Appearance
13.2 — B Cells and T Cells
Your immune system can recognise essentially any molecule — including synthetic chemicals that have never existed in nature, and proteins from organisms you will never meet. The estimated number of different antigen receptors your lymphocytes can produce is somewhere above 10¹¹.
Your genome contains about 20,000 genes.
Resolving that arithmetic is one of the more satisfying pieces of biology there is, and it won a Nobel Prize.
The problem, and the solution
You cannot have a gene for every receptor. There is not enough DNA, and there would be no way to anticipate molecules that do not yet exist.
Susumu Tonegawa showed in 1976 that lymphocytes rearrange their own DNA.
The receptor is not encoded as a single gene. It is assembled during the cell's development, by cutting out and splicing together one segment from each of several sets of alternatives.
A rough arithmetic: with roughly 40 segments in one set, 25 in another and 6 in a third, the combinations alone give 40 × 25 × 6 = 6,000. Do the same for the second chain of the receptor, multiply the two together, and add random insertion and deletion of a few nucleotides at each junction — which is imprecise on purpose — and the total exceeds 10¹¹.
This is somatic recombination, and it is the only place in the body where cells deliberately and permanently rearrange their own genome (Chapter 2.5). Tonegawa received the Nobel Prize in 1987.
Each lymphocyte performs this rearrangement once, early in its development, and then expresses that one receptor for life. So one cell recognises exactly one thing, and the body carries an enormous library of cells each waiting for its particular match.
Which means the response works by selection, not by instruction. The right cell already exists before the pathogen arrives; the pathogen simply finds it and causes it to multiply.
This is clonal selection, proposed by Frank Macfarlane Burnet in 1957, and it explains why the first response is slow — the matching cell must be found and then divide many times — and why the second is fast.
The self-tolerance problem
Generating receptors at random guarantees that some will recognise you.
So both cell types are screened, and the screening is brutal.
T cells are screened in the thymus (Chapter 12.7), in two stages.
Positive selection — a T cell must be able to recognise the body's own MHC molecules, or it is useless and is allowed to die. You have to be able to read the display system to inspect it.
Negative selection — a T cell that binds too strongly to self-antigens is ordered to kill itself (Chapter 1.8).
Around 98 percent of developing T cells die in the thymus. The system generates receptors at random and discards almost all of them, which is expensive and is the price of universal coverage.
And the thymus performs a genuinely remarkable trick to make negative selection work. It expresses genes from all over the body — insulin, thyroid proteins, myelin — in specialised thymic cells, so developing T cells encounter tissue-specific self-antigens they would otherwise never meet until they were loose in the body.
A single transcription factor called AIRE drives this, and people with a faulty AIRE gene develop autoimmune destruction of multiple endocrine glands at once, because the T cells that would have been deleted were never shown what to avoid.
B cells are screened in the bone marrow, by a similar process.
But central screening cannot be complete, because not every self-antigen can be shown to every cell.
So there is peripheral tolerance as a second layer:
The two-signal requirement (Chapter 13.1) — antigen without a danger signal switches a T cell off rather than on.
Regulatory T cells — a subset whose job is to suppress other T cells. People lacking them develop overwhelming multi-organ autoimmunity in infancy, which shows how much active suppression is going on continuously.
Anergy — cells that meet antigen in the wrong context become permanently unresponsive.
Autoimmune disease is the failure of these layers (Chapter 13.6).
T cells
T cells only recognise antigen presented on MHC molecules — they cannot see a free-floating protein at all. This is the fundamental difference from B cells, and it means T cells only respond to things that are inside cells or have been eaten by them.
Two MHC classes, two audiences.
MHC class I is on every nucleated cell, displaying fragments of that cell's own internal proteins. It is a continuous self-report of what is happening inside.
MHC class II is only on professional antigen-presenting cells — dendritic cells, macrophages, B cells — displaying fragments of what they have engulfed from outside.
And the two classes are read by the two T cell types:
Cytotoxic T cells (CD8) read class I. So they inspect every cell in the body and kill the ones displaying something foreign — virus-infected cells and cancer cells.
Helper T cells (CD4) read class II. So they receive reports from cells that have sampled the outside world, and they coordinate the response.
This division makes complete sense once stated as: class I says "look what is inside me", class II says "look what I found outside".
Helper T cells
They kill nothing and they are the most important cells in the adaptive system, because everything else depends on them.
They activate B cells to make antibody and to switch antibody class. They activate cytotoxic T cells.They activate macrophages to kill organisms they have engulfed but cannot destroy alone. They release cytokines coordinating the whole response.
And their central role is demonstrated by what happens when they are lost.
HIV infects CD4 helper T cells (Chapter 17.9). As the count falls, the entire adaptive system degrades — antibody responses, cell-killing, and macrophage activation all fail, even though those cells are themselves uninfected.
Below a CD4 count of about 200 per microlitre, opportunistic infections appear — organisms that healthy people carry harmlessly. The immune deficiency is not from the loss of a killer cell but from the loss of the coordinator.
Helper T cells specialise into subsets, each producing a different set of cytokines and each suited to a different threat.
Th1 — activates macrophages, for intracellular bacteria and viruses. Th2 — drives antibody and eosinophil responses, for parasites — and inappropriately, for allergy (Chapter 13.6). Th17 — recruits neutrophils, for fungi and extracellular bacteria. And heavily implicated in psoriasis and several autoimmune diseases — which is why drugs blocking its cytokines have been so effective in those conditions. T regulatory — suppresses the others.
The subsets cross-inhibit each other, so the response commits to one strategy rather than doing everything at once.
Cytotoxic T cells
They inspect and kill.
On finding a cell displaying foreign peptide on class I, they attach and deliver a lethal package: perforin, which makes pores in the target membrane, and granzymes, which enter through those pores and trigger apoptosis.
Note that they instruct the cell to die rather than bursting it (Chapter 1.8). This matters: apoptosis destroys the cell's contents neatly, including any virus inside, without spilling infectious particles into the tissue. Bursting the cell would release the virus.
One killer T cell can kill several targets in succession, then detach and move on.
B cells
B cells recognise antigen directly, in its native three-dimensional shape, without any presentation.
Which is why antibodies can bind sugars, lipids and complex shapes that T cells cannot see at all.
Their receptor is a membrane-bound antibody. When it binds, the B cell engulfs the antigen, processes it, and displays fragments on MHC class II — and then presents them to a helper T cell.
So a B cell needs T cell help for a full response, and the collaboration is specific: the B cell that captured the antigen presents it to the T cell that recognises a fragment of the same antigen.
With T cell help, the B cell undergoes three changes, all of which happen in specialised structures in lymph nodes called germinal centres.
Proliferation — becoming a large clone.
Class switching — changing which antibody class it makes without changing what it recognises (Chapter 13.3).
Affinity maturation — and this is the remarkable one.
The antibody genes are deliberately mutated at a rate about a million times higher than normal — somatic hypermutation. Most mutations make the antibody worse or useless, and those cells die because they can no longer bind well enough to receive survival signals.
Cells whose mutations improve binding survive and divide more.
So evolution by natural selection runs inside a lymph node, over days. Variation, differential survival, and heredity within a cell lineage (Chapter 3.2). Antibody affinity increases by 10 to 1,000-fold over the course of a response, which is why the antibodies produced in the second week of an infection are far better than those in the first.
Some B cells become plasma cells — antibody factories, packed with endoplasmic reticulum (Chapter 1.5), producing up to 2,000 antibody molecules per second and living days to weeks, though some survive in bone marrow for decades.
Others become memory B cells.
A small number of antigens can activate B cells without T cell help — repeating polysaccharide structures, as found on bacterial capsules. These "T-independent" responses produce antibody quickly and produce no memory and no class switching.
And that has a direct consequence for vaccines. Children under about two respond poorly to plain polysaccharide vaccines, which is exactly why the pneumococcal and Hib vaccines for infants are conjugate vaccines — the sugar is chemically attached to a protein, which recruits T cell help and converts a weak T-independent response into a strong one with memory (Chapter 13.5).
Where it all happens
Lymph nodes are not filters that happen to contain lymphocytes — they are meeting rooms (Chapter 7.8).
The odds of the right T cell meeting the right dendritic cell by chance in the whole body are essentially zero. So the system concentrates both in one place.
Dendritic cells carrying antigen migrate to the draining node. Lymphocytes circulate continuously through nodes, entering from the blood and leaving via the lymph, sampling as they go — each cell passing through a node roughly once or twice a day.
Within a node, T cells and B cells occupy separate zones and meet at the boundary.
And node swelling in infection is the visible result of the clone expanding — which is why a tender enlarged node near an infection is a sign the system is working, not a sign of something sinister.
The timeline
Understanding why the first infection makes you ill and the second usually does not.
Day 0 — infection. Innate defences engage immediately.
Days 0 to 4 — innate response contains it, dendritic cells carry antigen to nodes.
Days 4 to 7 — the matching T and B cells are found and begin dividing. A lymphocyte divides roughly every 6 to 12 hours, so one cell becomes thousands within days. First antibody appears, mostly IgM.
Days 7 to 14 — full response. Class switching to IgG, affinity maturation, killer T cells clearing infected cells. Symptoms peak and then resolve.
Weeks 2 to 4 — the response contracts. Most effector cells die by apoptosis, which is essential — otherwise every infection would leave the body permanently full of cells for a threat that has passed.
Memory cells persist.
On re-exposure: days, not weeks. Higher-affinity antibody from the start, produced in larger quantity, and often the infection is cleared before any symptoms appear at all.
The MHC and individual variation
MHC genes — called HLA in humans — are the most variable genes in the human genome, with thousands of variants at some loci.
And the variation is maintained deliberately. Each MHC variant presents a different set of peptide shapes, so a population with diverse MHC types cannot be wiped out by a single pathogen that happens to evade one type. This is balancing selection (Chapter 3.3), operating on the immune system itself.
Consequences:
Transplant rejection — MHC mismatch is the main target of rejection (Chapter 13.8).
Disease associations — specific HLA types are associated with specific autoimmune diseases, sometimes very strongly. HLA-B27 is present in around 90 percent of people with ankylosing spondylitis.
And drug reactions. HLA-B*57:01 predicts severe hypersensitivity to the HIV drug abacavir, and testing before prescribing has essentially eliminated it. HLA-B*15:02 predicts severe skin reactions to carbamazepine in Han Chinese and Thai populations (Chapter 3.3). These are among the clearest examples of genetic testing changing a prescription.
What the next page fixes
Chapter 13.3 covers the product of all this — the antibody molecule itself: five classes with five different jobs, how one molecule shape does so many things, and why antibodies have become one of the most important classes of drug in modern medicine.