Cut the protein in a mouse's food and something strange happens inside its fat. Patches of ordinary white fat start switching on the same genes that fire during cold exposure, when the body scrambles to generate heat. The fat begins to burn energy instead of just storing it. And according to a study published in Nature on 4 March 2026, the animal cannot pull off this trick alone. It needs its gut bacteria.
The finding comes from a large collaboration led by researchers at Keio University in Tokyo and the Broad Institute in Cambridge. They were chasing a basic question that has nagged at the field for years. Diet and the gut microbiome both shape metabolism, and they clearly influence each other, but exactly where those two forces meet inside the body has been hard to pin down.
When white fat starts acting like a furnace
Fat is not one thing. White adipose tissue hoards calories. Brown and beige fat do the opposite, spending energy to produce heat. Turning white fat beige, a process called browning, is a long-standing target for metabolism research because it could in principle help the body offload excess energy.
The team found that a low-protein diet triggered browning genes in white fat to roughly the same degree as classic stimuli like cold or drugs that mimic adrenaline signaling. That is a strong effect from a simple dietary change. But when they ran the same experiment in germ-free mice, animals raised with no microbes at all, the browning largely failed to appear. The diet was giving the instruction. Something in the gut was needed to carry it out.
To prove bacteria were responsible, the researchers assembled defined groups of strains and put them back into germ-free mice. Some strains came from the feces of low-protein-fed mice. Others came from healthy human volunteers whose brown and beige fat activity had been confirmed by PET imaging. Colonizing the animals with these bacterial consortia restored the browning response. The microbes were not bystanders. They were part of the machinery.
Two chemical messages, two organs
What the bacteria actually do turned out to involve two separate routes, and both matter. In the first, gut microbes shape the pool of bile acids in a way that activates a receptor called FXR inside adipose progenitor cells, the precursors that give rise to fat. That signal nudges those cells toward the beige, energy-burning fate.
The second route runs through the liver. Certain commensal bacteria carry a gene called nrfA and produce ammonia as they metabolize. That ammonia drives liver cells to make FGF21, a hormone with well-documented effects on metabolism. Knock out either the bile-acid-FXR arm or the ammonia-FGF21 arm and browning suffers. The two pathways are not backups for each other. Each does its own job, and the host needs both to respond fully to the low-protein signal.
It is a tidy picture of a body outsourcing part of its metabolic decision-making to the organisms living in its gut. The diet sets the conditions, the bacteria translate those conditions into chemical signals, and distant tissues like fat and liver read the signals and act.
What this does and does not tell us
The obvious caution is that this is a mouse study. The human element here is limited to bacterial strains sampled from people, not to any test of what a low-protein diet does to human fat. Mouse and human metabolism differ in ways that have humbled plenty of promising leads, and browning is notoriously easier to trigger in a small rodent than in a person. Chronic protein restriction also carries real risks in humans, so no one should read this as dietary advice.
What the work does offer is a concrete mechanism, spelled out down to specific bacterial genes and host receptors. That specificity is useful. If the bile-acid-FXR and ammonia-FGF21 axes hold up in further study, they point to defined targets rather than the vague hope that a healthier microbiome is somehow good for the waistline. The researchers isolated the strains, mapped the signals, and traced them to the tissues that respond. Whether any of it can be nudged in people is the next question, and this study does not answer it. It just draws a clearer map of where to look.
Comments