Intravascular Immune Surveillance by CXCR6+ NKT Cells Patrolling Liver Sinusoids

The liver receives roughly a quarter of your cardiac output every minute. Blood from the gut pours through the portal vein, carrying nutrients, metabolic byproducts, and anything that slipped past the intestinal barrier. For decades, immunologists treated the liver as a passive filter — a place where things were cleared, not where they were actively hunted. Then Geissmann and colleagues trained a microscope on a living mouse liver and watched immune cells crawling along the vessel walls. Not floating. Not drifting. Crawling. Deliberately. With purpose. That observation is the center of this paper, published in two thousand five. But to understand what they saw, you need to know what kind of cell was moving. Natural killer T cells, or NKT cells, are an unusual hybrid. They carry T cell receptors that recognize lipid antigens, not protein fragments. The antigens are presented by a molecule called CD1d, a non-classical version of the familiar major histocompatibility complex. Most organs have NKT cells in trace amounts, well under one percent of lymphocytes. The liver is different. NKT cells can make up thirty percent of hepatic lymphocytes. That extreme enrichment is what prompted the question: why here, and what are they doing? To answer that, Geissmann and colleagues built a genetic tool. They replaced the coding region of CXCR6, a chemokine receptor, which you can think of as a molecular zip code receptor that tells a cell where to go and stay, with a gene for green fluorescent protein. In these knock-in mice, any cell that normally expresses CXCR6 now glows green. Flow cytometry confirmed that CD1d-reactive NKT cells account for roughly seventy-five to eighty percent of the bright green cells in the liver, and that ninety-nine percent of NKT cells in the liver carry this marker. The liver's glowing cells are, overwhelmingly, NKT cells. Their counterpart molecule, the ligand CXCL16 that binds CXCR6, turned out to be expressed on sinusoidal endothelial cells, the walls of the liver's tiny blood channels. Now the microscope comes in. The team performed intravital fluorescence imaging, using a confocal microscope aimed at the surgically exposed liver of a living mouse, watching green cells move in real time through a coverslip window. What they saw was not cells floating freely through the bloodstream. NKT cells were crawling along the inner surface of the hepatic sinusoids, the capillary-like channels that thread between liver cells, at ten to twenty micrometers per minute. That's roughly one to two cell body lengths per minute. Deliberate, surface-bound movement. Not extravasation into tissue, not recirculation through lymph nodes. The cells stayed inside the vascular space and scanned the walls from within. The functional logic becomes clear when you look at what makes them stop. When the T cell receptor is engaged by antigen, the crawling ceases. Geissmann and colleagues injected alpha-galactosylceramide, a glycolipid that activates NKT cells through CD1d, and forty minutes later assessed the percentage of immobile cells, defined as those moving less than ten micrometers over a six-minute interval. The arrest was dose-dependent: five micrograms worked, but so did fifty nanograms. The cells are scanning sinusoidal surfaces in continuous motion, and they halt the moment they detect something. Crawl, scan, stop. That is the surveillance logic. The team then asked what happens when CXCR6 is gone entirely. In homozygous knockout mice, where no functional CXCR6 is made, the effect on hepatic NKT cells was selective and severe. The liver lost three to five times as many CD1d-reactive NKT cells compared to heterozygous littermates. Other organs were untouched: thymus, spleen, peripheral blood, bone marrow, and lung all showed normal NKT numbers. The defect was liver-specific. Here is the part that reframes how you think about chemokine receptors. When the team looked at how individual NKT cells moved in CXCR6-deficient mice, they found no difference. Mean crawling velocities were sixteen point five micrometers per minute in heterozygotes and eighteen point four in knockouts — statistically indistinguishable. The cells still arrested on T cell receptor stimulation. CXCR6 deficiency didn't change how cells moved. It changed how many survived long enough to be there. The ex vivo survival experiments make this concrete. After eighteen hours in culture, NKT cells from knockout mice died at dramatically higher rates than those from heterozygous mice, confirmed by Annexin V staining, a marker for cells entering apoptosis. No difference in proliferation was detected. When the team added recombinant CXCL16 to cultures of heterozygous cells, survival improved. The signal from sinusoidal endothelium, through CXCL16 to CXCR6, keeps NKT cells alive in the liver. It is not steering them. It is sustaining them. The adoptive transfer experiments sealed the causal argument. When the team co-transferred equal numbers of thymocytes from wild-type and from knockout donors into recipients lacking T cell receptors, after two to three days, the knockout-derived NKT cells were underrepresented specifically in the liver. The requirement for CXCR6 is cell intrinsic. It lives in the NKT cell itself, and without it, those cells cannot accumulate and persist in the liver. The downstream consequence — fewer patrolling cells, less coverage — shows up as a quantitative shift in hepatocyte surveillance. Geissmann and colleagues estimated that in heterozygous mice, each hepatocyte is visited by an NKT cell roughly once every fifteen minutes. In CXCR6-deficient mice, with the lower NKT density, that drops to once every forty-seven minutes. Same crawling speed per cell, but the aggregate coverage collapses because there are fewer cells doing the patrol. That gap translates directly into disease susceptibility. Concanavalin A, or ConA, is a plant lectin that induces T cell-dependent hepatitis when injected at twenty milligrams per kilogram. Serum transaminases, the liver damage markers AST and ALT, peak at twelve hours in control animals. In CXCR6-deficient mice, those markers were approximately eightfold lower at that time point. Histology showed substantially less bridging necrosis. The liver was protected — not because the NKT cells that remained behaved differently, but because their reduced numbers meant less amplified inflammatory damage when the system was provoked. Fewer sentinels, quieter fire. What this paper establishes, taken whole, is that intravascular patrolling inside the sinusoids is a distinct strategy for immune surveillance — one that differs from both tissue residency and lymph-node recirculation. The liver is the right place for it. It receives gut blood continuously, performs extensive lipid metabolism, and is a site of visceral infection by hepatitis viruses, malaria, and leishmania. CD1d is expressed on hepatocytes, Kupffer cells, sinusoidal endothelial cells, and dendritic cells. NKT cells recognize microbial lipid antigens — mycobacterial cell-wall components, leishmania lipophosphoglycan. Scanning the sinusoidal surface from the inside of the bloodstream is a logical way to intercept these antigens before they spread. CXCR6 and CXCL16 are the molecular partnership that makes this possible — not by guiding cells to their patrol route, but by keeping them alive once they’re there. The sinusoidal endothelium provides a continuous survival signal. Without it, the population thins, coverage falls, and the liver's response to immune challenge is blunted. Geissmann and colleagues were careful to note that NKT cells are probably not unique in using this strategy. Natural killer cells and cytotoxic T lymphocytes may also patrol sinusoids. But this paper was among the first to visualize immune patrolling directly, in a living organ, in real time — watching the cells move, watching them stop, and then removing one molecular handshake to show that the entire architecture of surveillance depends on it. 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.