GRAB a biology textbook and turn to the embryology chapter. You’ll almost certainly find a paragraph or two about the “progress zone” model, one of the most influential and firmly entrenched notions in biology. It explains how every four-limbed creature with a backbone, from alligators to zebras to humans, develops arms and legs. Some biologists think it also goes a long way towards explaining the entire mind-boggling process by which a single cell turns into a bouncing baby.
The PZ model was proposed 30 years ago by biologist Lewis Wolpert of University College London to explain how a bunch of undistinguished cells transform themselves, in strict sequence, into the various parts of a wing, arm or leg. It was simple, elegant and was backed up by good experimental evidence, and it quickly became part of the furniture. “Everyone assumed it was right,” says Gail Martin, a developmental biologist at the University of California in San Francisco (UCSF). And that assumption meant that after a while no one really questioned the PZ model, much less subjected it to tough scientific scrutiny.
But last year, two teams in the US independently published research suggesting that the PZ model is wrong. Although both insist they didn’t set out to challenge the orthodoxy, they say that what they found cannot be reconciled with the model and requires a new theory to explain it.
Advertisement
The results sent a shock wave through the world of developmental biology. Egos were bruised and defenders of the PZ came forward to refute the new findings, insisting that the new data fits with the old model. But the upstarts are sticking to their guns, and increasingly their colleagues are ready to think the unthinkable and ditch one of their most cherished theories.
This is more than an esoteric debate about the intricacies of limb development. Congenital limb abnormalities are common, and it is only by understanding normal development that we can see where things have gone wrong. More fundamentally, the PZ model is often invoked to explain how other aspects of an embryo’s development are orchestrated; recently it was proposed as a key mechanism in determining the main head-to-foot axis. So if the PZ model is wrong then a great chunk of developmental biology might need rethinking.
Look at your arm and you’ll see it is divided into three basic units: upper arm, lower arm and hand. Your leg, too, conforms to this pattern, as do all vertebrate limbs, from bats’ wings to whales’ tails.
The PZ model describes how this pattern emerges. In vertebrate embryos, limbs start out as limb buds, tiny bulges of unspecialised cells that grow progressively larger until they contain all the elements of a three-part limb. In chicks, this process starts around three days after fertilisation and takes about a week. In a human embryo it lasts a little longer, starting in week 5 and ending in week 8.
According to the PZ model, the key player in the process is the “progress zone”, a mass of rapidly proliferating but undifferentiated cells that sits just behind the tip of the limb bud. The PZ is the limb’s engine of creation, constantly extruding new cells from its trailing edge which go on to make the bones, muscles and cartilage of the fully-formed limb.
But how do these new cells know which parts of the limb they’re going to become? According to the PZ model, it’s all about timing. Every cell in the progress zone has an internal clock that started ticking as soon as the limb bud formed. New cells inherit their mother’s clock, but as soon as they are pushed out of the PZ the clock stops. At that point the cells’ fate is sealed. Cells whose clocks only ticked for a short time before leaving the PZ know to form part of the upper arm. Cells with a longer time on their clock go on to become lower arms, or longer still, hands.
The exact nature of this clock has never been determined, but according to the PZ model it is set ticking – and kept ticking – by signalling chemicals called fibroblast growth factors. These are secreted by a crest of specialised tissue on the very tip of the limb bud, the “apical ectodermal ridge”. The ridge sits on top of the progress zone and pumps growth factors into it. The extent of the zone is determined by the factors, because when a cell gets too far from the ridge the chemical signals can no longer reach it and its clock stops. This is a crucial part of the PZ model: cells only differentiate after they leave the progress zone.
Much of what we know about limb patterning comes from studies of chick embryos, and the PZ model is no exception. Biologists found they could cause dramatic changes to limb structure by removing apical ectodermal ridges from chick embryos. Cutting off the ridge early in development produced a limb with a normal upper bone, but nothing farther down – it was as if all the cells’ clocks had only been ticking for a short while. Removing it progressively later allowed more and more of the limb to develop normally. The results were taken to indicate that limb patterning happens piece by piece: the upper part first, then the lower part, and then the hand or foot.
But the new research calls this model into question. It proposes that there is no such thing as the progress zone, no ticking clock, and no section-by-section patterning. Instead, cells in the limb bud already know what they will become. The bud, in effect, is a tiny, preformed limb.
The first inkling of the new idea came last August, when Clifford Tabin and Andrew Dudley of Harvard Medical School in Boston, along with Marian Ros from the University of Cantabria in Spain, published some startling results in Nature (vol 418, p 539). Tabin and his colleagues weren’t looking to test the PZ model – far from it. They had set out to define the size of the progress zone more precisely. They did this by injecting permanent dye into the limb buds of three-day-old chick embryos at greater and greater depths. In this way they laid down traceable signals in the form of thin layers of dyed cells progressively further from the apical ectodermal ridge.
The PZ model predicts that dye placed anywhere in the progress zone will spread through the entire limb as cells proliferate. Tabin hoped to use this to find out where the boundary of the zone was. How deep did his dye layer have to be before it was no longer distributed throughout the finished limb?
What he saw came as a surprise. Regardless of the depth, the dye never spread through the limb. Instead, it stayed fixed in place. When he labelled cells near the outer tip of the limb bud, the dye ended up in the fingers. When he labelled cells halfway down, the dye ended up in the forearm. It seemed as though cells inside the progress zone, far from being undifferentiated, already knew where they would end up.
It was a troublesome result that needed checking out, so they designed another experiment. Ros took the very tips of wing buds from chick embryos at various stages and grafted them onto the stubs of limb buds that had been cut off. If the PZ model is right, grafts from early-stage embryos should have grown into complete limbs. But regardless of whether the tips came from early or late limb buds, Ros found that they only produced digits – a strong indication that their fate had already been sealed.
What was going on? The classic experiments that led to the PZ model would seem to rule out such early specialisation. Everyone agrees that removing the apical ectodermal ridge early on results in total loss of lower arms and hands. But if there are parts of the limb bud already predestined to become these structures, why do they disappear?
To reconcile the apparently contradictory results, Tabin turned to previous work suggesting that removing the apical ectodermal ridge triggers a burst of cell death in the limb bud. According to the PZ model, when the ridge is removed, the supply of growth factors is abruptly cut off so the internal clock is stopped too soon, meaning the lower parts of the limb never form. Tabin, though, wondered whether cell death alone could account for the truncations.
There was a way of finding out. He removed the ridge from limb buds at different stages of development and measured the extent of cell death. It turned out that the depth of cell loss was always the same – about 200 micrometres – regardless of when the ridge was removed. That could explain the experimental data supporting the PZ model: lose 200 micrometres’ worth of cells early on, when the limb bud is very small, and you’ve lost a huge proportion of your tiny pre-programmed limb. Perhaps only the upper part will remain. But later on, when the tiny limb has had a chance to grow, the loss of 200 micrometres is much less of a problem.
Tabin’s group has now decided to go out on a limb. “Before, I always thought of limb development in terms of the PZ model, but our work now suggests that there is no special zone, and it is not progressive,” says Ros. The group set about working out the details of an alternative theory.
Their “early specification” model is radically different from the PZ model. Instead of differentiating after they leave the progress zone, cells in the limb bud already know their fate. Some are destined for the shoulder, others for the hand. Once this pattern has been laid down, the different regions expand one at a time to form the complete limb.
The alternative model still has important gaps in it. In particular, it doesn’t say how and when the pre-patterning occurs. Tabin’s group thinks that cell specification takes just a few hours to sweep through the limb bud, starting about 72 hours after fertilisation with those destined to become the shoulder. “But we don’t really know yet what the mechanism for this is,” Ros admits.
Nevertheless, support for the new model has been quick to arrive. Even as Tabin was putting his new picture of limb development together, Martin at UCSF was independently doing experiments that would also cast doubt on the PZ model.
Martin was interested in exploring exactly how fibroblast growth factor (FGF) signals from the apical ectodermal ridge affected limb development. So she created mouse mutants in which the expression of genes for both FGF4 and FGF8 – two key members of the FGF family that can stand in for the apical ectodermal ridge when it is cut off – was knocked out during development.
In the PZ model, FGFs call the tune. They keep the cell’s internal clock ticking. Take them away early and limb development stops.
At first, what Martin found wasn’t so surprising: without FGF4 and FGF8, limb development ceased. But when she took a closer look at the underdeveloped paws of her mice, she saw something that couldn’t be reconciled with the PZ model. Although the back paws didn’t form at all, the front paws did. Thanks to a quirk in the experimental system, the front limb buds produced a transient pulse of FGFs before the genes were knocked out. And this created some very puzzling patterns.
As predicted by the PZ model, the front paws had normal upper-arm bones. But they also had something the PZ model would rule out: wrist, hand and finger bones, albeit smaller than normal.
Further analysis of the mutant mice convinced Martin that FGFs have nothing to do with internal clocks, but instead promote cell proliferation. All this fits nicely with Tabin’s new model: the limb bud contains precursor cells for all three limb parts, and they proliferate under the influence of growth-promoting FGFs. Last August, Martin’s team published its results alongside Tabin’s (Nature, vol 418, p 501). The gauntlet had been well and truly thrown down.
Wolpert was quick to pick it up. “This is all overblown,” he says. “Martin’s results have no real bearing on the model, and Tabin’s findings do not show the PZ model to be wrong.” In particular, Wolpert points to experiments from the late 1970s which he says still provide strong support for the PZ model. In these, he irradiated early limb buds with X-rays, killing most of the cells in the progress zone. The resulting limbs lacked upper parts but had normal fingers. The PZ model accounts for this by saying that the surviving cells occupy a depopulated progress zone for longer than usual, so none leaves early and no upper limb forms. Wolpert argues that Tabin’s new model doesn’t explain this result, and that Ros’s graft study actually backs it up. “Her findings are exactly what the [PZ] model predicts would happen,” he insists.
Tabin disagrees. He says the results merely reflect how far the wave of pre-patterning has travelled when the X-rays are turned on. If upper arm structures have already been patterned onto cells at this stage, they cannot be reconstituted by the surviving cells after irradiation. Fingers, though, are pre-patterned last so they sometimes survive.
This point is a key battleground between the old guard and the Young Turks. “Tabin’s explanation [of the irradiation experiments] is a bit weak,” says John Saunders, an embryologist at the Marine Biological Laboratory in Woods Hole, Massachusetts, who first described the apical ectodermal ridge more than four decades ago. Saunders says he’d like to see some evidence of genetic differences between cells in the early limb bud before he’s convinced. “This would definitely strengthen Tabin’s argument that the limb is pre-patterned,” says Saunders.
Other embryologists take a cautious view, too. Cheryll Tickle of the University of Dundee, who worked on the PZ model with Wolpert in the 1970s, says Tabin needs to explain the mechanism of pre-patterning. But she admits that the PZ model also has problems. For one, there is still no evidence for an internal clock. It has been suggested that cells may count the number of times they divide, or that an “oscillator” molecule somehow marks time, but no one knows for sure how cells measure the duration of their stay in the zone.
If, in the end, the PZ model needs to be discarded, it will be a shock to everyone in developmental biology, says Malcolm Logan of the National Institute for Medical Research in London. But he thinks the debate is overdue. “These results will stir up much-needed research,” he says. “The debate will finally force us to design a test to prove the PZ model right or wrong, once and for all.”
Already, alternative ideas are cropping up. One comes from Miguel Torres, a biologist at Spain’s National Centre for Biotechnology in Madrid, who proposes that a dual signalling system determines patterning: FGF from the apical ectodermal ridge guides the development of the lower limb, while retinoic acid from the body wall signals the upper bones to grow. Even a year ago, theories challenging the PZ model seemed like heresy. Now it’s open season.
