288 | Nature | Vol 578 | 13 February 2020
Article
in ITT after enteric-neuron-specific deletion of Ahr was comparable to
that observed in microbiota-depleted SPF mice (Fig. 3a, b). To confirm
that Ahr activity in enteric neurons regulates the physiological output
of intestinal neural circuits independently of extrinsic gut innervation,
we used ex vivo spatiotemporal mapping of colon preparations to
record the ENS-dependent colonic migrating motor complexes^10. The
frequency and organization of colonic migrating motor complexes
were reduced in AhrEN-KO mice (Fig. 3c, d), demonstrating that the cell-
autonomous activity of Ahr in enteric neurons regulates the peristaltic
activity of the colon. The CYP1A1-mediated clearance of natural AHR
ligands^16 predicts that constitutive upregulation of Cyp1a1 in enteric
neurons would phenocopy the effect of neuron-specific deletion of
Ahr on intestinal motility, thus demonstrating the ligand-dependent
activity of AHR in enteric neurons. To test this idea, we administered
the AAV9-CaMKII-Cre vector to mice homozygous for the Rosa26LSL-Cyp1a1
allele^16 , resulting in constitutive overexpression of Cyp1a1 specifically in
enteric neurons (termed ENCyp1a1 mice) (Fig. 3e–g). As expected, ENCyp1a1
mice had an increased ITT (Fig. 3h) that is similar to that observed for
AhrEN-KO mice (Fig. 3a), which indicates that dysregulation of AHR-
ligand metabolism in enteric neurons disrupts intestinal motility. To
examine further the potential role of AHR ligands in gut motility, we
supplemented the diet of ENCyp1a1 mice for four weeks with the AHR
pro-ligand indole-3-carbinol (I3C), which generates the high-affinity
ligand indolo[3,2-b]carbazole (ICZ)^24. Exposure to I3C diet rescued—to
a large extent—the dysmotility in ENCyp1a1 mice (Fig. 3i), which further
demonstrates that the neuron-specific and ligand-dependent activa-
tion of AHR signalling regulates intestinal peristalsis. We suggest that,
similar to ICZ, other natural ligands that originate in the gut lumen
and activate AHR in epithelial and immune cells in the gut wall^16 ,^25 are
also capable of reaching nearby enteric neurons and their projections,
modulating their transcriptional profile in an AHR-dependent manner.
Finally, to provide direct evidence that AHR signalling is implicated
in the regulation of intestinal motility by microbiota, antibiotic-treated
wild-type mice—which show reduced expression of Ahr in enteric neu-
rons (Extended Data Fig. 4m–r) and a longer ITT (Fig. 3b)—were injected
with AAV vectors expressing an AHR cassette under the control of the
CaMKII promoter (AAV9-CaMKII-Ahr; AAV-AHR) or control vectors
(AAV-control), and intestinal peristalsis was evaluated four weeks later.
Because depletion of the microbiota is likely to reduce the amount of
available AHR ligands^26 ,^27 , mice were also fed with I3C-supplemented
diet for one week before the motility assay (Fig. 3j). As expected, anti-
biotic treatment of mice injected with AAV-control showed a marked
increase in ITT, but injection of AAV-AHR resulted in a significant reduc-
tion of total transit time (Fig. 3k). However, the partial rescue that we
observed suggested the presence of additional microbiota-dependent
neuromodulators, such as serotonin (which is produced by enterochro-
maffin cells and modulates intestinal peristalsis)^8. Together, our experi-
ments demonstrate that AHR signalling in enteric neurons regulates
the motor output of intestinal neural circuits.
In this study, we reveal regulatory mechanisms of enteric neurons
that link the luminal microenvironment of the gut with ENS function.
The systematic comparison of neuronal transcriptomes that represent
distinct intestinal segments and microbiota states of mice enabled us
to identify the transcription factor AHR as fulcrum of an ENS-specific
surveillance pathway that regulates intestinal peristalsis in response
to microbial colonization. Furthermore, the identification of genesRosa26LSLCyp1a1 ENCyp1a1CYP1A1overexpressionNumberof CMMCs per 2,500 sITT (min) I.v. injectionofAAV9-CaMKII-Cre5weeksehITT (min)abCYP1A1AHR ligands
AHR
XRE
Overexpression2,5002,0001,5001,000
Time (s)50002,5002,0001,5001,000
Time (s)5000(^01020) Distance (mm) 30 40 50
(^01020) Distance (mm) 30 40 50 60
Control
AhrEN-KO
j k
I.v. injectionof
AAV-AHR
I3Cdiet
High
AHR
Low
AHR
I.v. injectionof
AAV-control
ITT (min)
AAV
control
AAV
AHR
Rosa26LSLCyp1a1 ENCyp1a1
Cyp1a 1
ITT (min)
c d
fg
P = 0.0003 P = 0.0065
P = 0.000096 P = 0.0028
ControlENCyp1a1
ControlAhr ControlAhEN-KOr
EN-KO
ControlENCyp1a1ControlECyp1a1N
0
100
200
300
400
500
0
200
400
600
800
Vehicle
Antibiotics
0
100
200
300
400
500
0
100
200
300
400
500
0
2
4
6
8
10
ITT (min)
i
P = 0.6239
P = 0.0159P = 0.004
0
100
200
300
400
(^500) Puri eddiet I3Cdiet
I3Cdiet
Antibiotics
Antibiotics
P = 0.00062
Fig. 3 | AHR signalling in enteric neurons regulates intestinal peristalsis.
a, Quantification (mean ± s.d.) of ITT in control and AhrEN-KO mice (two-sided
non-parametric Mann–Whitney U-test). n = 18 control and 17 AhrEN-KO mice.
b, Quantification (mean ± s.d.) of ITT in vehicle- and antibiotic-treated mice
(two-sided non-parametric Mann–Whitney U-test). n = 9 vehicle- and
8 antibiotic-treated mice. c, Representative ex vivo recorded colonic migrating
motor complexes from control (top) and AhrEN-KO (bottom) mice.
d, Quantification (mean ± s.d.) of colonic migrating motor complexes
(CMMCs) from control and AhrEN-KO mice (two-sided non-parametric Mann–
Whitney U-test). n = 6 control and 5 AhrEN-KO mice. e, The negative feedback
regulation of AHR signalling by CYP1A1 (left) is the basis for the experimental
design to assess the role of neuron-specific Cyp1a1 overexpression on
intestinal motility (right). f, g, Myenteric ganglia from the colon of
Rosa26LSLCyp1a1 (f) and ENCyp1a1 (g) mice hybridized with the Cyp1a1 RNAscope
probe (green). Dotted line defines the borders of myenteric ganglia and arrows
indicate positive neurons. Data represent two independent experiments. Scale
bars, 30 μm. h, Quantification (mean ± s.d.) of the effect of neuron-specific
Cyp1a1 overexpression on total ITT (two-sided non-parametric Mann–Whitney
U-test). n = 17 control (WT + A AV and R26LSLCyp1a1) and 19 ENCyp1a1 mice. i, I3C-
supplemented diet rescues the total ITT increase observed in ENCyp1a1 mice
(two-sided non-parametric Mann–Whitney U-test) (mean ± s.d.). n = 5 control
(purified diet), 4 ENCyp1a1 (purified diet), 9 control (I3C diet) and 8 (ENCyp1a1
(I3C-diet-fed) mice (female). j, Experimental design for expressing AHR in
enteric neurons of microbiota-depleted mice. Wild-type SPF mice were
injected with control (A AV-control) (top) or AHR-expressing (A AV-AHR)
(bottom) A AV vectors and treated with antibiotics. All mice were fed with
I3C-supplemented diet one week before ITT analysis. k, Quantification
(mean ± s.d.) of the effect of combinations of A AV-control and A AV-AHR vectors
with antibiotic treatment on total ITT (two-sided non-parametric
Mann–Whitney U-test). n = 10 mice per group.