Vitamin A doesn't reach your immune cells on its own

Vitamin A has sat on the "good for immunity" list so long that most of us stopped asking how it actually reaches an immune cell. A paper published this year answers that, and the answer is stranger than any supplement label suggests: the vitamin doesn't travel there on its own. It gets handed along, cell to cell — and the bacteria in the gut decide whether the handoff happens at all.

The study is "The gut microbiota directs vitamin A flux to regulate intestinal T cell development," by Srinivasan and colleagues at UT Southwestern Medical Center, published in Cell Host & Microbe in 2026 ¹. It was done in mice. Hold on to that — I come back to it.

The relay, one hand at a time

Dietary vitamin A is absorbed by the cells lining the small intestine and converted to retinol. Those same lining cells listen to the gut's bacteria — to the molecular patterns bacteria share across the community. That listening switches on serum amyloid A proteins in the lining. One resident is named: monocolonising germ-free mice with five species one at a time — among them a Bacteroidota gut symbiont and an Escherichia coli lab strain — the authors found that segmented filamentous bacteria, a Bacillota member, robustly induced SAA ¹. That screen scored SAA induction alone, not the flux downstream of it. It is not the whole story, though: mice without SFB still moved more vitamin A along the relay than antibiotic-treated mice did ¹. SAA was shown in 2014 to bind retinol directly ³; in this new work it is the courier that passes retinol from the lining down to the myeloid cells waiting underneath.

A loaded courier then has to move, and it does. Retinoid signalling inside those myeloid cells switches on CCR7, the receptor that reads the chemical signposts pointing toward the mesenteric lymph nodes — the immune waiting rooms that drain the gut. There the myeloid cell meets a naive T cell and makes the final handoff. Retinoid enters the T cell and turns on a genetic programme that tells that cell where to live. The destination is not the new part; retinoic acid was shown to imprint gut-homing on T cells back in 2004 ². The delivery route is.

The whole chain takes about three days ¹. The team followed radiolabelled retinol for ten days ¹ and watched the label appear in lining cells, then in myeloid cells, then in lymph-node T cells, then in T cells back in the gut wall — in that order. When mice drank antibiotics for ten days ¹, absorption into the lining was unchanged, but everything downstream of it stalled. Restore the microbiota, and the relay starts again.

Most of the vitamin never joins the relay — that is the point

One number is worth sitting with. Twenty-four hours ¹ after a dose, about 57 % ¹ of the labelled vitamin had gone to the liver, roughly 3 % ¹ was in the blood, about 22 % ¹ was still sitting in the gut lining, and under 2 % ¹ had reached the myeloid cells that carry it onward. The immune branch is a trickle off the main road. Which is exactly why the courier matters: when the share is that small, whether the handoff happens decides the whole outcome.

The relay has two switches, not one

The most elegant part of the paper is that it pulls the process apart. Bacterial molecular patterns do the first job: they induce SAA, load the myeloid cells, and send them walking. The authors built mice whose gut lining makes SAA regardless of which bacteria are present, and in those animals the loading and the walking still happened after antibiotics had stripped the microbiota out. But the vitamin still did not reach the T cells. That last handoff needed bacterial antigen and cell-to-cell contact. Sensing bacteria and recognising them turn out to be two separate switches on the same line.

The window is early life, around weaning

The relay is not running at full strength from birth. Between 2 and 6 weeks of age ¹ — around weaning, when a mouse's diet and its bacterial community both change at once — SAA in the gut rose, retinoid reached myeloid cells, and gut T cells accumulated. In germ-free animals, and in animals lacking SAA, that rise did not happen. Food and bacteria arrive together, and the immune system gets built on both at the same time.

What the study did not show — no people, no supplement

Here is the caveat, said out loud. This is mouse work — C57BL/6 animals, mostly 6–12 weeks ¹ old, plus germ-free and genetically modified lines. No humans, no clinical outcome, no dose of anything. Nothing here says a vitamin A supplement improves anyone's immunity. The one organism it does single out, SFB, is a mouse gut resident that nobody sells — and even in mice it accounts for only part of the effect.

The authors also name two gaps of their own. They did not directly measure which chemical form of vitamin A crosses into the T cell; retinoic acid degrades in light too quickly to catch in these samples, so its identity is inferred from the signalling rather than seen. And they did not define how the molecule physically moves between the two cells — the contact made during antigen presentation is a plausible site, not a demonstrated one.

What we are left with is a clean, drawable mechanism: food supplies the material, bacteria start the relay, and three days later a T cell knows where it belongs. In mice. That is a good deal more than we knew, and a good deal less than a headline.