Human Mucosal Associated Invariant T Cells Detect Bacterially Infected Cells

For decades, immunologists divided T cells into two groups — the adaptive cells that learn from experience and the innate cells that respond by instinct. Then a paper by Gold and colleagues drew a line straight through that boundary. They found a population of T cells carrying antigen-recognizing receptors like any adaptive cell, yet responding to bacterial infection with the speed and breadth of innate immunity. That paradox is where this lecture begins. Mycobacterium tuberculosis, the bacterium behind tuberculosis, remains one of the leading causes of infectious disease mortality worldwide. Controlling it requires a coordinated immune response, and CD8-positive T cells — the immune system's cytotoxic specialists — play a central role. These cells normally work by scanning cell surfaces for peptides displayed on classical HLA class I molecules, specifically the HLA-A, HLA-B, and HLA-C proteins. When a cell is infected, fragments of the pathogen get loaded onto these molecules and flagged for destruction. But human immunology studies kept turning up something odd: Mtb-specific CD8-positive T cells were present at high frequency not just in people with active or latent tuberculosis, but also in people with no history of exposure at all. That's strange. Adaptive T cells are supposed to require prior infection to expand. So what were these cells, and why were there so many of them in people who had never encountered the bacteria? Gold and colleagues set out to answer that by performing limiting dilution analysis — a technique that lets you estimate how many T cells in a sample respond to a given target, and then clone them individually to figure out exactly how they work. They screened an average of 128 clones per donor across 16 people, isolating 120 stable Mtb-reactive CD8-positive T-cell clones in total. Then they asked a simple but revealing question: are these cells classically restricted, meaning they depend on HLA-A, HLA-B, or HLA-C, or are they nonclassically restricted, meaning they use something else entirely? The answer split cleanly by infection status. In the five uninfected donors, 85 percent of Mtb-reactive CD8-positive clones were nonclassically restricted, with only 15 percent using the classical pathway. In donors with latent tuberculosis infection, the split was 71 percent nonclassical versus 29 percent classical. And in donors with active tuberculosis, the pattern flipped: 64 percent of clones were classically restricted. The association between classical HLA-Ia restriction and active disease was statistically significant, with a p-value of 0.009. Something nonclassical was dominating the immune response in people without active infection — and the classical response only came to the fore when disease had taken hold. To find out which molecule was presenting antigen to these nonclassical clones, the team ran a series of blocking experiments. They tested antibodies against pan HLA-I, HLA-E, HLA-G, and the CD1 family — and none of them inhibited the nonclassical clones. Then they tried an antibody against MR1, a molecule so conserved that it is essentially identical across all mammals. That one worked. Anti-MR1 blocking abolished interferon gamma production across multiple independent clones, and a second anti-MR1 antibody produced the same result. The restricting element was MR1 — a nonpolymorphic HLA class I-like molecule whose function, at the time of this study, was genuinely unknown. That's where mucosal associated invariant T cells — or MAIT cells — enter the story. MAIT cells had been identified as a population, but no one knew what they did. Gold and colleagues showed that all 14 randomly selected MR1-restricted Mtb-reactive clones expressed the canonical V alpha seven point two T-cell receptor chain. An additional 28 nonclassical clones also expressed V alpha seven point two. Genotypic analysis of six representative clones showed all six used the same germline segment — hAV72 — five of six used the same joining segment, hAJ33, and all six encoded CDR3 alpha loops of the same conserved length. This is the molecular signature of a semi-invariant T-cell receptor, the calling card of MAIT cells. One of the most telling experiments involved TAP — the transporter associated with antigen processing, the machinery that loads peptides onto classical MHC molecules. Gold and colleagues blocked TAP using a viral inhibitor protein called ICP47. In a classical HLA-B08-restricted clone, ICP47 produced over 85 percent inhibition of the response. In the MR1-restricted MAIT clones, ICP47 had no effect at all. This TAP independence tells you something important: whatever antigen MR1 is presenting, it's not a standard intracellular peptide. The team traced the antigen source to the mycobacterial cell wall — specifically the delipidated cell wall fraction, which was strongly antigenic, while culture filtrate proteins were not. Protease digestion of that fraction substantially reduced antigenicity, suggesting a proteinaceous component associated with the cell wall. MAIT cells weren't just Mtb-specific either. Every MR1-restricted clone tested recognized dendritic cells infected with Mycobacterium smegmatis and Escherichia coli. Some clones responded to Salmonella typhimurium and Staphylococcus aureus. That breadth of cross-reactivity across bacterial species — not shared by classically restricted clones — is a hallmark of innate-like sensing. These cells are not tuned to one pathogen; they detect a feature common to many bacteria. Now for the clinical geography. Where do MAIT cells actually live, and what happens to them during disease? The team examined lung tissue from two organ donors and found that 13 percent and 22 percent of lung-resident V alpha seven point two-positive CD8-positive cells produced tumor necrosis factor alpha — a key inflammatory cytokine — in response to Mtb-infected antigen-presenting cells. In peripheral blood across 19 donors, the equivalent figure ranged from 0 to just over 10 percent, with a mean of just over 3 percent. MAIT cells are concentrated in the lung, the very site where Mtb replicates, and they're functionally more active there than in circulation. In peripheral blood, MAIT cells tell a different story depending on disease status. In ex vivo assays using Mtb-infected lung epithelial cells as antigen-presenting cells, the frequency of MR1-dependent V alpha seven point two-positive CD8-positive T cells producing tumor necrosis factor alpha was 0.092 percent in uninfected subjects, 0.185 percent in those with latent infection, and just 0.011 percent in people with active tuberculosis. The difference between active tuberculosis and uninfected donors was statistically significant at a p-value of 0.0025. MAIT cells are depleted from peripheral blood precisely when the disease is worst. The finding that lung epithelial cells can act as MR1-presenting antigen-presenting cells adds another layer. Traditionally, T cell activation has been studied in the context of professional antigen-presenting cells — macrophages and dendritic cells. But Gold and colleagues showed that Mtb infection induces detectable surface expression of MR1 on the A549 human lung epithelial cell line, and that Mtb-infected A549 cells activated MR1-restricted clones in an MR1-dependent manner. Primary human large airway epithelial cells infected with Mtb produced the same result: MR1-dependent interferon gamma production from an MR1-restricted clone, blocked by anti-MR1 antibody. The epithelial lining of the airway — the first surface the bacteria encounter — is itself capable of flagging infection to MAIT cells. Taken together, this is a coherent story with a clear mechanistic thread. MAIT cells sit at the lung mucosa, bearing a semi-invariant receptor that recognizes a conserved bacterial antigen presented by MR1. They're cross-reactive, TAP-independent, and capable of responding to infected epithelial cells without needing a professional antigen-presenting cell to sound the alarm. In uninfected people, they're abundant and readily detectable. In active tuberculosis, they're gone from the blood — either pulled into the lungs, consumed by the response, or destroyed. That last question remains open. Is the peripheral depletion of MAIT cells in active tuberculosis a cause of progression to disease, or a consequence of it? Gold and colleagues raise this directly, and the data don't resolve it. What the data do establish is the function of a previously mysterious immune population: MAIT cells detect bacterially infected cells via MR1, broadly and rapidly, at the mucosal front line. Because MR1 is nonpolymorphic — essentially identical across all humans — a MAIT cell response doesn't depend on which HLA type a person carries. That universality is what makes this finding matter beyond tuberculosis. A sensing system that operates the same way in every human, responding to many bacteria, expressed at mucosal barriers where infections begin — that's not a footnote. That's a mechanism worth building on. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.