In the News

November 1, 1997


Starting Over

Some animals can regenerate limbs or even most of their bodies. How?


Last summer, on the city's hottest day in a century, a construction platform beneath a Washington, D.C., bridge collapsed, dropping a 15-ton steel beam that killed one man and pinned a second by his legs. With a blowtorch, rescue personnel eventually freed the injured man and rushed him to a hospital, where physicians had to amputate his left leg and transplant a back muscle into his right leg.

Such misfortunes have long prompted physicians to dream of restoring damaged or amputated limbs by inducing them to regenerate. Other injuries have spurred doctors to imagine regrowing crushed spinal cords and dead heart tissue.

These medical flights of fancy draw inspiration from the well-known ability of certain animals to perform precisely those feats. Consider the urodeles, a class of vertebrates that includes newts and salamanders. These animals possess an enviable talent for regrowing arms, legs, tails, heart muscle, jaws, spinal cords, and more. Some simpler organisms can even be sliced and diced, with each piece giving rise to a complete new animal.

Surprisingly, given the obvious medical appeal of regeneration, scientists know relatively little about the process. Researchers have produced detailed descriptions of regeneration, but they don't understand the molecular signals driving this physiological tour de force.

The problem lies largely in the animals that scientists have traditionally chosen to study, such as mice, frogs, fruit flies, and nematodes. Though investigators have amassed a wealth of knowledge about these common laboratory animals and have many ways to study and manipulate them genetically, the adults of these species don't come close to matching a urodele's regenerative powers.

While most scientists have tried to overcome the mysterious barriers to regeneration that exist in the traditional laboratory animals, a few investigators have taken the road less traveled. They have decided that to unearth the genes and proteins vital to regeneration, they must study animals that can actually regenerate.

"We're trying to understand the molecular basis of regeneration," says Alejandro Sánchez of the Carnegie Institution of Washington in Baltimore.

The basics of limb regeneration have been evident for more than a century. First, the animal heals the wound at the site of the missing limb. Then, various specialized cells at the site, such as bone, skin, and blood cells, lose their identity in a process called dedifferentiation. The resulting blastema, a mass of unspecialized cells, proliferates rapidly to form a limb bud. The cells ultimately take on specialized roles as the new limb takes shape.

"Salamanders can turn back time," says David L. Stocum of Indiana University-Purdue University in Indianapolis. "The trick they have is dedifferentiation of mature cells. They regenerate a lot of tissues this way."

For some reason, people and other non-urodele vertebrates lack this ability to create a blastema. They repair a wound and stop. Rather than rely on dedifferentiation, the few human tissues that can regenerate -- such as blood and the liver -- turn to a small number of unspecialized cells set aside during embryogenesis. These so-called stem cells maintain the ability to proliferate rapidly and indefinitely.

In recent years, Susan V. Bryant and David M. Gardiner, both at the University of California, Irvine, have turned to salamanders called axolotls to study how a blastema transforms itself into a limb. The process depends on many of the same genes employed when an embryo originally creates a limb, but the blastema doesn't turn those genes on and off in exactly the same pattern.

Take the HoxA genes. From developmental studies of many animals, scientists have shown that this cluster of genes helps pattern a growing limb. Which HoxA genes in a cell turn on ultimately determines the cell's position in the final appendage.

Remarkably, the order of the HoxA genes on their chromosome mirrors the genes' order of activation. In the developing limb bud of the arm, for example, HoxA genes at one end of the chromosome turn on first, marking cells intended for the upper arm. As the bud grows, successive HoxA genes become active, identifying cells destined to form the lower arm. Finally, the HoxA genes farthest down the chromosome take their turn, signaling which cells will become part of the hand.

Bryant and Gardiner have found that this Hox code exists in regenerating axolotl limbs, but with a twist: The timing of gene activation does not follow the genes' order on the chromosome. For example, after a limb is amputated, both the HoxA-9 and HoxA-13 genes become active a day or two into the 3-week regenerative process. In the embryo, HoxA-13 follows HoxA-9 and isn't turned on until the very end of limb patterning.

The blastema first specifies which cells will form the axolotl "hand," concludes Bryant. "Then, it's a process of filling in the gap. That allows you to make exactly what's missing." Studies of another developmental gene cluster, the HoxD genes, bolster this view, she adds.

Axolotls have shed light on another curious difference between developing and regenerating limbs. In the embryo, limbs take shape well before nerves arise. Yet a denervated axolotl limb can't initiate the process of regeneration, says Bryant. Moreover, scientists can denervate a regrowing limb midway through regeneration without stopping it, suggesting that there's a window of time in which regeneration is dependent on the nerves.

Investigators have speculated that nerves release some vital regeneration factor, but its identity has proved elusive. Recently, Bryant, Gardiner, and their colleagues found the axolotl version of a fruit fly gene called distalless, or dll. In flies, the gene is needed for proper leg development.

This gene turns on in a regenerating axolotl limb just as the limb loses its dependence on nerves for regrowth, the investigators reported in the November 1996 Development.

They also discovered that denervating a regenerating limb before this transition prevents dll from turning on, whereas later denervation does not turn off the gene. Suspecting that limb regeneration requires dll activity, the investigators sought to determine why the gene fails to turn on after early denervation.

They eventually hit upon a protein called FGF-2. When they implanted beads coated with FGF-2 into a regenerating limb, dll activity stayed normal after early denervation. The scientists further established that nerves indeed make FGF-2.

While a limb blastema can produce its own FGF-2, the scientists believe it needs an initial supply from nerves before it becomes self-sufficient. "We think the nerves provide the FGF-2 to prime that transition," says Bryant.

Consequently, an absence of nerve-derived FGF-2 may be one of the barriers to limb regeneration in most vertebrates. Triggering an amputated arm to regrow will not be simply a matter of providing this growth factor, but FGF-2 may provide part of the eventual recipe for regeneration in people.

"The hope is that there won't be too many things to replace and that you can get to the point where the [regeneration] cascade begins to roll," says Bryant.

Sánchez loves axolotls. In fact, he keeps two as pets in an office aquarium. Yet Sánchez believes the salamanders will keep secret most of their regenerative tricks.

The inability to mutate specific genes in axolotls and other urodeles, or to add specific genes to the animals, hinders attempts to determine the significance of any new gene that researchers suspect of aiding regeneration.

Though the work of Bryant and her colleagues is "heroic," says Sánchez, "you reach a ceiling of how much you can do with the organisms."

"The genetics [of urodeles] is poor, their generation time is long, and there are no transgenics yet. That's why it's a struggle for people to keep working in the area," acknowledges Bryant.

Consequently, Sánchez intends to pursue his regeneration studies primarily with planaria, simple worms that have their own legendary regenerative ability. Chop one of the inch-long creatures into 300 pieces, and in a matter of days 300 planaria will be swimming around. Indeed, S·nchez' group occasionally uses that strategy to expand the worm colonies.

Sánchez' plan is to identify the genes that enable planaria to regenerate. Why? "Most likely we have those genes as well," he says.

That statement would have sounded brash 20 years ago. Yet scientists studying worms, primarily those called Caenorhabditis elegans, have increasingly found that genes from these simple animals resemble genes in more complex creatures. In addition, the related genes are frequently used in a similar manner.

As a result, Sánchez is operating on the assumption that the regeneration strategy used by planaria resembles the one employed by urodeles and may provide insight into the human failure to regenerate. There's some evidence supporting this contention. When cut in half, the worms form blastemas similar in architecture to those seen when researchers amputate the limbs of a urodele.

Planarian regeneration does have clear differences, however. The most significant seems to be that the worms do not dedifferentiate cells to create the blastema. Instead, planaria turn to cells called neoblasts.

Scattered within the planarian body, neoblasts apparently remain in an unspecialized, stem cell state, which enables them to differentiate into any cell type. Wherever planaria are cut, the neoblasts migrate to the site and form a blastema by themselves.

Sánchez' colleague Phillip Newmark has kept planarian neoblasts alive in petri dishes for several weeks, an advance the researchers hope will allow them to genetically alter the animals. The investigators will try to slip genes into the neoblasts or mutate existing genes, then add the cells to worms whose own neoblasts have been destroyed with radiation. When such worms regenerate, any new cells should derive from the genetically engineered neoblasts, says Sánchez.

Efforts to identify planarian regeneration genes have already offered some promising leads, Sánchez' group reported at last summer's International Congress for Developmental Biology in Snowbird, Utah. With a technique called subtractive hybridization, the researchers collected fragments of genes turned on in blastemas when a planarian regenerates its tail, its head, or both and compared them to genes normally active in the head and tail. Genes active only in the blastemas presumably participate in regeneration, notes Sánchez.

The investigators detected dozens of DNA fragments specific to regenerating blastemas, some belonging to genes that become active in the initial hours of the process. In several cases, the researchers have found that the fragments are parts of genes found previously in other animals.

One such gene encodes an enzyme that degrades the extracellular matrix, a mesh of proteins and other molecules that surrounds cells. This enzyme may trigger regeneration by releasing growth factors bound to the matrix or by eliminating obstacles to a cell's proliferation and movement, says Sánchez.

Is the notion of regenerating human limbs or diseased tissue a pipe dream or a realistic expectation for the 21st century? It's far too early to know whether there are insurmountable differences between animals that can regenerate and those that cannot, says Jeremy P. Brockes of University College London.

He and his colleagues have recently shown that some factor present in blood serum, newt or any other kind, can induce newt muscle cells to dedifferentiate. The muscle cells of nonregenerating vertebrates don't respond to the serum, however, indicating that they are not sensitive to this still unidentified factor.

If feasible, overcoming this insensitivity might prove invaluable. Cardiac specialists might then learn to trigger damaged heart muscle to replace itself. Such an advance could be possible even if scientists never obtain sufficient expertise to regenerate complex structures like limbs, suggests Brockes.

Given the potential value of regeneration in medicine, Stocum finds it puzzling that so many scientists choose to study cell and organ transplantation or work on building artificial hearts, livers, and limbs rather than investigate animals that can regrow parts of themselves.

"With regeneration, you don't have to worry about [finding] donor cells, and you don't have to worry about immune rejection," he says. "Regeneration is clearly superior to any kind of transplantation or artificial tissues."

To link to an animated version of limb regeneration, visit



Mullen, L.M., et al. 1996. Nerve dependency of regeneration: The role of Distal-less and FGF signaling in amphibian limb regeneration. Development 122(November):3487.

An animated illustration of limb regeneration is available at

Further Readings:


Brockes, J.P. 1997. Amphibian limb regeneration: Rebuilding a complex structure. Science 276(April 4):81.

Dinsmore, C.E. 1996. Urodele limb and tail regeneration in early biological thought: An essay on scientific controversy and social change. International Journal of Developmental Biology 40:621.

Stocum, D.L. 1997. New tissues from old. Science 276(April 4):15.



Susan V. Bryant
Developmental Biology Center
Department of Developmental and Cell Biology
University of California
Irvine, CA 92697-2274

Charles E. Dinsmore
Department of Anatomy
Rush Medical College
Rush University
524 Academic Facility
600 S. Paulina Street
Chicago, IL 60612

Alejandro Sánchez
Carnegie Institution of Washington
Department of Embryology
115 W. University Parkway
Baltimore, MD 21210

David L. Stocum
Department of Biology
School of Science
Indiana University-Purdue University
355 N. Lansing Street
Indianapolis, IN 46202

copyright 1997 Science Service