In 2012, a team at Tufts University led by biologist Michael Levin grew eyes in the wrong place. Working with Xenopus laevis tadpoles, Levin, along with Vaibhav Pai and Dany Adams, injected mRNA into cells in the gut and tail region, cells with no developmental business ever becoming part of a visual system, and used it to shift those cells’ membrane voltage. Paired with a dose of Pax6, the gene long associated with eye development across animal species, the voltage change was enough. Functional eye structures, with lens and retinal tissue, grew on the tadpoles’ tails and bellies, built from cells that had never been anywhere near a developing head.
The striking part wasn’t Pax6. Biologists already knew Pax6 could nudge tissue toward eye fate. The striking part was that Pax6 alone, without the voltage shift, produced ectopic eyes far less reliably. The cells needed to be told, electrically, that they occupied a position where an eye was plausible, before the genetic instructions could take hold in a way that built an actual organ. Voltage came first. Gene expression followed it. That ordering is the part that still unsettles researchers who built their careers on the assumption that DNA runs the whole show.
What Levin’s lab was studying has a name: bioelectric signaling, the use of voltage differences across cell membranes and between groups of cells to carry positional and developmental information. It’s not a new discovery so much as a rediscovered one, and it’s reshaping how researchers think about growth, wound healing, and what regenerative medicine might eventually be able to do that gene editing alone cannot.
Every Cell Runs on Voltage, Not Just the Ones in Your Brain
You already know that neurons fire electrical signals. What’s less widely understood is that nearly every cell in your body, skin cells, liver cells, the cells lining your gut, maintains a voltage across its membrane at all times, generated by pumps and channels that control the flow of sodium, potassium, calcium, and chloride ions in and out of the cell. This resting membrane potential typically sits somewhere between negative 20 and negative 90 millivolts relative to the outside of the cell, depending on cell type. It’s not a byproduct of metabolism. It’s a maintained, regulated state, and cells spend a meaningful fraction of their energy budget keeping it where it needs to be.
The difference between this and a neuron firing an action potential is speed and purpose. A neural spike is a millisecond event that carries a single bit of information down an axon. The bioelectric patterns Levin’s lab and others study are slow by comparison, stable over hours, days, or weeks, and they exist as spatial maps spread across groups of cells rather than single-point signals. Harold Saxton Burr, a Yale anatomist working in the 1930s and 1940s, was among the first to measure these standing electrical fields around developing embryos and injured tissue, calling them “L-fields,” decades before anyone had the molecular tools to explain what generated them. Lionel Jaffe, working at Purdue and later the Marine Biological Laboratory in Woods Hole, picked up the thread in the 1970s, measuring ionic currents flowing through developing eggs and showing those currents predicted where an embryo’s axis would form.
Every cell is a battery. Most of them never fire, they just hold a charge and wait.
What Levin’s generation of researchers added was the ability to deliberately rewrite those voltage maps, using ion channel drugs, optogenetics, or injected mRNA, and then watch what the tissue built next.
The Flatworm That Keeps Growing Two Heads
Planarian flatworms are a favorite subject in this field because they regenerate almost any part of their body from a fragment, and they do it fast enough to run repeated experiments on the same lineage. In a 2012 study, Levin and Tufts colleague Taisaku Nogi used a gap junction blocker called octanol to briefly disrupt the electrical connections between cells near the wound site of a cut flatworm. Gap junctions are small channels that let neighboring cells share ions and small molecules directly, effectively wiring groups of cells into a shared electrical domain. Block them at the right moment, and the worm’s cells lose track of which end used to be the head.
The result: worms that regenerated two heads instead of one head and one tail. That alone wasn’t shocking, prior work had produced two-headed worms through other means. What made the experiment matter was what happened next. Researchers cut these two-headed worms again, with no further chemical treatment, and they kept regenerating as two-headed animals, over multiple rounds of amputation. The flatworm’s genome hadn’t changed. Its cells had the same DNA as any normal, one-headed planarian. What had changed, and what persisted, was the bioelectric pattern encoding where a head belonged.
The worm’s genome never changed. Its voltage map did, and the anatomy followed the map.
That’s the conceptual hinge of the whole field. Genes supply the parts list. The bioelectric state of a tissue appears to supply something closer to a blueprint, a pattern that tells identical genomic machinery where to put a head, a tail, or in the Xenopus case, an eye.
How a Voltage Difference Actually Reaches a Gene
The mechanism connecting membrane voltage to gene expression runs through chemistry, not magic. Changes in a cell’s voltage alter which ions and small molecules flow through its gap junctions and channels, including calcium, serotonin, and butyrate, each of which has downstream effects on intracellular signaling cascades. Calcium influx, for instance, can activate transcription factors directly. Voltage also affects the activity of voltage-sensitive phosphatases, enzymes that add or remove phosphate groups from other proteins and, in doing so, switch entire signaling pathways like Notch, Wnt, and BMP on or off. Those are the same developmental pathways geneticists already study closely. Bioelectric signaling doesn’t bypass genetics, it sits upstream of it, acting as a coordination layer that tells groups of cells which genetic programs to run and where.
This is also why gap junctions matter so much in this story. They let voltage states synchronize across many cells at once, turning a patch of tissue into something closer to a single electrical unit with a shared instruction set, rather than thousands of cells each guessing independently what they should become. Researchers sometimes describe these patches as bioelectric circuits, borrowing the language of electronics, though the comparison is loose. Nothing here is being clocked or computed in any formal sense. It’s closer to a slow-moving chemical gradient that happens to be carried by charge instead of by diffusion alone.
A Frog Leg That Grew Back After One Day of Treatment
The clearest recent demonstration of practical stakes came out of Tufts and Harvard’s Wyss Institute in January 2022. Nirosha Murugan, working with Levin’s lab, built a small wearable device, essentially a silicone cap containing a five-drug cocktail, and fitted it onto the amputated hind legs of adult Xenopus laevis frogs. Adult frogs, unlike tadpoles or axolotls, don’t normally regenerate lost limbs. They heal over with a scar and stop. The device delivered its drug combination, targeting ion channels, inflammation, and serotonin signaling, for just 24 hours before being removed.
Over the following eighteen months, the treated frogs regrew a paddle-like appendage containing new bone, muscle, and nerve tissue, enough structure to restore some swimming function, from a single day of bioelectric intervention. No gene editing was involved. No stem cells were transplanted. The treatment worked by resetting the local voltage and signaling environment of the wound long enough for the tissue’s existing developmental program, apparently still present in frog genomes despite frogs having lost the habit of using it, to pick up where it normally stops. Levin later co-founded a company, Morphoceuticals, built specifically around pursuing this approach toward mammalian tissue.
| System | Bioelectric Intervention | Researchers / Institution | Result |
|---|---|---|---|
| Planarian flatworms | Gap junction blocker (octanol) at wound site | Nogi and Levin, Tufts, 2012 | Worms regenerated two heads, trait persisted through repeated cutting |
| Xenopus tadpoles | Ion channel mRNA plus Pax6 in gut/tail cells | Pai, Adams, Levin, Tufts, 2012 | Functional ectopic eyes grown outside the head |
| Adult Xenopus laevis limbs | 24-hour wearable five-drug bioreactor cap | Murugan, Levin, Tufts/Wyss Institute, 2022 | Partial limb regrowth with bone, muscle, nerve over 18 months |
| Human skin wounds | Measurement of endogenous injury current | Colin McCaig, University of Aberdeen | Electric field shown to guide keratinocyte migration toward wound |
A Frog Leg Is Not a Human Arm
Here’s where the honest skepticism has to come in. Every striking result described above happened in a frog or a flatworm, organisms with far simpler immune systems, far less aggressive scarring responses, and in the planarian’s case, a body plan that’s already built for constant regeneration. Mammals heal wounds by closing them with scar tissue fast, which is useful for survival and terrible for anyone hoping to coax a mammalian limb into regrowing. Nobody has regrown a human limb by changing its voltage. Not yet, and not for lack of people trying.
Nobody has regrown a human limb by changing its voltage. Not yet, and not for lack of trying.
There’s also a genuine scientific debate underneath all of this about causation. In some systems, voltage changes clearly precede and direct gene expression, as in the Xenopus eye work. In others, critics have pointed out that voltage shifts can be downstream consequences of gene expression changes rather than the trigger for them, and untangling which direction causation runs in a given tissue takes careful, slow experimental work that isn’t always done. A few of the field’s more dramatic claims, including some planarian results, have proven harder for outside labs to replicate exactly, which is a normal part of science working as intended but worth saying plainly rather than glossing over.
The bioelectric medicine that already exists in clinics today is far more modest than regrowing organs, and it’s worth distinguishing it from the Levin lab’s ambitions. Vomaris Wound Care’s Procellera dressing, cleared by the FDA in 2009, embeds a grid of tiny silver and zinc dots that generate a weak electric field across a wound when moistened, speeding closure and reducing infection risk. That’s bioelectricity doing a narrower job: nudging a healing process the body already knows how to run, not instructing it to build something new.
Where the Near-Term Payoff Actually Sits
The more plausible near-term use of this science probably isn’t growing new limbs on people. It’s reading and gently rewriting the voltage maps already running processes your body performs constantly: closing wounds, remodeling bone, managing the border between healthy and cancerous tissue, since depolarized membrane states have been linked to tumor cell behavior in multiple cancer types. Those are jobs the tissue already knows how to do, where a bioelectric nudge has to finish a sentence the cell already started rather than teach it a new language from nothing. The frog leg experiment is proof that the underlying biology can still be coaxed into ambitious reconstruction given the right electrical and chemical cues. Whether that same coaxing works on a creature that spent its evolutionary history optimizing for scars over structure is the open question the next decade of this research actually has to answer.
Credit: Karl Callwood on Unsplash