The microbiota promotes the maturation and homeostasis of the immune system, in part through the release of microbial products. Early-life microbial colonization has been shown to play a fundamental role in the development of the immune system and imparts long-lasting effects on host fitness. However, despite the importance of this early host-microbiota dialogue, little is known about the microbial-derived signals and antigens involved. This question is of particular importance for mucosal-associated invariant T (MAIT) cells, which are predominantly located in tissues colonized by the microbiota and characterized by their recognition of microbial-derived intermediates of vitamin B2 (riboflavin) synthesis. Because riboflavin synthesis is broadly conserved among bacteria and fungi, MAIT cells are thought to be particularly dependent on the microbiota.

Although MAIT cells are the predominant innate-like lymphocyte subset in humans, there is remarkable variability in their abundance between individuals. How commensal-derived antigens contribute to the variability in MAIT cell abundance and their function has not been established. Furthermore, the extent to which MAIT cells promote tissue physiology remains to be determined.

MAIT cells were highly abundant within human and mouse skin. However, genetically identical mice housed in distinct cages showed striking variability in the proportion of MAIT cells. By contrast, animals in the same cage had similar frequencies of MAIT cells, supporting the hypothesis that these differences were associated with distinct microbiota. MAIT cells accumulated in barrier tissues between 2 and 3 weeks of age, bolstering the idea that MAIT cells develop during a very specific temporal window and in response to defined microbial exposure. The isolation of early-life intestinal commensals and subsequent colonization of neonatal germ-free mice with defined bacteria induced MAIT cell development. Conversely, colonization later in life failed to promote their development within tissues, indicating that microbial exposure must occur during an early-life window, imprinting MAIT cell abundance for life. Commensals that induced MAIT cell development were capable of synthesizing riboflavin, demonstrating antigen necessity, whereas the development of MAIT cells after treatment with a riboflavin derivative indicated sufficiency. After their development in response to early-life commensals, MAIT cells represented a dominant type-17 effector subset in the skin, and cutaneous MAIT cells distinctly expressed a transcriptional program associated with tissue repair. Cutaneous MAIT cells were tissue-resident and required the cytokine interleukin-23 (IL-23) for their homeostasis. These lymphocytes were capable of responding locally to skin commensals in a manner that required IL-1 and IL-18 as well as antigen presentation mediated by the major histocompatibility complex molecule MR1. MR1-mediated presentation of riboflavin metabolites was necessary and sufficient for MAIT cell recognition of skin commensals and further enhanced the tissue-repair program of these lymphocytes. Within the skin, MAIT cells were distinctly localized at the interface of the dermis and epidermis, in close proximity to the basal layer. Topical application of a riboflavin derivative selectively increased MAIT cells in the skin and was sufficient to promote cutaneous wound healing, demonstrating that MAIT cells contribute to skin physiology.

Our work demonstrates how early microbial encounters have long-term effects on the composition of the immune system. We show that MAIT cells are induced …