ture.
Indeed, the cells of the skin that even-
tually give rise to feathers may acquire their
identity along the medial-lateral axis, later
revealed by their color, by receiving some
signals from adjacent embryonic struc-
tures. To directly test this idea, the authors
used grafting experiments, swapping small
pieces of embryonic tissue between spe-
cies with slightly different patterns of color
stripes. Swapping the neural tube between
species had no effect on coloration. By con-
trast, swapping a few somites changed the
agouti expression pattern, and in turn the
local feather coloration, in the chimeric
chicks. Hence, skin cells acquire their po-
sitional identity from a landmark signal
produced by the underlying somites. The
pattern variation may in turn result from
changes in the strength of this signal, or in
how the skin cells interpret it.
The study of Haupaix et al. stands out as
a direct demonstration that color pattern-
ing on the coat of vertebrates relies on spa-
tial patterning information from embryonic
structures. Rather than contradicting the
idea that local cellular interactions or reac-
tion–diffusion phenomena govern the es-
tablishment of color patterns, the existence
of an embryonic blueprint provides a new
framework within which to consider these
processes. Indeed, the embryonic blueprint
may guide or bias how skin cells interact. Ar-
chitects design buildings and decide where
structural elements such as walls, windows,
pipes, and chimneys are placed. These
structural landmarks impose positional
constraints for later decoration. Designers
have to work within these constraints, but
they can still independently create a unique
appearance to each building. By analogy,
spatial embryonic structures may provide a
template that first translates into skin land-
marks, which may in turn set specific initial
conditions to the local cellular interactions
that will define the final coloration pattern.
The work of Haupaix et al. opens a
new era where self-organizing processes
are no longer systematically opposed to
embryonic patterning. Yet to reach the
integration of these processes and fully un-
derstand color pattern formation and color
pattern evolution, research is needed to
investigate how the early embryonic blue-
print and the late cellular interactions are
coordinated. Identifying the nature of the
somitic signal patterning the skin will be a
first step in this direction. j
REFERENCES
1. S. Kondo, H. Shirota, Semin. Cell Dev. Biol. 20, 82 (2009).
2. M. Inaba et al., Science 335, 677 (2012).
3. A. P. Singh et al., Nat. Cell Biol. 16, 604 (2014).
4. N. Haupaix et al., Science 361, eaar4777 (2018).
10.1126/science.aau5103
SCIENCE sciencemag.org
NEUROSCIENCE
A gut feeling
Gut-brain signaling involves neurotransmission
from gut sensory epithelial cells
By Benjamin U. Hoffman1,2 and
Ellen A. Lumpkin1,3
T
he anatomist Friedrich S. Merkel pre-
dicted in 1880 that sensory systems are
composed of epithelial cells and sen-
sory nerves, which together transform
environmental cues into neural signals
that trigger our rich sensory experi-
ences (1). We now know that this hypothesis
mostly holds true for the canonical senses
of vision, hearing, taste, and touch. Perhaps
surprisingly, the peripheral outposts of these
classical sensory systems (eyes, ears, tongue,
and skin) are dwarfed by the human body’s
largest sensory organ—the gut. Enteroen-
docrine cells, which are rare epithelial cells
that decorate the gut lining, have long been
suspected to be sensory receptor cells that
inform the brain about ingested nutrients
(2). Since their description, these cells were
assumed to play a role in metabolism and gut
physiology by releasing slow-acting peptide
hormones that stimulate neurons through-
“…findings overturn a
decades-old dogma that
enteroendocrine cells
signal exclusively through
hormones.”
out the gut and in the brain. On page 1219
of this issue, Kaelberer et al. (3) challenge
this view by demonstrating that gut entero-
endocrine cells locally excite sensory nerves
through release of the neurotransmitter glu-
tamate. A recent study of enterochromaffin
cells, a subset of enteroendocrine cells, also
found that gut signals are transmitted at
epithelial-neural synapses through release of
the neurotransmitter serotonin (4). Together,
these findings overturn a decades-old dogma
that enteroendocrine cells signal exclusively
through hormones.
Kaelberer et al. focused on a subset of
mouse enteroendocrine cells marked by
1
Department of Physiology and Cellular Biophysics, Columbia
University, New York, NY 10032, USA. 2Medical Scientist
Training Program, Columbia University, New York, NY 10032,
USA. 3Department of Dermatology, Columbia University, New
York, NY 10032, USA. Email: eal2166@columbia.edu
Published by AAAS
expression of two neuropeptide precursor
proteins, cholecystokinin (CCK) and pep-
tide YY (PYY), which are cleaved to produce
hormones that control hunger (5, 6). The
authors confirmed that these enteroendo-
crine cells possess anatomical hallmarks of
presynaptic cells and that they form direct
synaptic connections with vagal and spinal
sensory neurons in vivo (3, 7). Moreover, in-
fusing sucrose, or table sugar, into the gut
caused vagus nerve activation in an entero-
Downloaded from http://science.sciencemag.org/ on October 7, 2018
endocrine cell–dependent manner in vivo.
Given the location and relative scarcity
of enteroendocrine cells, their function is
difficult to study in the intact gut. To over-
come this challenge, the authors used an in
vitro co-culture model of enteroendocrine
cells and sensory neurons. Stimulation of
enteroendocrine cells with D-glucose, a
component of sucrose, triggered excitatory
ionic currents and action potentials in
their associated neurons. Further studies
revealed that enteroendocrine cells activate
sensory neurons within tens to hundreds of
milliseconds, a time scale typical of synap-
tic transmission rather than neuropeptide
signaling.
Other epithelial-neuronal synapses, such
as inner-ear hair cells and retinal pho-
toreceptors, employ glutamate as a neu-
rotransmitter. Using genetically engineered
biosensors, Kaelberer et al. demonstrate
that D-glucose stimulates glutamate release
from enteroendocrine cells. Importantly,
blocking glutamate-gated ion channels that
mediate synaptic transmission abolished
D-glucose–induced currents in vagal neu-
rons in vitro. In vivo studies showed that
glutamate-gated ion channels are required
for early (<60 s) vagus responses to dietary
sugars, whereas neuropeptide receptors are
responsible for vagus activity on a longer
time scale. Thus, enteroendocrine cells are
chemosensory cells that activate vagal sen-
sory neurons via release of glutamate (see
the figure). Kaelberer et al. hypothesize
that synaptic transmission confers precise
spatial and temporal information about gut
contents. Such sensory feedback could al-
low neural control of rapid processes that
modulate gut physiology, such as muscle
contraction, whereas neuropeptides such
as CCK are responsible for long-lasting be-
havioral outputs of gut stimulation, such
as satiety (5, 6). A remaining issue is what
21 SEP TEMBER 2018 • VOL 361 ISSUE 6408 1203
INSIGHTS | P E R S P E C T I V E S
behaviors or physiological processes rely on
subsecond neural signaling from the gut.
The finding that gut enteroendocrine
cells form glutamatergic synapses identifies
a first relay in the neural circuit through
which the brain detects gastrointestinal
contents, such as nutrients and microbes,
to control satiety, metabolism, and di-
gestion. Glutamate is arguably the most
ancient excitatory neurotransmitter: Gluta-
mate receptors are encoded in the genomes
of primitive nervous systems, such as in
ctenophores (marine invertebrates), and of
Trichoplax, an animal that emerged before
nervous systems evolved (8, 9). It is interest-
ing that the mammalian gut has co-opted
Epithelial sensory cells
In addition to secreting neuropeptides, entero-
endocrine cells rapidly convey information about
nutrients in the gut by releasing neurotransmitters
to excite vagal and spinal sensory neurons.
Nutrients Satiety
Stretch Metabolism
Irritants Infammation
Microbes Digestion
Gut stimuli
Enteroendocrine cell
Fast neurotransmitters
(glutamate, serotonin) Sensory
Slow neuropeptides (CCK, PYY) neuron
this ancient neurotransmitter to mediate
nutrient sensing, which is perhaps the ear-
liest sense to arise during evolution.
Enteroendocrine cells are remarkably
heterogeneous, with at least 12 molecularly
distinct subtypes that express a multitude of
sensory receptor proteins and neurotrans-
mitters (10, 11). Indeed, enterochromaffin
cells, which lack expression of CCK and
PYY, employ serotonin to convey stimuli
such as stretch, irritants, and metabolites
to vagal sensory neurons (4, 12). An intrigu-
ing possibility is that enteroendocrine cells
use discrete neurotransmitter systems to
communicate with functionally distinct
1204 21 SEP TEMBER 2018 • VOL 361 ISSUE 6408
subsets of vagal and spinal sensory neurons
(13). Such a system could allow detection of
different types of ingested nutrients such
as sugars, fats, or proteins, or could partici-
pate in neural control of the progression of
contents through the gastrointestinal tract.
Perhaps most relevant to human health
is the possibility that synaptic transmis-
sion between gut epithelial cells and sen-
sory neurons provides a signaling node
through which microbes interact with our
nervous system. The gut provides an array
of essential functions for humans includ-
ing not only nutrient sensing, metabolism,
and absorption, but also a barrier to infec-
tion from the multiplicity of organisms that
compose the gut microbiome. Enteroendo-
crine cells express toll-like receptors, which
detect bacterial products to activate innate
immune responses and maintain the epi-
thelial barrier. Activation of these receptors
on enteroendocrine cells induces release of
neuropeptides, including CCK (14). These
data suggest that enteroendocrine cells
might not only serve as detectors of gut mi-
crobiota, but also participate in proinflam-
matory and other immune responses.
The findings of Kaelberer et al. pose a
number of tantalizing open questions. For
example, what are the molecular mecha-
nisms of neurotransmitter release in en-
teroendocrine cells? Which postsynaptic
receptors mediate glutamatergic transmis-
sion in vagal neurons? Moreover, how do
enteroendocrine cells modulate neuronal
signals in the presence of environmental ir-
ritants or in intestinal disorders? The abil-
ity to investigate the first relay station of
the gastrointestinal–nervous system circuit
in vitro, in conjunction with genetic target-
ing techniques, provides an exquisite plat-
form to answer these questions. Furthering
our understanding of the molecular mecha-
nisms that mediate signaling along the gut-
brain axis provides the foundation for the
development of future therapeutics to treat
gastrointestinal disease. j
REFERENCES
1. F. S. Merkel, Über die Endigungen der sensiblen Nerven in
der Haut der Wirbeltiere (H. Schmidt, Rostock, 1880).
2. F. Feyrter, Über diffuse endokrine epitheliale Organe (J.A.
Barth, Leipzig, 1938).
3. M. M. Kaelberer et al., Science 361, eaat5236 (2018).
4. N. W. Bellono et al., Cell 170, 185 (2017).
5. L. R. Beutler et al., Neuron 96, 461 (2017).
6. Z. Su, A. L. Alhadeff, J. N. Betley, Cell Rep. 21, 2724 (2017).
7. D. V. Bohórquez et al., J. Clin. Invest. 125, 782 (2015).
8. A. Krishnan, H. B. Schiöth, J. Exp. Biol. 218, 562 (2015).
9. L. L. Moroz et al., Nature 510, 109 (2014).
10. A. L. Haber et al., Nature 551, 333 (2017).
11. D. Grün et al., Nature 525, 251 (2015).
12. C. Alcaino et al., Proc. Natl. Acad. Sci. U.S.A. 115, E7632
(2018).
13. E. K. Williams et al., Cell 166, 209 (2016).
14. M. Bogunovic et al., Am. J. Physiol. Gastrointest. Liver
Physiol. 292, G1770 (2007).
10.1126/science.aau9973
Published by AAAS
ARCHAEOLOGY
Bang or
whimper?
The evidence for collapse of
human civilizations at the
start of the recently defined
Meghalayan Age is equivocal
By Guy D. Middleton1,2
T
he International Commission on
Stratigraphy recently announced the
creation of a new unit in the scale of
Downloaded from http://science.sciencemag.org/ on October 7, 2018
geological time, the Meghalayan Age,
from 4200 years before present, or
2200 BCE, to the present. The Commis-
sion explains that this period began with a
two-centuries-long megadrought that caused
the collapse of civilizations in Egypt, Greece,
Syria, Palestine, Mesopotamia, the Indus Val-
ley, and the Yangtze River Valley (1). However,
there is little archaeological evidence for such
sudden, widespread civilizational collapse.
Since at least 1971, scientists have repeat-
edly argued that a major drought caused
civilizational collapse at numerous locations
at this time (2). Some scholars have amassed
impressive amounts of data to demonstrate
the existence of a megadrought and have
linked this causally with civilizational col-
lapse (3). However, detailed archaeological
and historical analysis, including recent in-
vestigations of chronology and paleoclimate,
suggests that rather than simultaneous civili-
zational collapse, different kinds of changes
occurred in different parts of the world at dif-
ferent times, all of them less abrupt than once
thought (4–9). The environmental and cli-
matic determinism behind the megadrought-
collapse narrative fails to account for specific
historical circumstances, the power of hu-
man agency to drive substantial change, and
the translation of environmental factors into
cultural and sociopolitical contexts. Current
evidence, therefore, casts doubt on the utility
of 2200 BCE as a meaningful beginning to a
new age in human terms, whether there was
a megadrought or not.
To understand in more detail what hap-
pened around 4200 years ago, consider the
GRAPHIC: V. ALTOUNIAN/SCIENCE
situation in several of the locations for which
collapse has been suggested. In Egypt, there
1
Czech Institute of Egyptology, Charles University, Prague,
Czech Republic. 2School of History, Classics and Archaeology,
Newcastle University, Newcastle upon Tyne, UK.
Email: gdmiddletonphd@gmail.com
sciencemag.org SCIENCE
A gut feeling
Benjamin U. Hoffman and Ellen A. Lumpkin

Science 361 (6408), 1203-1204.

DOI: 10.1126/science.aau9973

Downloaded from http://science.sciencemag.org/ on October 7, 2018

ARTICLE TOOLS http://science.sciencemag.org/content/361/6408/1203

RELATED http://science.sciencemag.org/content/sci/361/6408/eaat5236.full
CONTENT

REFERENCES This article cites 12 articles, 3 of which you can access for free
http://science.sciencemag.org/content/361/6408/1203#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title
Science is a registered trademark of AAAS.