Texas A&M researchers show in animal amputation models that two growth factor signals can redirect wound healing from scar tissue to regrowth of bone, joint, ligament, and tendon, a reframe of mammalian healing rather than a finished human therapy.
For most mammals, a serious wound ends in a scar. That is the default. Texas A&M University researchers now report in an animal amputation model that this default can be steered, briefly and pharmacologically, toward something that looks more like rebuilding: bone, joint, ligament, and tendon regrowing where a limb was lost. The work, summarized by ScienceDaily on June 17, 2026, does not regrow a working limb. What it does is harder to dismiss. It suggests that mammals were never missing a regenerative program at all. They were suppressing one.
The reason this matters is the framing it replaces. For decades, the standard story has been that humans and other mammals lost the ability to regrow complex tissue the way salamanders regrow tails and zebrafish regrow fins, an evolutionary trade that scar-based healing more than makes up for. The Texas A&M study, described in the ScienceDaily release as preclinical animal work, instead argues that the program is still there, just switched off by the same wound response that closes the injury in the first place. If that reading holds, the question stops being "can we build a new regenerative biology?" and becomes "can we find the right switch and the right delivery to release the one mammals already have?"
The intervention itself is modest, which is part of the point. The team applied two growth-factor signals, BMP2 (a bone-shaping protein) and FGF2 (a fibroblast growth factor that drives tissue patterning), to an amputation site in a two-step protocol. The first step, the BMP2 signal, lays a bone-like template at the wound. The second step, the FGF2 signal, recruits the connective-tissue machinery needed to build a working joint and the ligaments and tendons that connect it. In the animal study, this redirected the wound's default scar response, the dense collagen plug mammals use to seal injuries quickly, and replaced it with a structured regrowth of multiple tissue types at the amputation site. No gene editing. No stem-cell transplant. Just two protein signals delivered into a wound at the right moment.
Salamanders are still the comparison the reader deserves, but they belong in one paragraph, not the lede. A salamander's stump forms a blastema, a mass of dedifferentiated cells that essentially redraws the missing limb from scratch. Mammals do not make a blastema. What the Texas A&M protocol appears to do is something narrower and arguably more useful: rather than rewind cellular identity, it tilts the existing mammalian healing cascade away from fibrosis (the technical name for scar formation) and toward the multi-tissue rebuilding that already exists in a mammal's toolkit. That distinction is the article. A blastema would be a new technology. A re-routed wound response is an instruction set the body can already read.
The limits of the claim are real, and they live in the same paragraph as the result. ScienceDaily describes the work as a preclinical proof of concept in animal models, not a human therapy. The regrown bone, joint, ligament, and tendon are not identical copies of what was lost, and the study has not yet been independently replicated. The protocol depends on delivering the right growth factors at the right time after injury, which is its own unsolved delivery problem in a human patient. None of this erases the result. It just sets the scale: this is the kind of finding that expands the research surface, not the kind that lands in a clinic next year.
That expanded surface is the constructive read. A reframe from "regeneration is a lost superpower" to "regeneration is a suppressed default" is something many more labs, and eventually many more clinical problems, can act on. Scarring is not only a wound-healing issue. It is a problem in every organ that fibroses after injury: skin, heart, liver, lung, kidney. Joint injury and tendon repair are two of the most common unmet needs in musculoskeletal medicine. If the same two-signal logic can be tuned for those tissues, the result is a generation of work, not a single product. The Texas A&M team has not solved that. They have moved it from a speculative moonshot to a defined experimental program, which is the most that good biology usually delivers in one paper.
What to watch next is the underlying study, which the ScienceDaily summary is based on. The peer-reviewed paper, reported as published in Nature Communications by the Texas A&M College of Veterinary Medicine and Biomedical Sciences group, will need to confirm which animal model was used, how many animals were studied, which limb segment was restored, and how completely the regrown joint functioned under load. Until those details are public, the headline result is the protocol, not the promise.