A bacterium taken from the gut of a Japanese tree frog (Dryophytes japonicus) cleared every tumor in a group of mice after a single injection into the bloodstream, a complete response in 100 percent of the treated animals. In the same experiment, an anti-PD-L1 antibody, a type of immune checkpoint inhibitor, and liposomal doxorubicin, a standard chemotherapy drug, ran as controls, and neither matched the result. The team, led by Eijiro Miyako at the Japan Advanced Institute of Science and Technology, published the finding in the journal Gut Microbes on December 10, 2025.
The frog itself is unremarkable to look at: bright green skin that can shift to brown or gray depending on its surroundings, a dark spot below each eye, and a body about 1.2 to 1.6 inches (3 to 4 centimeters) from snout to vent. Miyako’s team was not studying the frog. They were studying what lives inside it. They isolated 45 bacterial strains from the intestines of Japanese tree frogs, Japanese fire belly newts (Cynops pyrrhogaster), and Japanese grass lizards (Takydromus tachydromoides), screened all 45 for antitumor activity, found nine that had it, and watched one strain from the tree frog’s gut, Ewingella americana, come out as the most effective of the nine.
How a Gut Microbe Finds a Tumor
Ewingella americana is a facultative anaerobe, meaning it can grow with oxygen or without it, a trick most bacteria can’t manage. Tumors are chronically starved of oxygen: they outgrow the blood vessels feeding them, leaving a hypoxic core most organisms find useless. The bacterium found that core very useful. Within 24 hours of the injection, bacterial counts inside the tumors had risen roughly 3,000-fold, while healthy, well-oxygenated organs stayed clear.
Once established, the bacteria did two jobs. They killed cancer cells directly. And their presence acted as an alarm, pulling in T cells, B cells, and neutrophils, and raising inflammatory signaling molecules including TNF-alpha and IFN-gamma, which pushed cancer cells into programmed cell death. The tumor’s own hiding place had turned into the one part of the body actively working against it.
Why This Isn’t Just Luck
None of this reads as luck once you lay out what the bacterium was actually reading. Tumors are hypoxic, they build leaky, disorganized blood vessels that are easier to slip into than normal ones, they raise a chemical shield using the CD47 “don’t eat me” signal that normally keeps immune cells off cancer cells, and they produce their own distinct metabolic byproducts. A gut bacterium that evolved to survive wherever oxygen runs low and the local chemistry turns strange had, without meaning to, built a detector for exactly that combination of conditions.
The researchers’ broader reasoning for looking here in the first place is worth sitting with: spontaneous tumors are considered rare in amphibians and reptiles, despite lives full of physiological stress and constant microbial exposure, which is what led the team to ask whether gut microbes might be doing some of the protecting.
The safety numbers come from the same paper. The bacterium’s half-life in blood was about 1.2 hours, and it had vanished from the bloodstream entirely by 24 hours. No colonization turned up in the liver, spleen, lungs, kidneys, or heart. The only cost was mild, temporary inflammation that resolved within 72 hours, with no chronic toxicity over 60 days of observation.
The Myth This Kills
Most gut-microbiome cancer research trains people to expect a slow, indirect story: feed the microbiome, transplant stool, wait for a bacterial population to shift and hope disease responds somewhere downstream. This is not that. The researchers isolated a single strain, cultured it, and injected it straight into the bloodstream, where it did the work directly and then cleared out within hours, colonizing nothing. It also answers the standing worry attached to any bacterial cancer therapy, that putting a live organism into the blood risks an uncontrolled infection. Here, the organism’s disappearance was tracked and reported as carefully as its tumor-killing.
What Nobody Knows Yet
This is one study, one mouse model, one cancer type: colorectal. Whether Ewingella americana does anything against breast cancer, pancreatic cancer, or melanoma is untested, though the team has named those as directions for future work. Nobody has worked out the best way to deliver it, whether it is safe or effective alongside existing chemotherapy or immunotherapy, or how any of this maps onto a human immune system and a human tumor, neither of which behaves exactly like a mouse’s. There is even a loose end in the frog itself: the tree frogs sampled were classified as Dryophytes japonicus, but a 2025 taxonomic split carved a genetically distinct sister species, Dryophytes leopardus, out of the eastern populations once grouped under that name. Which population actually supplied the gut samples is a detail still being sorted out. The researchers themselves describe the whole result as a proof of concept, not a treatment.
What You Noticed
Most people who meet a Japanese tree frog clock it as small, green, harmless, and move on. This one was carrying, in its own intestines, a bacterium that erased tumors in a lab mouse with a single injection. Nothing about the frog’s life selected for that. It was just sitting there, in the gut of an animal nobody thought to check, waiting for someone to look.