A Metal in the Liver Sets How Well Insulin's Message Gets Through
Insulin signaling is one of the tidiest diagrams in cell biology. The hormone lands on its receptor, the receptor trips a relay of enzymes, and at the end of the chain an enzyme called Akt tells the liver to stop pouring sugar into the blood. The awkward part is that the chain does not always behave like a chain. In insulin resistance, the far end can run out of step with the near end, doing more or less than the incoming signal should cause it to do. Nobody has had a good explanation for it.
A preprint posted to bioRxiv on September 7, not yet through peer review, offers a candidate that is not another protein. Jennifer Gamarra, Rebecca Haeusler and colleagues at Columbia University report that the amount of manganese inside a liver cell is a key determinant of how well Akt works. The team includes co-authors at Baylor College of Medicine, Duke, and St. John's University. The effect runs without going through PI3K, the relay that normally carries insulin's message down the chain.
The distinction being drawn is between a switch and a dial. Manganese does not turn Akt on. It makes each molecule of Akt work harder once insulin has already switched the enzyme on, which the preprint calls an increase in Akt's catalytic efficiency. The group's earlier conference abstract, presented at the American Diabetes Association meeting in June 2024, showed the effect in exactly that shape: the proteins Akt acts on picked up more of the chemical tags Akt puts on them, while Akt itself picked up no more of its own. It was not switched on harder. It was working better.
A broad survey of those tags turned up something else. Manganese and insulin have additive effects rather than competing with each other, so raising both does not force a trade-off between them. That makes the metal look less like background nutrition and more like a second lever on the same machinery.
The genuinely new part is not the effect but its schedule. Manganese in the liver is not a fixed background level, the preprint reports: it falls during fasting and rises with feeding. The reason offered is that carbohydrate intake controls how much the cell makes of a particular pump, Slc30a10, whose job is to move manganese out of the liver cell. Make less of the pump and the metal builds up inside; make more and it drains away. The 2024 conference abstract records the pump side of that cycle in wild-type mice, where Slc30a10 sat low during fasting and rose rapidly after refeeding. The preprint goes one step further and says the manganese level itself swings with the cycle. That step is new and, so far, unreviewed.
None of this arrived overnight. The link between manganese and Akt was on the record before this preprint. The group's own 2024 abstract states it plainly, and the work has been presented in public since that June. What the preprint adds is the physiology: not that manganese can reach Akt, but that the liver appears to move the metal on a nutrient schedule, which would make it a control input rather than an accident of diet.
The preprint's closing move deserves a slow read. It proposes that the mechanism suggests a molecular explanation for glucose-lowering effects of manganese observed in humans. The hedge is the authors' own, and it is the right one, because the human record is stranger than that phrase sounds. A case-control study of 3,228 people in China, published in Environmental Health Perspectives in 2016, found the association is U-shaped rather than a straight line. Measured against the middle third of the range, people in the lowest third of plasma manganese carried roughly 1.9 times the odds of type 2 diabetes, and those in the highest third roughly 1.6 times. Both ends were worse than the middle. That paper opens by calling manganese "both an essential element and a known toxicant."
Neither line of work tests taking manganese. The human studies are observational, comparing people who happen to sit at different points on the range, and they cannot say which way the arrow runs. The animal work points the other way entirely. In that earlier abstract, the mouse model in which glucose tolerance improved was a model of manganese excess, built by stripping the Slc30a10 pump out of liver and intestine so the metal could not leave. When that same transporter fails in people through inherited mutation, manganese accumulates in the body and causes a movement disorder. Manganese is sold over the counter as a supplement, and none of this work asks what happens when someone raises their level from outside.
The preprint is a first version posted on bioRxiv, funded by NIH and NSF grants, with no competing interests declared and no journal version yet. It was posted this week, so no one outside the group has yet had time to test it. What is on the table is a proposed mechanism in animals and cells, and the question it leaves open is whether manganese levels in the human liver move with meals as this work reports they do in animals.