T cell development and activation

T cells are the coordinators and the assassins of adaptive immunity. Helper T cells decide what kind of response the body makes and help B cells and macrophages do their jobs; cytotoxic T cells find and kill cells that harbour viruses or have become cancerous; regulatory T cells keep the whole system in check. This lesson follows T cells from their origin in the thymus, where they acquire their receptors and are tested against self, through their activation in lymph nodes by dendritic cells, to their differentiation into the subsets that shape each immune response. Lesson 7 then examines what effector T cells do and how memory forms.

The T cell receptor

The T cell receptor (TCR) resembles one arm of an antibody. Most T cells carry an αβ TCR, a heterodimer of an α and a β chain, each with a variable and a constant immunoglobulin domain; a minority carry a γδ TCR. Like antibody genes, TCR genes are assembled by V(D)J recombination using RAG1 and RAG2: the β chain from V, D and J segments, the α chain from V and J segments. Junctional diversity at the joints, again concentrated in CDR3, generates an enormous potential repertoire, estimated at more than 10^15 distinct αβ receptors, although an individual has perhaps 10^8 to 10^10 different TCRs at any time. Unlike B cells, T cells do not undergo somatic hypermutation or class switching, and the TCR is never secreted.

The TCR recognizes a composite surface formed by a peptide and the MHC molecule holding it. Structures solved since 1996 show that TCRs generally dock diagonally across the peptide–MHC surface, with their germline-encoded CDR1 and CDR2 loops contacting mainly the MHC helices and their hypervariable CDR3 loops sitting over the centre of the peptide. The affinity of a TCR for peptide–MHC is low, with dissociation constants typically in the micromolar range, a thousand times weaker than many antibodies. Sensitivity comes from other sources: the CD4 or CD8 coreceptor binds the MHC and brings the kinase Lck to the complex; adhesion molecules hold the cells together; and repeated brief binding by many TCRs over minutes allows a T cell to respond to as few as one to ten peptide–MHC complexes on a target cell.

The TCR itself has short cytoplasmic tails and cannot signal. It is associated with the CD3 complex, made of γ, δ and ε chains, and a ζ chain dimer, whose cytoplasmic tails carry immunoreceptor tyrosine-based activation motifs (ITAMs). When the TCR binds peptide–MHC, Lck phosphorylates the ITAMs, which recruit the kinase ZAP-70, which in turn phosphorylates adaptor proteins and sets off pathways that raise intracellular calcium, activate the transcription factors NFAT, NF-κB and AP-1 and reorganize the cytoskeleton. The immunosuppressant drugs cyclosporine and tacrolimus, essential in organ transplantation (lesson 11), block the phosphatase calcineurin, which is required to activate NFAT.

T cell development in the thymus

T cell progenitors leave the bone marrow and enter the thymus as cells lacking CD4, CD8 and a TCR, called double-negative cells. In the thymic cortex they proliferate and rearrange first the β chain. A functional β chain pairs with a surrogate pre-Tα chain to form a pre-TCR, which signals the cell to stop rearranging β, to proliferate and to express both CD4 and CD8, becoming a double-positive thymocyte. Double-positive cells then rearrange the α chain, and can try several successive α rearrangements on each chromosome, until a complete αβ TCR appears on the surface. Each day an adult mouse thymus produces tens of millions of double-positive cells, yet only about 2 to 5 per cent survive to leave the thymus. The rest die by apoptosis because they fail one of two tests.

The first test is positive selection. Cortical thymic epithelial cells display self peptides on MHC class I and class II. A double-positive thymocyte whose TCR binds weakly to one of these complexes receives a survival signal; one whose TCR cannot bind self MHC at all, as is the case for most random receptors, dies by neglect within a few days. Positive selection ensures that surviving T cells can recognize peptides on the person's own MHC molecules, which explains MHC restriction. It also determines lineage: thymocytes selected on class II become CD4 single-positive cells and those selected on class I become CD8 single-positive cells, a decision governed by the transcription factors ThPOK and Runx3. Cortical epithelial cells use a special thymoproteasome to generate the peptides used for positive selection.

The second test is negative selection. Thymocytes that bind self peptide–MHC strongly are dangerous, because they could attack the body. In the thymic medulla, dendritic cells and medullary thymic epithelial cells present a wide range of self peptides, and thymocytes that bind them with high avidity are deleted by apoptosis or diverted to become regulatory T cells. To expose developing T cells to proteins that are normally found only in particular organs, such as insulin in the pancreas or thyroglobulin in the thyroid, medullary epithelial cells express thousands of tissue-restricted antigens under the control of the transcription factor AIRE (autoimmune regulator). People with mutations in AIRE develop autoimmune polyendocrine syndrome type 1, with autoimmune attack on several endocrine organs and chronic candidiasis, a striking demonstration of the importance of central tolerance (lesson 8).

The selection model implies that the fate of a thymocyte depends on the strength of its TCR signal: too weak and the cell dies; moderate and it survives; too strong and it is deleted. This affinity model raises an apparent paradox, since the same receptor must be selected on self yet be able to respond to foreign peptides. The resolution is that a TCR that weakly binds some self peptide–MHC is likely to bind strongly to some of the vast number of possible foreign peptides on the same MHC, and that thresholds for activation differ between thymocytes and mature T cells. Selection is also leaky: some self-reactive T cells escape, and peripheral mechanisms must control them.

import random
# Toy model of thymic selection: each thymocyte gets a random TCR whose strongest
# binding to self peptide-MHC is drawn from a skewed distribution (arbitrary units).
random.seed(380)
N = 200_000
neglect, positive, negative, treg = 0, 0, 0, 0
for _ in range(N):
    s = random.expovariate(1.0) ** 2        # most receptors bind self MHC very weakly
    if s < 5.3:
        neglect += 1                         # too weak: death by neglect
    elif s < 7.9:
        positive += 1                        # positively selected, survives
    elif s < 9.0:
        treg += 1                            # strong but below deletion: some become Tregs
    else:
        negative += 1                        # deleted by negative selection
for name, k in (("death by neglect", neglect), ("survive as conventional T cells", positive),
                ("diverted to regulatory T cells", treg), ("deleted (negative selection)", negative)):
    print(f"{name:34s} {k / N:6.1%}")

death by neglect                    90.1%
survive as conventional T cells      4.0%
diverted to regulatory T cells       1.0%
deleted (negative selection)         4.9%

The thresholds in the code were chosen only to give proportions of the right order: most thymocytes die by neglect because random receptors rarely fit self MHC well, a few per cent survive as conventional T cells, a smaller fraction with relatively strong self-reactivity become regulatory T cells, and a similar fraction is deleted. Real estimates suggest that in mice roughly 5 per cent of double-positive cells mature, and that negative selection removes a substantial share of those that are positively selected, although the exact proportions remain debated.

The thymus shrinks with age, and by middle age much of it is replaced by fat. The output of new naive T cells falls, and in older adults the T cell pool is maintained largely by division of existing cells, reducing its diversity. This contributes to immunosenescence, the decline of immune function with age that makes older people more vulnerable to infections and less responsive to vaccines. Children born without a thymus, as in complete DiGeorge syndrome, have almost no T cells and can be treated by transplanting thymus tissue.

Activating naive T cells: three signals

Naive T cells circulate between the blood and secondary lymphoid organs, guided by the chemokine receptor CCR7 and the adhesion molecule L-selectin, and survive for months to years with the help of interleukin-7 and contact with self peptide–MHC. To become an effector, a naive T cell must meet a dendritic cell presenting its specific peptide. Activation requires three kinds of signal, a model that explains much about how responses are controlled.

Signal 1 is antigen: the TCR binds peptide–MHC on the dendritic cell, forming a structured contact zone called the immunological synapse, in which TCRs cluster at the centre surrounded by a ring of adhesion molecules. Signal 2 is costimulation: the molecule CD28 on the T cell binds CD80 (B7-1) or CD86 (B7-2) on the dendritic cell, which are increased when pattern recognition receptors detect microbes. Signal 1 without signal 2 tends to make T cells unresponsive, a state called anergy, or to kill them, which is one reason why self antigens presented by resting dendritic cells do not trigger responses. Signal 3 is cytokines, released by the dendritic cell and other cells, which direct the T cell toward a particular type of effector.

Once activated, a T cell produces the cytokine interleukin-2 and expresses the high-affinity IL-2 receptor, including the α chain CD25, which drives its own rapid proliferation and that of its neighbours, the clonal expansion described in lesson 1. Activation also changes the T cell's homing molecules, so that effector cells leave the lymph node and enter inflamed tissues. In parallel, activated T cells begin to express inhibitory receptors such as CTLA-4 and PD-1, which provide negative feedback; CTLA-4 binds CD80 and CD86 more strongly than CD28 does and competes with it. Lesson 8 shows how these checkpoints maintain tolerance, and lesson 12 how blocking them treats cancer.

Helper T cell subsets

Naive CD4 T cells can differentiate into several kinds of helper cell, each specialized for a different type of threat. The concept arose in 1986, when Tim Mosmann and Robert Coffman found that mouse helper T cell clones fell into two groups, Th1 and Th2, based on the cytokines they made. Since then further subsets have been added. Each subset is induced by particular cytokines present during activation, is controlled by a master transcription factor and produces characteristic cytokines that mobilize particular effector cells. The subsets are not rigidly fixed, and some cells change their profile, but the framework remains useful.

Th1 cells are induced when dendritic cells detect intracellular bacteria or viruses and make interleukin-12. Their interferon-γ activates macrophages to kill microbes they have engulfed, as in tuberculosis, and promotes switching to IgG subclasses that opsonize and fix complement. People with inherited defects in the IL-12–interferon-γ axis are highly susceptible to mycobacteria, including the weakened BCG vaccine strain, a condition called Mendelian susceptibility to mycobacterial disease. Th2 cells, induced by interleukin-4 in responses to helminths and allergens, produce interleukin-4, which drives IgE class switching, interleukin-5, which recruits eosinophils, and interleukin-13, which increases mucus production and gut contractility, together helping expel worms. Th17 cells, identified in 2005, recruit neutrophils and maintain barriers against extracellular bacteria and fungi; patients with defects in the IL-17 pathway suffer chronic mucocutaneous candidiasis, and antibodies blocking IL-17 or IL-23 are highly effective treatments for psoriasis.

Tfh cells, which express the chemokine receptor CXCR5 that draws them into B cell follicles, provide the CD40 ligand and interleukin-21 that germinal centre B cells need, as described in lesson 5, and they are the subset induced by most vaccines that work through antibodies. Regulatory T cells, characterized by FOXP3, suppress other immune cells and are the subject of lesson 8. Their discovery earned the 2025 Nobel Prize in Physiology or Medicine.

CD8 T cells

Naive CD8 T cells are activated by dendritic cells presenting peptides on MHC class I, often through cross-presentation. For many responses, CD4 help is required to generate effective and long-lived CD8 memory. Help is delivered indirectly: a CD4 T cell recognizing antigen on the same dendritic cell activates it through CD40 ligand, a process called licensing, so that the dendritic cell expresses more costimulatory molecules and cytokines and can fully activate CD8 T cells. Activated CD8 T cells proliferate and become cytotoxic T lymphocytes carrying granules of perforin and granzymes. Lesson 7 describes how they kill.

Unconventional T cells

Not all T cells recognize peptides on classical MHC molecules. γδ T cells, which make up 1 to 10 per cent of blood T cells but are far more common in the skin and gut epithelium, recognize a variety of antigens, including phosphorylated metabolites called phosphoantigens made by stressed cells and some bacteria, often without MHC. Natural killer T (NKT) cells use a semi-invariant TCR to recognize lipid antigens presented by the MHC-like molecule CD1d. Mucosal-associated invariant T (MAIT) cells, abundant in human blood and liver, recognize vitamin B metabolites made by bacteria and yeasts and presented by the molecule MR1. These cells respond quickly, more like innate cells, and bridge innate and adaptive immunity.