How Your Immune System Recognizes Almost Anything | Generate, Test, Retry, Cache

This is the immune system episode of “A Programmer's View” How can your immune system recognize a molecule that was only made this morning, one found nowhere in nature? Not with a list of every molecule. It builds detectors at random, tests every one, gives some a second try, and keeps the winners as a cache. Treat it as a design problem. You must recognize almost any molecule that could ever get into you, and you have about twenty thousand genes to build with. The obvious answer is a lookup table, one detector worked out in advance for every foreign molecule, and for years some scientists thought the body kept one. It fails twice: twenty thousand lines are nowhere near enough, and a table can only answer questions someone has already asked. So where does the variety come from? Your DNA does not store finished antibody tips. It stores pieces of them in three small boxes. Each new antibody-making cell (a B cell) takes one piece from each box at random, joins them, and cuts out the DNA in between for good, so no two of these cells carry quite the same DNA. Where two pieces meet, a few letters are snipped off and a few new ones added at random. Count those in, and at least a hundred billion different antibodies could be built. Every cell makes its guess before meeting anything at all: generate and test. But what stops a random detector fitting you? At first, nothing: in young human B cells, more than half of the new antibodies stuck to the body's own material. So new cells are tested before they are released. T cells are tested in the thymus, a small organ behind the breastbone, a test environment where thymus cells hold up small pieces of the body. A cell that cannot grip them at all is not kept; one that grips too tightly is not kept either. Nobody runs the test; each cell's fate depends only on its own grip. In young mice about 98 in 100 new cells never leave, most for being useless rather than dangerous. B cells, tested in the bone marrow, get a second try. And the thymus tests on real data: its cells switch on thousands of genes that normally work only in other organs. When one cell finally meets the molecule its guess was made for, it copies itself into a crowd of identical cells, and some copies are kept for years. A lookup table after all, but one that only holds answers already found: a cache. B cells then make small random changes to the gripping tip, and the ones that grip tighter are favoured, round after round, like a game of hot and cold. What comes up Recognizing a synthetic molecule found nowhere in nature What the immune system is Why a lookup table cannot work: about 20,000 genes The old idea of one gene per antibody Antibodies and the gripping tips Gene segments in three boxes, joined at random (V(D)J recombination) DNA cut out for good: no two B cells alike Random letters at the joints, at least 100 billion possible antibodies Generate and test Self-reactive antibodies in young human B cells T cells and the thymus as a test environment Too weak, too tight, just right: positive and negative selection No examiner: each cell's own grip decides About 98 in 100 new T cells never leave the thymus (mice) A second try for B cells in the bone marrow (receptor editing) Real samples of other organs inside the thymus Clonal expansion, memory cells and the cache Affinity maturation: hot and cold, round after round Timestamps 0:00 A molecule made this morning 1:00 The obvious answer: a lookup table 1:43 Building blocks: where the variety comes from 3:06 Generate and test 3:33 What stops it fitting you? 3:54 The thymus: a test environment 5:09 A second try 5:26 Real samples 5:57 A match: copies and a cache 6:43 Hot and cold 7:10 Still running References GENCODE human statistics: https://www.gencodegenes.org/human/st... Tonegawa, Nobel Lecture 1987: https://www.nobelprize.org/uploads/20... IMGT, numbers of human IG and TR genes: https://imgt.org/IMGTrepertoire/Locus... Janeway's Immunobiology, 5th ed. (NCBI Bookshelf): https://www.ncbi.nlm.nih.gov/books/NB... and https://www.ncbi.nlm.nih.gov/books/NB... Alberts, Molecular Biology of the Cell, 4th ed.: https://www.ncbi.nlm.nih.gov/books/NB... Wardemann et al., Science 2003, self-reactive antibodies in early human B cells Sawicka et al., Front. Immunol. 2014: https://www.frontiersin.org/articles/... Derbinski et al., Nat. Immunol. 2001; Anderson et al., Science 2002; Sansom et al., Genome Res. 2014: genes of other organs in the thymus Molecular structures: RCSB PDB entries 1IGT, 1AO7 and 1AO6 (CC0)