Malte Wagenbach — July 2026
In 2022, a team led by Nirosha Murugan, David Kaplan and Michael Levin took an adult African clawed frog, amputated its hind leg, and grew most of it back.
That sentence should stop you. Adult frogs cannot do this. After metamorphosis they lose the ability entirely — cut off a limb and they seal the wound with a useless cartilage spike and scar tissue, the same dead-end our own bodies default to. Yet these frogs regrew a functional leg: bone, blood vessels, nerves, even toe-like structures at the end. The animals could feel touch on the new limb and used it to swim.
Here is the part that reorganised how I think about medicine. The intervention lasted twenty-four hours. A small silicone cap over the stump, a hydrogel loaded with five drugs, one day. Then it came off — and the leg kept building itself for eighteen months.
Twenty-four hours of instruction. Eighteen months of construction. Nobody was steering the growth. The frog's own cells knew what a leg was supposed to look like the entire time. They just needed permission to remember.
⸻
The layer under the chemistry
Almost everything we do in biomedicine targets what Levin calls the hardware — the genome, the proteins, the biochemical pathways. Edit a gene, block a receptor, deliver a molecule. It has given us extraordinary things. But it treats the cell as a bag of chemicals and nothing more.
Levin's life work points at a second layer running on top of the chemistry: a kind of physiological software. Cells are not just chemical factories — they are tiny batteries. Every cell holds a voltage across its membrane, and cells are wired to their neighbours through channels that let those voltages talk. The result is a bioelectric network — not neurons, not a brain, but a genuine information-processing system spread across ordinary tissue.
That network carries a pattern. And the pattern is a blueprint — a stored idea of the shape the tissue is supposed to be.
This is not a metaphor he reaches for lightly. His lab has done the experiments. In planarian flatworms — the little creatures that regrow from almost any fragment — Levin's team altered the bioelectric signalling and got worms to regenerate two heads, or the head of a different species entirely, without touching a single gene. Change the voltage pattern, and the worm builds a different body from the same DNA. In frog embryos they induced eyes to grow on the gut, hearts and limbs in the wrong places — by editing the electrical state, not the genome.
The genome is the parts list. The bioelectric layer is closer to the instructions for what to assemble. And instructions, unlike hardware, can be rewritten in an afternoon.
⸻
What the frog cocktail actually did
Back to the leg. The clever thing about that experiment is what the five drugs were not trying to do. They were not micromanaging the growth. You cannot dictate where every cell goes over eighteen months — the information required is astronomical. Instead the cocktail removed the reasons the body gives up.
One drug crushed the initial inflammation, flipping the tissue from "defend" mode to "repair" mode. One suppressed the collagen machinery that lays down scar. One drove nerve growth. One provided general growth signal, and one — retinoic acid, a classic morphogen — supplied the coarse spatial "which way is which" signal for laying out a limb.
None of that grows a leg. What it does is clear the roadblocks and hand control back to the dormant developmental program the cells were carrying all along. The scar, it turns out, is not the body's best effort at healing. It is the body interrupting a better process it still knows how to run.
That reframing matters more than any single molecule. It says the information for regeneration is not lost with age. Adult cells still hold the target shape. The problem is a behavioural default — wound, panic, scar — and defaults can be overridden.
⸻
So: can we heal 100 times faster?
This is the question everyone actually wants answered. If the body runs on software, can we turn up the clock speed — heal wounds ten, a hundred times faster? I spent a while convinced the answer was a clean yes. It is not. It is two different answers to two different questions, and the gap between them is the whole story.
Ask about a single cell, and the answer is a hard no.
A cell crawling across tissue is not magic — it is soft-matter physics. It moves by building a scaffold of actin at its front edge and gripping the surface with molecular clutches. That whole system has a top speed set by chemistry we cannot cheat: how fast actin can polymerise, how fast the grip molecules bind and release, how fast ATP is burned. Push a cell to move 100 times faster and the grips simply tear off — it spins in place and goes nowhere. There is a physical throttle, and it is bolted on hard.
Division is worse. A human cell takes the better part of a day to divide, and that clock is set by the speed of copying DNA and folding chromosomes without wrecking them. Demand a cell divide every fifteen minutes and you do not get fast healing — you get shredded DNA, scrambled chromosomes, and cancer or cell death. There is even a floor on how small a cell can shrink before its own nucleus can no longer fold the genome inside. The single cell is a machine running near its physical limits already. There is no dial marked 100x.
Ask about the whole system, and the answer is an emphatic yes.
Because most real healing is not throttled by cell speed. It is throttled by everything going wrong. A diabetic foot ulcer can sit open for months or years — not because the cells are slow, but because the tissue is stuck: locked in chronic inflammation, jammed in a mechanical traffic-jam where crowded cells freeze each other in place, receiving no coherent signal about what to rebuild. The effective healing rate is close to zero. It is not slow. It is stalled.
You do not speed that up by making cells faster. You speed it up by un-sticking the process — clearing the inflammation, breaking the jam, restoring the pattern signal so the tissue knows what shape to aim for. Turn a wound that would never close into one that closes in weeks, and you have not broken any law of physics. You have removed the thing standing on the brakes. Measured against the old outcome — never — that is effectively a hundredfold gain. The speed was always there. What was missing was permission and direction.
That is the real promise. Not a faster cell. A system that stops sabotaging itself.
⸻
Steering a crowd instead of shoving it
There is a second thread here I find just as beautiful, and it is about how you deliver the signal.
Wounds carry their own natural electric field. When skin breaks, the leaking cells create a small voltage that points from the healthy edge toward the centre of the wound — a built-in "this way" arrow that guides cells inward. This has been known for a long time, and the obvious move is to amplify it: blast the whole area with current. It mostly fails. Brute-force fields fry tissue and make wound edges pull back. You are trying to command a crowd by shouting at all of it at once.
A group around Daniel Cohen and L. Mahadevan built the opposite. Their system — named SCHEEPDOG, after the way a sheepdog moves a whole flock with a few well-placed nudges at the edge — applies small, precise electric fields only at the wound's border. The stimulated cells pull, and because cells are physically linked to their neighbours, that pull propagates inward. A local nudge steers the whole sheet.
Then they found the catch, and the fix, and it is the jamming problem again. Stimulate continuously and healing surges for about three hours — then the cells pile up, jam, and the whole thing grinds to a halt. So they pulsed it. Two separated pulses instead of one steady push, giving the tissue time to mechanically relax between shoves. The double-pulse rhythm dramatically outperformed constant stimulation. You are not overpowering the tissue. You are conducting it.
There is a deep lesson buried in that. The old model of control is force: overwrite the system's behaviour with something stronger. The bioelectric model is closer to communication: find the language the tissue already speaks and give it one clean instruction at the right moment. It is the same shift I keep seeing everywhere at the frontier — from silicon that computes by imitating the brain's architecture to proteins we now design from scratch instead of borrowing from evolution. We are learning to work with the grain of living systems instead of against it.
⸻
The mammal problem
None of this matters for us unless it crosses into mammals, and here the picture is genuinely encouraging.
Some of the tricks already work. A repurposed anti-parasite drug, ivermectin, can coax ordinary skin cells in mice to take on a nerve-supporting role, driving richer nerve growth into a healing wound. A wearable "bioelectronic bandage" that pumps a common antidepressant — which happens to nudge the right ion channels as a side effect — directly into mouse wounds shifted the immune response from inflammatory toward reparative and lifted the rate of skin regrowth by roughly forty percent. These are not exotic new molecules. Many are drugs we already have, used to send a bioelectric message rather than hit a chemical target.
But there is a wall, and it is us specifically. Humans are tight-skinned mammals. Our skin is under mechanical tension, and tension drives scarring — it is why we form keloids and hypertrophic scars where looser-skinned animals do not. The African spiny mouse can regrow whole patches of skin, hair follicles and all, with no scar, in part because its skin carries far less tension.
For a while that looked like a hard biological law: high tension, no real regeneration. Then researchers looked at dolphins. Fraser's dolphin lives under constant water shear and high skin tension — and still regenerates the deep, complex architecture of its skin, blood vessels and all, from massive wounds, something no human has done since before birth. The law is not a law. Tissue can regenerate under tension. Something in the dolphin's bioelectric and mechanical signalling keeps the regenerative program running where ours quits. Decode that, and the prize is enormous: the permanent replacement of scarring with genuine regeneration in humans.
⸻
Where this is going
Levin and Kaplan spun the work into a company, Morphoceuticals, now run by veteran drug developers and backed by real venture money. Their ambition is a map they call the druggable bioelectrome — an atlas of the electrical patterns of healthy, sick and regenerating tissue — and a strategy that inverts the usual pharma playbook. Instead of building a new molecule to micromanage one gene, they take existing, often already-approved drugs that nudge ion channels, and use them to push a cell network into a regenerative subroutine it already knows. First targets are the unglamorous, brutal ones: amputee stump health, the diabetic foot ulcer. Then, further out, organ and limb regeneration in people who have lost them.
I do not know how much of this survives contact with human clinical trials. Biology is where beautiful theories go to get humbled. But the core idea has already changed how I read the whole field.
For a century we thought the body was chemistry, and healing was a matter of finding the right molecule. The bioelectric view says something else: the body is also information — a running program with a stored picture of what it is supposed to be. Injury does not erase that picture. It interrupts the program. And a program that is merely interrupted can, in principle, be resumed.
We spent a hundred years learning to read and edit life's hardware. The next hundred may be about learning to speak its software — to walk up to a wound, or a stump, or a failing organ, and simply remind the tissue what it already knows how to build.
The frog remembered its leg the whole time. It was waiting for someone to ask.
⸻
This is for informational purposes only. For medical advice or diagnosis, consult a professional.