ÐÓ°ÉÔ­´´

Next time you tuck into a helping of crab bisque or lobster thermidor
ponder the following. The stomachs of these retiring crustacea have all
but revolutionised conventional thinking about nervous systems. In recent
years the tiny nerve centres that drive the grinding motions with which
the animals’ stomachs digest food have become central to theories about
how nerve cells ‘talk’ to each other. So central, in fact, that several
neurobiologists have devoted their careers to studying them.

Why so much interest? Part of the answer is that these nerve centres
generate rhythmic movements; or to be more specific, rhythmic stomach movements.
Because of this, the nerve centres have proved to be an ideal model for
researchers curious about the neural basis of our own rhythmic movements
– breathing, chewing, walking and so on – all of which result from rhythmic
patterns of electrical activity generated within the human brain and spinal
chord. In lobsters and crabs the four nerve centres – ganglia – that control
stomach rhythms are structurally simple and can easily be studied in isolation
from the living animal. One in particular, the stomatogastric ganglion (STG),
has been investigated in such detail over the past decade or so that it
is now arguably the best understood network of neurons in the whole of biology.

Nerve cells, or neurons, comprise a cell body surrounded by several
short fibres – dendrites – which receive incoming messages, and usually
one long fibre – axon – which carries electrical impulses to other cells.
At its tip, the axon branches, and the nerve endings form junctions – synapses
– with the dendrites or cell bodies of other cells. When an electrical impulse
arrives at the nerve endings it triggers the release of chemicals called
neurotransmitters. These cross the tiny gaps at the synapses and bind to
receptors on the membrane of the next cell, either increasing or decreasing
its electrical excitability. If there is enough neurotransmitter the receiving
cell fires an electrical impulse which, in turn, passes along the axon.
The last neuron in the chain of response, the motor neuron, has nerve endings
close to the target organ so that its neurotransmitters will stimulate
muscle contraction or glandular secretion, for example.

The basic mechanism by which neurons communicate is quite simple, but
in humans even the simplest action or thought probably requires millions
of interactions between nerve cells. It is the coming together of these
cells into networks that give nervous systems their flexibility. Research
into lobster and crab nervous systems revolves around a handful of key questions.
To what do networks of neurons like those of the lobster’s stomach owe their
innate ability to generate rhythms? How are such networks ‘programmed’?
Do they generate one, and only one, rhythm, or are they more versatile?

Tackling such questions in mammals is hampered by the sheer complexity
of their motor systems. The quadriceps muscle in a human leg, for instance,
is controlled by a hundred or more motor neurons, each of which projects
a fibre to its target muscle. A lobster muscle, by comparison, is controlled
by no more than five motor neurons. And the entire nervous system controlling
the lobster’s stomach contains no more than 1000 neurons, divided up into
four nerve centres or ganglia (see Figure 1): complex, certainly, but a
better bet for experimental analysis than the motor systems of rats and
mice.

Neurobiologists’ fascination with the stomach rumblings of crustaceans
stretches back to the 1960s, when the late Donald Maynard, then at the University
of Oregon, first spotted the four ganglia that make up the stomatogastric
system of lobsters. The rhythmic nature of their output was clear to Maynard,
as was their role in controlling stomach movements. But what were the neural
mechanisms producing the rhythms? More than twenty years later, and after
much dogged research, many of the pieces of the puzzle are now in place.

The starting point takes us back to food. Lobsters and crabs swallow
their food in large chunks, which pass via the oesophagus and cardiac sac
to the gastric mill region of the stomach, a muscular sac containing three
interlocking teeth that break up the chunks. From there the finely divided
mush passes into the pylorus, where it is filtered before being sent to
the midgut. All the muscles responsible for these movements are controlled
by the four ganglia of the stomatogastric system.

We know that the STG – the best defined of the four nerve centres –
controls the gastric mill and pylorus, each of which has its own distinctive
pattern of rhythmic movements. The STG contains just 30 neurons, linked
into well-defined circuits by networks of fibres and synapses. All but
two of these neurons are motor neurons, directly controlling the contractions
of individual muscles as well as being part of the network that produces
the whole pattern of rhythmic activity. All the neurons have an identifiable
shape and a distinctive electrical signature tune (which can be recorded
using a microelectrode). Each neuron also uses a specific neurotransmitter,
or combination of neurotransmitters, to pass messages to other neurons.

But the real beauty of the STG is that it continues to generate rhythmic
patterns of electrical activity even after it has been removed from the
animal. If the nerve centre is placed in a glass dish, it can be kept alive
by a salt solution that mimics the blood. In recent years this approach
has been exploited by several teams, notably those of Allen Selverston at
the University of California in San Diego, Maurice Moulins at CNRS-Universite
de Bordeaux, Eve Marder at Brandeis University in Boston and Ronald Harris-Warrick
at Cornell University. Today their results are challenging the orthodox
view of the way motor systems function.

Until recently, most theories were dominated by the notion that motor
rhythms are ‘hard wired’; that is, they are produced by inflexible neural
circuits whose neurons are programmed at birth to produce a specific type
of rhythm. But this concept has fallen in the face of evidence that every
neuron in the stomatogastric networks, and indeed the networks themselves,
can behave differently under different conditions. The neurons contain within
their membranes multiple ‘personalities’. This finding is now being extended
beyond the stomatogastric neurons and may well be a characteristic of all
nervous systems. Hence, the STG generates not just one gastric and one pyloric
rhythm but several variations of each. Its neurons can also participate
with those in other ganglia to produce rhythmic movements involving the
whole of the stomach, such as swallowing.

Orchestral themes

This striking neural flexibility seems to depend on a fascinating family
of neurotransmitters known as neuromodulators. Broadly speaking, neurotransmitters
can be divided into two camps: fast acting/short lasting types and slow
acting/long lasting types. The former hurtle across synapses, strike receptors
confined to a small target area on the neuron opposite and provoke a rapid
response – a sudden brief change in the voltage across the neuron’s membrane.
But the slow-acting neuromodulators, like hormones, act over a wider region.
Having been discharged by a neuron, neuromodulators scatter over the surfaces
of several target neurons. As they bind to receptor molecules on these
surfaces, they trigger a cascade of biochemical reactions, the usual outcome
of which is a long-lasting change in the behaviour of the target neurons
– influencing, for instance, the strength, frequency or likelihood of their
activity.

Neuromodulators have been found in all nervous systems from the lowliest
worm to humans. Because the nervous systems of vertebrates are so complex,
researchers have long struggled to gather information about the cellular
architectures and patterns of connections within their motor rhythm networks.
But enough is known for us to be confident that vertebrate nervous systems
share many basic features with their better characterised invertebrate counterparts,
including their ability to generate different patterns of electrical activity
under the influence of different neuromodulators. So invertebrate systems
seem to provide good models for unravelling the fundamental functions of
neuromodulators.

Many neuromodulators adjust the responsiveness of neurons to incoming
messages from fast neurotransmitters: in a sense, they are to a neuron what
a tuner is to a radio. In networks of neurons like those of the STG, neuromodulators
act to vary the way neurons ‘talk’ to each other. In so doing they enable
a single network of neurons to express a variety of electrical output patterns.
If you think of the network as an orchestra, the neuromodulators allow it
to play variations on a basic theme.

In the stomatogastric system, this basic theme is established by the
anatomical patterns of connections between neurons, which are forged during
the animal’s embryonic development and remain thereafter essentially fixed.
Some of these connections allow signals to pass from one nerve centre to
another. For instance, Marder and her team have discovered that at least
15 neuromodulators are released into the STG by input neurons stationed
in neighbouring nerve centres. Some of these neuromodulators, such as proctolin,
are peptides; others, such as 5-hydroxytryptamine, are amines. And each
has a different effect on the way the pyloric and gastric mill neurons behave.

It is largely because of neuromodulators that the various networks of
the STG can generate such a remarkable range of rhythms. The neural network
that controls the pylorus is a prime example. The stable, three-phase motor
rhythm this network generates (see Figure 2) is merely its basic theme:
neuromodulators can trigger variations in the speed with which the pattern
cycles, the point in the cycle when each neuron fires, and the intensity
of the resulting bursts of electrical impulses. And each characteristic
affects the way the pylorus moves.

The long-lasting effects of neuromodulators mean researchers can test
the impact on motor rhythms of each substance simply by bathing isolated
ganglia from the stomatogastric system in neuromodulator solutions. Teams
in several laboratories have now shown that different neuromodulators alter
the STG’s motor rhythms in distinctive ways. Some affect the network controlling
the pyloric sac, others the gastric mill network, and many affect both.

Such diversity, though, is not the only reason the networks of the STG
are so flexible. Equally important is the ability of individual input neurons
to secrete more than one type of neuromodulator at a time and in varying
quantities. This much is clear from experiments done some years ago by one
of us (Nusbaum) with Marder and by Paul Katz and Harris-Warrick. The aim
was to study the rhythms produced when neuromodulators are delivered the
natural way, from input neurons. Both groups found that stimulating input
neurons to discharge their chemical cargoes sometimes had different effects
on the STG’s rhythms than those produced when the neuromodulators were delivered
to the ganglion artificially, in the form of solutions. The reason turned
out to be that input neurons can release more than one neuromodulator.
More recent research suggests that input neurons can in fact regulate whether
they release one or more neuromodulators.

Bursting into rhythm

One of the most exciting breakthroughs came in the late 1970s and early
1980s, when Selverston and his colleagues traced the rhythmic activity of
the STG’s pyloric network to a single neuron. Dubbed the ‘anterior burster’,
this neuron produces rhythmical bursts of electrical impulses, or action
potentials, which travel out through the pyloric network, driving its varied
rhythms. The main reason the neuron fires rhythmically is that its membrane
contains a complement of different ion channels that open and close in a
sequence that alternately drives the neuron to fire impulse bursts and then
suppresses this activity.

Neurobiologists now know that neuromodulators such as dopamine – more
familiar as the substance that is depleted in the brains of people with
Parkinson’s disease – play a crucial part in varying the frequency and intensity
of the anterior burster’s basic rhythmic output. Neuromodulators also influence
the way this output ‘spreads’ through the network. Several teams have shown
that neuromodulators alter the strength of both neurotransmitter-mediated
synaptic responses and the rarer electrical synaptic communication known
as electrical coupling. The latter controls the spread of electrical rhythms
from the anterior burster neuron directly to certain motor neurons in the
network, by electrical signals without the use of a neurotransmitter. Dopamine,
for example, as well as boosting the rhythmic activity of the anterior
burster neuron, simultaneously decreases its electrical coupling with motor
neurons, so producing a distinct variant of the pyloric motor pattern.

Although many neural rhythms are driven by a central, bursting neuron,
others are not. The network that controls the heartbeat of the medicinal
leech, for example, has no bursting neurons. Its rhythms result instead
from the way in which its neurons communicate through ‘inhibitory’ synapses.
The passing of messages through such synapses switches off, rather than
stimulates, a receiver neuron. When two neurons are each connected to the
other by inhibitory synapses, as they are in the leech heart network, each
can switch off the other, with the result that only one of the pair can
be active at any one time. The neurons thus fire alternately, with electrical
activity ‘bouncing’ from one to the other. In effect, the neurons act as
a biological version of a simple electric oscillator.

By simulating the leech heart network on computer, one of us (Calabrese),
working with Erik DeShutter of California Institute of Technology in Pasadena,
has found that subtle interactions between the synapses and ion channels
of the network’s neurons generate electrical rhythms in pairs of neurons
that inhibit each other. Each of these interactions can be influenced by
neuromodulators, so the potential for flexibility is enormous.

Another surprise is that flexibility is not restricted to rhythm-generating
networks in the central nervous system. It also occurs at the junctions
between motor neurons and muscles. One of us (Calabrese) and colleagues,
for instance, have discovered that motor neurons stimulating the leech heart
release a set of peptides (as well as a fast-acting neurotransmitter) which
stimulate rhythmic contractions in the heart muscle. Other leech neurons
use the same set of peptides to adjust the strength of the heartbeat. Similarly,
the strength and rate of the heartbeat in mammals is modulated by noradrenaline
and acetylcholine released from nerves.

Influence from afar

A more complex example is emerging from research into the feeding behaviour
of the marine snail, Aplysia. Irving Kupfermann at Columbia University,
New York, and Klaudius Weiss at Mount Sinai Medical Center have discovered
that the muscles which move the animal’s mouthparts are stimulated by several
motor neurons, each loaded with a fast-acting neurotransmitter and a different
complement of neuromodulatory peptides. This molecular menagerie is capable
of stimulating a remarkable variety of muscle rhythms by reinterpreting
the commands generated in the central nervous system. The overall message
is clear. Synapses and neuromodulators situated well outside an animal’s
central rhythm networks can play a vital part in the rhythmic movements
that are actually generated.

Recent work in the mammalian central nervous system paints the same
picture. Mammalian motor neurons were, until quite recently, depicted as
the foot soldiers of motor systems. They received their marching orders
from central neural networks and accurately conveyed this information to
their muscle targets. As we have already discussed for invertebrate systems,
it is now clear that mammalian motor neurons are not always passive followers
of central commands. Sometimes they actively reinterpret these commands.
For example, the team of Hans Hultborn, Jorn Hounsgaard and Ole Kiehn and
colleagues at the University of Copenhagen in Denmark have shown that particular
subsets of these motor neurons dramatically enhance their responses to inputs
when they are under the influence of neurons located in the midbrain that
release noradrenaline and 5-hydroxytryptamine, apparently maintaining posture.

The question undergoing most radical revision is, what exactly constitutes
a neural network? The neurons that control the movements of a lobster’s
pylorus generate different motor rhythms in response to different neuromodulators.
The network’s structure remains the same, as does the number of neurons
it contains. It is the properties of the neurons and the strengths of their
synapses that change. To go back to the musical analogy, the rhythms are
like variations on a melody played by one section of the orchestra.

But this conservative picture, in which each network operates separately
and autonomously, doesn’t tell the whole story. Recent studies suggest
that under the influence of certain neuromodulators, the distinctions between
separate networks become fuzzy: some neurons may temporarily participate
in the rhythm of another network. It is like one oboe picking up the melody
of the violins.

This fuzziness is best seen in the stomatogastric networks that control
the gastric mill and pylorus. Traditionally, researchers have viewed them
as being separate systems. But it now seems that the networks may operate
in a fluid fashion. Under the influence of certain neuromodulators, many
neurons belonging to the gastric mill network become active members of the
pyloric network, and vice versa. Moreover, several neurons can be active
in both networks simultaneously. Perhaps the best way to view the STG is
as a single pool of neurons from which different subsets are selected for
different tasks – a state of affairs that neuroscientists refer to as multiple
task processing.

This fluid style of organisation extends beyond the networks of the
STG. Now, Patsy Dickinson at Bowdoin College in Maine, working with Marder,
has found that the networks that control a lobster’s cardiac sac and gastric
systems can collaborate to produce a single new rhythm (under the influence
of particular neuromodulators, of course).

More recently, Pierre Meyrand and John Simmers, working with Moulins,
have discovered that networks in all four nerve centres of the stomatogastric
system can work together to produce a single novel pattern of electrical
activity. The researchers think this pattern enables the animal to swallow,
something requiring all four compartments of the stomach to be simultaneously
active. The collaboration appears to result from the activity of a single
neuron located near the oesophageal ganglion, also part of the stomatogastric
system. When this neuron stops firing, the four separate motor patterns
reappear.

Perhaps, then, it is better to think of the stomatogastric system as
a single network of neurons, various parts of which can function alone as
the need arises. There is, however, a caveat. Most of these findings come
from studies of nerve centres removed from a lobster or crab. Do they tell
us anything about the animal’s real behaviour? Several teams, including
those of Hans-Georg Heinzel of the University of Bonn in Germany, Selverston
and Moulins have tackled this question by monitoring stomach movements and
neural activity in the living animal. The stomach is deep inside the animal
so this is far from easy. One solution is to use an endoscope inside the
stomach, linked to a video camera, to watch the movements of the teeth in
the gastric mill. Results gathered in this way confirm many of the findings
from dissected nerve centres. The concept of the flexible neural network
– or rather multiple task processing – seems to be a biological reality.

The challenge for the future is to understand how all the many possibilities
embedded in these networks are orchestrated so that the animal always does
the right thing at the right time. Here, inputs from sensory neurons are
likely to be important. This is clear from recent research on the stomatogastric
system of crabs. The motor patterns produced by the system are strongly
influenced by sensory inputs that provide moment-to-moment information about
the animal’s environment. This, it seems, helps to fine tune the animal’s
movements to its environment.

Neural flexibility of the kind seen in the control of lobster and crab
stomachs is not just a speciality of a few crustaceans. Neuroscientists
are finding it in many other nervous systems that generate rhythms in both
invertebrates and vertebrates. It may be that multiple task processing,
where combinations of different neuromodulators can ‘select’ from a large
and flexible neural network exactly the right combination of neurons for
a particular job, is common to most nervous systems. Certainly this is an
economical way of expanding the functional repertoire of a nervous system
without the need to evolve more neurons. Perhaps most neurons can function
as members of several different networks, rather than being ‘hard-wired’
into specific networks. If so, the stomach nerves of the lobster may have
revealed one of the secrets behind the enormous flexibility of the human
brain.

Michael Nusbaum is assistant professor in the Neurobiology Research
Center and Department of Physiology and Biophysics at the University of
Alabama at Birmingham. Ronald Calabrese is professor in the Department of
Biophysics at Emory University, Atlanta, Georgia.