T-cell exhaustion

Ask about this lesson

In one sentence

T-cell exhaustion is a distinct differentiation program that T cells enter under persistent antigen stimulation, marked by several inhibitory receptors, graded loss of effector function and a stable TOX-driven gene and chromatin pattern.

The intuition

Think of a firefighter called to the same fire every hour for months. Over time they pace themselves: they still turn up, but they hold back, and their habits harden. Exhaustion is that kind of adaptation, not a cell that is simply tired after one shift. The analogy breaks because the cell's restraint is written into its gene-control machinery, and some subsets keep dividing and supplying new cells.

How it works

Persistent antigen starts it. In chronic viral infection and in cancer, T cells keep meeting their target without the infection or tumor clearing. Exhausted T cells then lose robust effector functions, display several inhibitory receptors at once and switch on an altered gene program (Wherry and Kurachi, 2015).
TOX commits the cell. The transcription factor TOX is induced by repeated signaling. In mice, exhausted cells did not form without TOX, while effector and memory cells formed normally (Khan et al., 2019).
The program has subsets. Progenitor exhausted T cells (Tpex) express T-cell factor 1 (TCF1) and programmed cell death protein 1 (PD-1), keep dividing and give rise to terminally exhausted cells, which are commonly identified by TIM-3 or CD39. In chronic viral infection in mice, the progenitor cells supplied the burst of new cells after PD-1 blockade (Im et al., 2016). In mouse tumors, terminally exhausted cells controlled tumors less well and did not respond to anti-PD-1; in people with melanoma, a higher share of progenitor exhausted cells went with longer responses to checkpoint blockade (Miller et al., 2019).
The chromatin pattern is sticky. In mice, blocking programmed death-ligand 1 (PD-L1) revived exhausted cells and re-engaged effector genes. But their chromatin barely changed, and they became exhausted again when antigen stayed high (Pauken et al., 2016).

Marker or feature	Usually found on	Caution
PD-1	Recently activated and exhausted cells	Alone it does not mean exhaustion
TCF1 (gene TCF7)	Tpex, naive and memory cells	Alone it does not assign Tpex identity
TIM-3, CD39	Terminally exhausted cells	Also seen on some activated tumor-reactive cells
TOX	Exhausted cells	Measured by intracellular staining or RNA, not routine pathology
CD103 with CD39	Tissue-resident memory-like CD8 cells in tumors	A residency program that overlaps with exhaustion markers

Why it matters in cancer

Checkpoint inhibitors act on this program. Immune-checkpoint inhibitors such as pembrolizumab block PD-1. In mouse models, the cells that expand after blockade come mostly from Tpex.
New arrivals may matter as much. In paired human skin-cancer biopsies before and after anti-PD-1, many expanded clones were new to the tumor (Yost et al., 2019). Revival of resident cells and replacement by new clones can both contribute.
Breast cancer has a tissue-resident subset. In single-cell profiling of breast cancers, tumors rich in tumor-infiltrating lymphocytes contained CD8 tissue-resident memory cells with high checkpoint and effector proteins. A gene signature from these cells was associated with better survival in early triple-negative breast cancer (Savas et al., 2018).

Common confusions

Exhaustion vs anergy. Anergy follows priming without proper costimulation and is mainly a block in interleukin-2 production and division; see anergy versus exhaustion.
Exhausted vs useless. In mouse tumors, progenitor exhausted cells kept several functions, persisted and controlled tumor growth better than terminal cells (Miller et al., 2019).
Exhaustion vs "dysfunction". "Dysfunction" is the umbrella word for any T cell failing to control a tumor; exhaustion is one specific program under it.
PD-1-positive vs exhausted. Activation also raises PD-1.

How it is measured

Flow cytometry panels combine PD-1, TIM-3, TCF1 and TOX. Single-cell RNA sequencing with T-cell receptor sequencing links a clone to its program, which is how clonal replacement was found. Functional cytotoxicity assays test what the cells can still do.

Try it

A tumor sample has many PD-1-positive, TCF1-positive CD8 T cells, and few TIM-3-positive ones. Are these cells exhausted, and will PD-1 blockade work?

Answer: The pattern fits a Tpex-rich population, but PD-1 and TCF1 together also fit recently activated or memory cells; TOX, receptor specificity and response history would help. Even a confirmed Tpex pool does not predict a clinical response by itself; the evidence linking Tpex to response comes mainly from mouse models and melanoma cohorts.

Explain it back

"Exhaustion is a ___ program caused by ___; PD-1 blockade can ___ but does not ___." One answer: "differentiation; persistent antigen; revive effector function in some cells; erase the exhausted chromatin pattern."

Takeaway

Exhaustion is a stable, graded program, not tiredness; checkpoint blockade works with it rather than undoing it.

Sources and scope

Source-checked November 7, 2026, against PubMed abstracts. Most mechanism evidence is from mouse chronic-infection and tumor models; human evidence is from biopsies and single-cell cohorts. Expert and learner review pending.