering vagal activation and potentially synergising with the actions of locally elevated peptide hormone concentrations.

Fig. 1

Quinacrine fluorescence and VNUT staining of GLP-1 secreting cells. a GLUTag cell under brightfield illumination and b 480 nm fluorescence excitation after incubation with 5 µM quinacrine. c RFP fluorescence of an L-cell from a Glu-Cre x Rosa26tdRFP reporter mouse and d 480 nm fluorescence of the cells shown in c after incubation with 5 µM quinacrine. e Example image from total internal reflection fluorescence (TIRF) microscopy of GLUTag cells after incubation with quinacrine (5 µM). Images were collected every 50 ms for 30 s. Inset shows zoomed view of vesicle in the white box, the yellow circle is the mask within which intensities were measured. f Representative profiles of quinacrine intensity for the image in e during transient fluorescence increase events (black) and failed secretion (grey). Transient increases occurred in 21 out of 35 cells examined with a median frequency of 14.9 spikes mm⁻² s⁻¹ in responsive cells (IQR: 5.0–27.3 spikes mm⁻² s⁻¹, n = 24 movies from seven independent experiments). g Intensity profile of the vesicle labelled * in e and f, with images showing vesicle signal at various time points during the spike: i = prior to spike, ii = peak of the spike, iii = during fluorescence dissipation, iv = return to initial signal. h GLUTag cells, i mouse ileum (left) or colon (right) and j human colonic cultures immunostained for the vesicular nucleotide transporter (VNUT, green, top panel) and GLP-1 or PYY (magenta, middle panel). Bottom panel represents merged images. Scale bars represent 10 μm, apart from g (3 μm)

Fig. 2

Measurements of ATP release from GLP-1 secreting cells. a ATP levels measured in GLUTag cell supernatants after stimulation for 10 min with AngII (1 µM) or forskolin (10 µM) + IBMX (10 µM) + 10 mM glucose (F/I/10 G); n = 15 wells, data obtained from five independent experiments. b ATP levels in supernatants of GLUTag cells transiently transfected with GqDREADD, incubated for 10 min with or without CNO (10 μM); n = 14–20 wells from four independent experiments. c ATP levels in supernatants from primary colonic cultures derived from GLU-Cre/GqDREADD mice, treated with CNO (10 μM), and forskolin/IBMX (F/I, 10 μM each) as indicated. d GLP-1 secretion from primary colonic cultures used in c, expressed as percentage total GLP-1 content in the cells. n = 10 wells for conditions without CNO and n = 11 wells for conditions with CNO from three independent experiments for c and d. e ATP levels measured in primary human colonic cultures after stimulation for 60 min with AngII (1 µM); n = 9, data obtained from four independent experiments. All experiments were performed in the presence of 100 μM POM-1 (GLUTag, a, b) or 100 μM POM-1 and 100 μM ARL-67156 (primary cultures, c–e). Individual points represent measurements from single wells and lines on the graph represent median ± interquartile range for a–c and mean ± SEM for d and e. *p < 0.05, ***p < 0.001 by one-way ANOVA with Dunnett’s multiple comparisons test. ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 by Kruskal–Wallis ANOVA with multiple comparisons test. ^$p < 0.05 by unpaired t-test

Fig. 3

ATP secretion measured using sniffer patches. Recordings from sniffer patches excised from HEK293 cells transfected with P2X₂ receptors. a Example trace of a sniffer patch placed adjacent to a GLUTag cell that was stimulated by forskolin/IBMX/glucose (10 μM/10 μM/10 mM; F/I/10G). 100 μM ATP was applied at the end of the recording to activate the sniffer patch directly. b Peak current amplitudes in response to forskolin/IBMX/glucose (blue triangles) and 30 mM K⁺ (blue circles) from multiple cells recorded as in a; sniffer patches were placed either adjacent to the perfusion outlet (control—triangles and circles overlaid) or nearby a GLUTag cell. c Example trace of a recording from a sniffer patch placed adjacent to a GLUTag cell transiently transfected with Gq-DREADD, with addition of ATP, CNO and AngII as indicated by the bars. d Summary data recorded as in c in response to CNO from GLUTag cells either untransfected (Control) or Gq-DREADD-transfected. e Example trace of a sniffer patch placed adjacent to a primary colonic L-cell identified by cell-specific Venus expression driven by the proglucagon promoter. AngII and ATP were perfused as indicated. f Summary data from sniffer patches in response to various stimuli of endogenous receptors of L-cells, when placed either adjacent to the perfusion outlet (control—symbols overlaid) or in close proximity to a primary L-cell. Various symbols represent the stimulus applied. Y-scale bars represent 0.5 nA and X-scale bars represent 20 ms in a, c, e

Fig. 4

Ussing chamber recordings from murine colonic mucosa. a Haematoxylin and eosin stained colonic epithelium following removal of the serosa and outer muscular layer (left) or additional removal of the submucosal and muscularis mucosal layer (right); scale bars represent 50 μm. Distribution of b transepithelial resistance (TER) and c peak I_sc amplitude induced by bilateral application of forskolin (10 µM) at the end of each experiment, colour coded according to the test drug added earlier in the experiment (coefficient of variation, CV TER = 15%, CV basal I_sc = 28.2%, CV forskolin peak I_sc = 17.5%). Individual data points represent individual epithelial preparations and lines represent mean ± SEM (n = 25). d I_sc recordings during addition of ATP (50 μM) to the basolateral compartment. Grey traces represent individual experiments and the black trace represents the mean of n = 3 recordings. Mean ± SEM ΔI_sc = 29.5 ± 6 µA cm⁻². Y-scale and X-scale bar represent 5 μA and 200 ms, respectively. e–h Superimposed I_sc recording traces from individual recordings during basolateral application of AngII (1 µM). Purinergic receptor inhibitors were added to the basolateral compartment in f–h for 10–12 min prior to AngII addition, as indicated above the traces. Grey traces indicate individual recordings and black traces represent the mean response. Y-scale and X-scale bar represent 10 μA and 2 min, respectively. i, j Amplitudes of the positive peak I_sc (i), and the I_sc depression (j), for the recordings depicted in e–h. Data points represent individual epithelial preparations and lines represent mean ± SEM. Control (Ctrl, n = 5), PPADS (100 μM, n = 4), Suramin (100 μM, Sur, n = 4) and AR-C 118925XX (5 μM, AR-C, n = 6). Statistical analysis performed using one-way ANOVA and Dunnett’s multiple comparisons test, **p < 0.01

Fig. 5

Ca²⁺ imaging of nodose (ND) neuron and Gq-DREADD-transfected GLUTag co-cultures. a, c Images of cells in our co-culture system: brightfield (top), fluorescence at 550 nm excitation (middle), and fluorescence at 488 nm excitation (bottom). Scale bars represent 50 µm. b, d Intracellular Ca²⁺ levels, represented as the ratio of Fura-2 fluorescence at 340 and 380 nm excitation, of the cells shown in a and c: mCherry-positive Gq-DREADD transfected GLUTag cell (top trace) and YFP-positive ND neuron (lower trace) identified using a NeuroD1-Cre/YFP reporter mouse. Drugs were applied as indicated above the traces. e Areas under the curve of the Fura-2 ratio change of ND neurons in response to CNO (10 µM) with and without PPADS (100 µM) pre-treatment. Values are expressed relative to the area of the response induced by ATP (100 µM) in each cell. Only ND neurons where a Ca²⁺ response could be evoked by CNO after wash out of the PPADS were included. Individual data points represent each ND neuron (n = 13). Statistical analysis performed using a paired t-test, *p < 0.05. f Percentage block of CNO responses in ND neurons by 100 µM PPADS, derived from the data shown in e. Individual data points represent each ND neuron and lines represent mean ± SEM (n = 13). Statistical analysis performed using one-sample t-test, ***p < 0.001. g Fura-2 imaging of ND neuron cultured with primary L-cells expressing Gq-DREADD under the control of the proglucagon promoter. Inset image to left is a brightfield image of the co-culture system. Drugs were applied as indicated above the traces. h Amplitude of Fura-2 ratio changes of ND neurons in response to CNO (10 µM) with and without PPADS (100 µM) pre-treatment. Values are expressed relative to the max response elicited by ATP (100 µM) in each cell. Individual data points represent each ND neuron (n = 7). Statistical analysis performed using a paired t-test, **p < 0.01. i Percentage block of CNO responses in ND neurons by 100 µM PPADs, derived from the data shown in h. Individual data points represent each ND neuron and lines represent mean ± SEM (n = 7). Statistical analysis performed using one-sample t-test, ***p < 0.001

Fig. 6

Expression analysis and pharmacological characterization of ATP receptors in nodose (ND) ganglion neurons. a P2rx subunit expression levels (2^−ΔCt values) of ND neurons from intact ganglia (black circles), acutely dissociated neurons (black squares), and after 3 days in vitro cultures (black triangles). Samples for each type of preparation were prepared from ND ganglia pooled from 2 to 3 mice, repeated three independent times. Individual data points represent independent preparations and lines represent mean ± SEM (n = 3). b Heat map of P2rx subunit expression from individually picked ND neurons. Each column represents a single ND neuron. Range indicator for heat map on left. Sample GLP1R negative (c) and GLP1R-positive (d) NeuroD1-EYFP neuron immunostained for P2X₃ (Alomone P2X₃ antibody APR-016 in c, Neuromics P2X₃ antibody GP10108 in d) and GLP1R. Scale bars represent 20 µm. e Scatterplot of % block of exogenous ATP (100 µM) application by 100 µM PPADs (grey filled circles, n = 27), 100 µM suramin (grey filled squares, n = 22), 30 µM AF-353 (grey filled triangles, n = 48) or 1 µM Ro51 (half grey filled triangles, n = 72) as measured by Fura2 Ca²⁺ imaging. Individual data points represent individual neurons and lines represent median ± interquartile range. f Intracellular Ca²⁺ levels, as the measured Fura2 signal (340/380 nm ratio), of an mCherry-positive Gq-DREADD transfected GLUTag cell (top trace) and GLP1R-positive ND neuron (lower trace) identified using a GLP1R-Cre/GCaMP3 reporter mouse. Drugs were applied as indicated above the traces. g Amplitude of Fura-2 ratio changes of ND neurons in response to CNO (10 µM) with and without Ro51 (1 µM) pre-treatment. Values are expressed relative to the max response elicited by ATP (100 µM) in each cell. Individual data points represent each ND neuron (n = 14). Dark purple symbols represent identified GLP1R-positive ND neurons. Statistical analysis performed using a paired t-test, ** p < 0.01. h Percentage block of CNO responses in ND neurons by 1 µM Ro51, derived from the data shown in g. Individual data points represent each ND neuron and lines represent mean ± SEM (n = 14). Statistical analysis performed using one-sample t-test, *** p < 0.001

Fig. 7

Schematic model of purinergic signalling from L-cells to local targets. ATP released together with GLP-1 and PYY from L-cells, exemplified here by angiotensin receptor stimulation, can activate purinergic receptors located on neighbouring enterocytes, potentially modulating fluid secretion by opening of Ca²⁺-sensitive K⁺ channels. ATP released from L-cells also targets local nerve terminals of vagal afferent neurons deriving from the nodose ganglion (brown cell) which express purinergic P2X₃ receptors, providing ATP neurotransmitter mediated signalling pathway from the intestinal epithelium to the brainstem