How does the immune system avoid potentially damaging responses against the body's own molecules? The answer lies partly in the ability of dendritic cells to sample their surroundings selectively.
Animal tissues are groomed by phagocytes, migratory cells that routinely ingest moribund neighbouring cells, infecting microbes and particulate debris. In the immune system, specialized phagocytes called dendritic cells recycle the molecular fragments from ingested proteins, transporting the scraps to their plasma membranes for display on their external surface as antigens. If the fragments are derived from a microbe, then the dendritic cell adds indicators of the alien status of its meal onto its surface. The displayed antigens are recognized by circulating T lymphocytes, cells that can activate specific immune responses if they later find the same alien molecules presented elsewhere in the body1. The voracious appetites of phagocytes suggest a dilemma concerning mixed meals: phagocytosis of two particles, one that can trigger an immune response and another that cannot, could launch a misguided response against benign molecules.
On page 808 of this issue, Blander and Medzhitov2 show that phagosomes, the intracellular organelles that contain ingested particles and antigenic proteins, can work independently of one another inside the same cell to determine the fate of their enclosed antigens. This subcellular regulation suggests a mechanism for focusing the immune response and provides clues for the rational design of vaccines.
Immature dendritic cells in peripheral tissues ‘sample’ their environment by phagocytosis, initiating specific immune responses when they sense microbes or tissue damage. Ingested material is usually degraded completely as phagosomes mature into phagolysosomes. The maturation, or ‘differentiation’, of dendritic cells can be stimulated in part by signalling through Toll-like receptors (TLRs), a family of cell-surface molecules that recognize molecular patterns common to microorganisms.
TLR4 recognizes lipopolysaccharide, a conserved and abundant surface molecule of Gram-negative bacteria such as Escherichia coli. Through TLR4, the lipopolysaccharide ligand stimulates dendritic-cell maturation and the re-routing of the cell's intracellular traffic, such that fragments of proteins generated in phagolysosomes are bound by ‘MHC class II’ proteins to make antigen-presentation complexes that move to the cell surface3. T lymphocytes capable of recognizing specific peptide antigens in MHC class II complexes continually scan surfaces of mature dendritic cells, and recognition of their cognate ligand typically leads to activation of an immune response specific to that antigen (Fig. 1).
Blander and Medzhitov2 examined whether TLR4-mediated differentiation affects the fate of all phagosomes in the cell or only those containing the TLR4 ligand. They allowed dendritic cells from mice to ingest bacteria or dying (apoptotic) cells containing protein antigens, such that antigen processing could occur in the presence or absence of phagosome-restricted TLR4 signalling. They then compared the antigen-presentation responses when antigen and lipopolysaccharide were eaten together either on the same particle or in distinct particles inside the same cell.
Protein antigens delivered via E. coli, which would be expected to activate TLR4, were fully processed into peptide-loaded MHC class II complexes and were presented to antigen-specific T lymphocytes at the dendritic-cell surfaces. However, proteins delivered by phagocytosis of apoptotic cells, which lack TLR4 ligands, were not presented to T lymphocytes, even when those same dendritic cells contained lipopolysaccharide in other phagosomes. Antigen could be successfully processed from phagosomes containing apoptotic cells if those cells were given lipo-polysaccharide before their ingestion. This indicated that activation of TLR4 in the phagosome resulted in selective loading of antigen from that phagosome.
To address possible confounding effects of comparing different kinds of cells as phagocytosed particles, the authors performed analogous experiments using protein antigens attached to synthetic microspheres, with or without added lipopolysaccharide. Antigens were fully processed and presented when the microspheres they were delivered on also contained lipopolysaccharide. By contrast, antigens associated with lipopolysaccharide-free particles did not reach the cell surface, even in cells that had lipopolysaccharide-labelled particles in other phagosomes. Characterization of isolated microparticle-containing phagosomes showed that all of them, with or without lipopolysaccharide, had matured into phagolysosomes. The essential difference was that the phagosomes with lipopolysaccharide particles contained stable MHC class II molecules and showed indications of the advanced stages of antigen processing (that is, proteolysed invariant chain proteins).
Therefore, the organizational unit for this information processing is not the entire cell but rather the individual organelle. Such phagosome autonomy was indicated by a study of antigen presentation by MHC class I proteins from phagosomes4, although the independence of the phagosomes was not established as thoroughly as in Blander and Medzhitov's work2. The localized selection of antigens for presentation implies that TLR4 signals are generated inside the phagosome, at least with respect to directing the maturation of MHC class II complexes. It also suggests that TLR4 can direct traffic of the membrane organelles that transport protein cargoes from phagosomes to the cell surface. TLR4 signalling was reported to inhibit the trafficking that leads to formation of phagolysosomes5, although this point remains controversial6,7.
It is somewhat surprising that the contents of phagosomes remain isolated from each other inside the cells used by Blander and Medzhitov, as studies of other phagocytes have indicated that phagosomes and endocytic organelles readily merge together or exchange contents8,9. Perhaps TLR4 or other lipopolysaccharide-associated activities isolate lipopolysaccharide-rich phagosomes from other kinds of phagosomes.
Several questions emerge from this work, the answers to which could guide the understanding of immune recognition and the design of vaccines. Is the organizational unit smaller than one phagosome? Perhaps patches of phagosomal membranes are sufficient to organize TLR4-dependent MHC class II maturation and trafficking. How are TLR4-mediated signals that stimulate dendritic-cell maturation integrated with the local signals that drive antigen processing and presentation? Vaccines elicit immune responses using combinations of antigen, particles and TLR ligands, so might vaccine formulations be improved by ensuring the persistent association of these ingredients inside phagosomes? Finally, are there counteracting but as yet undetected signals coming from the apparently inert phagosomes? Maybe every nibbled bit affects the quality of the meal.
Banchereau, J. & Steinman, R. M. Nature 392, 245–252 (1998).
Blander, J. M. & Medzhitov, R. Nature 440, 808–812 (2006).
Mellman, I. & Steinman, R. M. Cell 106, 255–258 (2001).
Gueremonprez, P. et al. Nature 425, 397–402 (2003).
Shiratsuchi, A. et al. J. Immunol. 172, 2039–2047 (2004).
Blander, J. M. & Medzhitov, R. Science 304, 1014–1018 (2004).
Yates, R. M. & Russell, D. G. Immunity 23, 409–417 (2005).
Knapp, P. E. & Swanson, J. A. J. Cell Sci. 95, 433–439 (1990).
Berthiaume, E. P., Medina, C. & Swanson, J. A. J. Cell Biol. 129, 989–998 (1995).