Ellen Heber-Katz, a scientist at The Wistar Institute in Philadelphia, used to study autoimmunity—that was until she noticed something strange in the mice she was using to model lupus: The small holes that she had poked in their ears to distinguish the animals from one another kept closing. At first she thought her postdoc, Lise Clark, had forgotten to make the holes in the first place. But Clark clearly remembered doing it. Together, Heber-Katz and Clark pierced new holes. Within days, they closed, too. “Every day they got smaller and smaller and then just disappeared,” Heber-Katz says. And, there was no scar—the tissue was perfect. They wondered: “If we could find out what it was that was creating this response, we could treat wounds that way!”

That was in 1995. Heber-Katz couldn’t wait to study the phenomenon in depth. But her colleagues counseled against it, urging her not to waste her expertise in the field of autoimmune disease and transplant rejection. They advised her to pursue the disappearing ear holes as “an aside,” like a hobby. But Heber-Katz knew she had stumbled onto something big, and she just had to go after it full force. “I realized, since I didn't know anything about wound healing, I had better go to a meeting about it,” she says. So she did, and there an expert told her: “Oh no, mouse ear holes absolutely do not close.” So Heber-Katz kept her finding secret. “I was really on cloud nine,” she says.

Some mammalian organs, like the liver, are capable of regeneration. Others, like the heart and the brain are unable to replace lost or injured cells. When tissue can’t regenerate, it is replaced with scar tissue. In contrast, salamanders and worms can regenerate lost or damaged tissue, leaving little trace, if any, of an injury. This appeared to be exactly what was happening in Heber-Katz’s mice.

To understand what was going on, Heber-Katz, started examining the cells filling in the holes in culture. When she compared the “healer” cells from the lupus mice with “nonhealer” cells from normal mice, she noticed several differences. Healer cells were dedifferentiated—meaning they had become like immature cells that had not yet decided what type of cell they were. They were also multinucleated—a characteristic of cells about to divide. Furthermore, they expressed a whole host of stem cell markers. What they did not express compared to their nonhealing counterparts was a single gene: p21.

To test whether p21 could be regulating healing, Heber-Katz got her hands on some mice genetically engineered to lack the gene. She punched little holes in their ears and waited (this procedure is thought to cause no pain in the mice). Sure enough, the holes disappeared. When Heber-Katz looked at the filled-in holes under the microscope, she noticed a breakdown in the basement membrane—the thin layer of cells that connects the dermis (the cushiony subcutaneous connective tissue) to the epidermis (the skin's outer layer). Instead of an organized basement membrane, there was a blastema— a circular arrangement of highly proliferative, undifferentiated cells that grew until the tissue was replaced—without scarring.

On March 15—15 years after Heber-Katz first noticed the disappearing ear holes—the study was published in Proceedings of the National Academy of Sciences. Heber-Katz is glad she followed the serendipitous lead. “I think it has really paid off,” she says. “Every day we came into the lab and found something new with these mice, it was just the most incredible adventure.” Heber-Katz says it was hard to get funding to study something new, especially because her background was in a different field. “I'm really glad we committed ourselves to doing this, but it hasn't been easy,” she says.

The study raises many new questions: What does p21 regulate? And do cells dedifferentiate (retreat to a stem cell–like state) without it? Reducing the expression of p21 and its upstream regulator p53 (a tumor suppressor) was recently found to enhance the production of induced pluripotent stem cells—a previously mature cell reprogrammed to act like a stem cell. “The most exciting thing is this is really the first gene we can point at than can be used for therapy,” she says.

Heber-Katz and her team plan to study different ways of blocking p21 to enhance healing and to regenerate tissue. Other efforts to regenerate human tissue have been in the works for more than a decade, but those primarily rely on growing stem cells from a subject on a synthetic scaffold in the presence of growth- and differentiation-promoting chemicals, and then placing them back in the body. Heber-Katz’s approach would involve temporarily silencing p21 expression in the wound itself. But because genes can regulate multiple processes, more work is needed to determine the safety of modifying the cell cycle in such a way.

The case of the disappearing ear holes is a perfect example of how easy it is to miss exciting leads when you’re not looking for them. Heber-Katz says another scientist working with the mouse model of lupus also noticed the closing ear holes, but thought of it as an inconvenience rather than a eureka moment, saying, “These darn mice, you have to keep going upstairs to punch holes in their ears!” Heber-Katz says she understands how small, unexpected findings often go undocumented. “It's hard enough to focus on one thing,” she says. "I think people saw it but they didn't really watch it, they were focused on something else.”

Some very significant medical discoveries were made by accident—penicillin, cisplatin (a chemotherapeutic drug), and even Viagra. Just think of all the flukes—and all the revelations they might impart—that continue to go unnoticed.