How did your fingers and toes know how long to grow, so that even before you
were born, your little fingers were smaller than your middle fingers, and your
big toes bigger than the others? Come to that, how do the parts of any animal or
plant “know” when they’ve reached the right size and to stop growing? Despite
finding a whole collection of genes that tell each cell whether it’s part of a
hand, a head, a wing or a leaf, scientists know remarkably little about how
everything grows in the proper proportions to give organisms their
characteristic shapes. It’s as if every structure in the body is following a
plan, but where is it and how does it work?
On the face of it, this doesn’t seem such a tough problem. Surely to make a
middle finger you just need to make a few more cells than you do to make a
little finger. And to make a wing, perhaps you just need to count out a certain
number of wing cells. But new research is showing that counting isn’t
everything. Bruce Edgar and his colleagues from the Fred Hutchinson Cancer
Research Center in Seattle, have a lab full of perfectly normal-looking fruit
flies (Drosophila), with normal-sized wings and wing veins in all the
right places. Normal, that is, until you look closer. Because they all have the
wrong number of wing cells.
It’s perhaps not surprising to find that flies can’t count. Newts can’t
either. Half a century ago, researchers were breeding newts with strange-sized
cells. Normal cells have two copies of every chromosome, and so are called
“diploid”. But disrupt the normal process of cell division and you can get newts
with tetraploid cells containing twice the normal amount of genetic
material—four copies of each chromosome. These tetraploid cells grow to be
much larger than diploid cells. And haploid cells, with only one copy of each
chromosome, are about four times smaller than tetraploids. But haploid and
tetraploid newts still grow to exactly the same size. The haploid newt makes up
for its tiny cells by having four times more of them. So, at least in
amphibians, the relative size of body parts doesn’t rely on counting cells.
Advertisement
Playing catch-up
But there’s more than one way to count a cell. Rather than counting the
number directly perhaps organisms time how long it takes to produce them. The
length of the growth phase, which affects how many cells are produced, should
dictate proportions in a predictable way—the longer the growth phase, the
bigger the body part. But a mutant strain of Drosophila ruled out this
idea two decades ago.
This strain of fly sometimes has slow-growing cells in one half of its wing
but normal cells in the other half. If timing were all-important, the two halves
should stop growing at the same time, leaving the slow half dwarfed. But that’s
not what happens. The slow half eventually catches up and the wings end up
looking quite normal. So timing isn’t vital either.
If proportions don’t depend on counting cells or timing growth, the newt
experiments leave another possibility open. Maybe size can be guided by counting
chromosomes or particular “marker” genes in each finger, toe or wing, rather
than the number of cells they contain. Tetraploid newts may have four times
fewer cells than haploid newts, but each cell has four copies of the genome, so
overall the tetraploids and the haploids have the same total number of
chromosomes.
But last summer, Edgar and his colleagues showed that even indirect cell
counts aren’t needed. They managed to change the number of cells in one half of
a Drosophila wing by manipulating the genes that control the rate of
cell division. If the cells divide too quickly, then the wing ends up with too
many cells. If they divide too slowly, it ends up with too few. If the size and
shape of the wing depends on counting cells—either directly or
indirectly—then the two halves of these wings should end up very different
sizes.
But, amazingly, the experiments have no effect on size. “We have adults that
have too many cells in the wings but the pattern of wing veins and the size of
the wings are completely normal,” says Edgar. “And we can do the reverse and
have wings with fewer cells, and the pattern and size is still normal.”
If the wing is a jigsaw with cells as its pieces, then Edgar and his
colleagues have shown that the relationship between the pieces and the puzzle is
far more flexible than we thought. Flies with too many cells in their wings
simply use each cell as a smaller piece of the whole, while flies with too few
wing cells stretch each to make a larger piece of the puzzle.
The experiment tears away the last arguments against the conclusion that
developing animals rely on counting cells to measure their size. Edgar believes
that organisms must have a technique for measuring lengths and volumes. “That’s
the simple interpretation and it makes sense in terms of our results,” he says.
There must be some sort of “molecular tape measure” that cells use to size up
organs, limbs and so on as they follow their body plan. But how does it
work?
Researchers hunting for this tape measure are starting to look for clues in
the mechanisms that dictate cell identities and body plans. Experiments on the
products of genes that dictate the layout of Drosophila wings—for
example, those setting what’s up and down, or front and back—now suggest
that these genes may also promote the growth required to realise these
patterns.
Like all flies, the Drosophila egg hatches into a maggot (larva),
which grows and then pupates, to metamorphose into the adult fly. In newly
hatched fruit fly larvae, the cells that eventually form the wings of the adult
fly are laid out in single layers known as “imaginal discs”. Initially, each
wing disc consists of around fifty cells, but by the time the larva turns into a
pupa there are about 50 000. As the wing cells proliferate inside the larva they
form a structure rather like a deflated balloon. At this stage the developing
wing is inside-out, but when the adult emerges from the pupa it turns the whole
structure the right way round, pumps it up by filling the veins with blood, and
flattens it into the mature double-layered wing.
The pattern of cell types in the wing—for example, where it forms a
hair or a vein cell—is controlled by two proteins, whose concentrations
tell the cells where they are. The two proteins are called Decapentaplegic (Dpp)
and Wingless (Wg). In the wing disc, Dpp is produced by a ribbon of cells
running along the centre, at the border between what will become the front and
back halves of the wing. And a stripe of Wg-making cells marks the position of
the fold that will eventually create the edge of the wing, bordering its upper
and lower surfaces. The rows of Dpp-producing and Wg-producing cells cross at
what becomes the wing tip, where cells make both proteins.
It is now becoming clear that Dpp and Wg not only determine the pattern and
structure of the wings, they are also vital for controlling their growth and
ultimate size. By genetically engineering or breeding fly mutants with extra
groups of Dpp and Wg-producing cells in the developing wing, several research
groups have produced flies with extra wing-tip-forming patches. These patches
stimulate neighbouring regions of the wing disc to produce an extra wing growing
straight out of the original wing—and the extra wing reaches roughly
normal proportions.
Spheres or stripes
So these pattern-forming proteins can also stimulate growth in a wing, but
how about growth more generally? Could they control the size of other
structures? Peter Lawrence, Bénédicte Sanson and Jean-Paul Vincent
of the Medical Research Council Laboratory of Molecular Biology in Cambridge
have found that stripes of Wg do seem to promote growth earlier in
Drosophila’s development, when it is still an embryo inside the egg. Just
as in the developing wing, Wg and Dpp are produced by stripes of cells,
corresponding to the radial segments of the larva. Intriguingly, the research
points to a way in which protein stripes could establish a molecular ruler,
stimulating growth until a developing organ reaches its proper size.
Lawrence’s team studied embryos with several gene mutations that prevent them
forming the proper body segments and the stripes of Wg-producing cells. Such
embryos fail to elongate into normal sausage-shaped larvae and instead form tiny
spheres. Even when the Wg-producing gene is put in all the cells of the deformed
embryos, they stay tiny and spherical. So it takes more than the presence of Wg
to trigger an increase in length. However, when Lawrence and his colleagues
introduced the Wg-producing gene into stripes of cells, as they normally appear,
the mutant embryos grew longer—and regained their normal sausage
shape.
“I think it’s done by the steepness of gradients,” says Lawrence. He believes
the secret may lie in a cell detecting a big difference in the amount of Wg on
one side compared with the other. Only when there are discrete stripes of
Wg-producing cells will steep gradients of Wg concentration develop. If all the
cells are producing Wg, the protein will be evenly distributed and no gradients
will develop, and Lawrence’s group has showed that only when there are steep
gradients of Wg do the embryos change from spheres into sausages.
According to Lawrence, the same method could explain how animals control
other proportions. When a Drosophila wing disc is small, for example,
there will be steep gradients of Dpp levels, with a peak along the centre and
very low concentrations at the edges. As the wing grows, the gradients would
become stretched and so the difference in concentration detected by each cell
from one side to the other would drop. So Lawrence suggests that cells measure
the size of the gradient and stop growing when it is sufficiently gentle. This
would provide a way to control the absolute size of an organ and allow animals
to grow into the correct shape without resorting to a cell count. “It is
potentially independent of cell numbers,” he says.
How much of the work on Drosophila can be extrapolated to the
control of shape in other animals, such as humans? Clearly the growth programme
of insects is completely different from that of vertebrates like us. In
particular, our shape is already set in the womb and we grow roughly in
proportion thereafter. But early fetal pattern formation—when dimensions
such as the relative lengths of fingers are set—may be governed by
concentration gradients. In fact, close relatives of the Drosophila Wg
and Dpp proteins help to establish the patterns of cell-types in vertebrate
embryos. “There are an amazing number of commonalities between systems,” says
Lawrence. “So there is a good chance that the growth mechanisms that will be
discovered in the lower animals will turn out to act elsewhere.”
Indeed, biological tape measures that rely on concentration gradients may be
shared not only by flies and other animals, but also by plants. In 1995,
researchers in Belgium and France reported the effects of inhibiting cell
division in tobacco plants. They found that the plants grew almost to normal
size despite having far fewer cells than normal. The drop in cell numbers was
compensated for by an increase in cell size. And in November last year, a group
at the University of Pennsylvania in Philadelphia found that tobacco plants
whose leaf cells have been made more than usually sensitive to the plant growth
hormone auxin compensate for the resulting cell enlargement by reducing cell
numbers. Here too, the leaves end up the normal size.
But if gradients are the basis of the rulers that measure proportions, it’s
not clear exactly how they work. Does the initial steepness of the
gradient—dictated both by the amount of protein released and by the area
over which it spreads—determine the ultimate size of the body part? For
example, in humans, there could be a gradient of a growth-promoting protein
along the embryonic middle finger that is initially steeper than the gradient
along the embryonic little finger to set the relative lengths. And is there a
ruler for each dimension, one for finger length and others for diameter?
The final solution is bound to be complex. Although Edgar’s experiments show
that developing animals do not need to count cells to guide their shape, it is
clear that under more natural conditions, embryos and larvae keep strict control
of cell numbers. Edgar’s experiments loosened this control without affecting
body shape, but throw the right spanner into the workings of cell division and
the machinery for shape control also falters.
Imaginal disc cells in Drosophila larvae produce the gas nitric
oxide, which, among other things, hinders cell proliferation. In 1996, a group
of researchers from the Cold Spring Harbor Laboratory in New York State showed
that giving larvae bigger than normal doses of nitric oxide slows cell division
and makes them grow into flies with stunted legs and wings. Inhibiting nitric
oxide production lifts the brakes on proliferation, and makes flies with
extra-large legs and wings. Changes in cell numbers don’t seem to be compensated
for by changes in cell size— unlike the flies in Edgar’s lab.
The ability of flies to control their shape despite disruptions in cell
division seems to depend heavily on the conditions of particular experiments,
and integrating the different strands of research may take some time. Such a
vital thing as shape control is bound to use a belt and braces approach. But one
thing is certain, evolution designed the molecular ruler long before the human
race swatted its first fruit fly.
