# TIL therapy

Tumour-infiltrating lymphocytes: neoantigen recognition without genetic engineering, lymphodepletion as the freeing of a cytokine niche, and the structural conflict between dose and differentiation.

Why the right T cells are already inside the tumour — and why growing enough of them makes them work less well.

Source: https://en.bioecon.ru/docs/health-biomedicine/therapeutics-platforms/til-therapy/
Updated: 2026-09-04



CAR-T therapy begins from the premise that specificity is absent and must be built: a receptor against a pre-chosen surface antigen is engineered into the cell. That gives it reproducibility, and it gives it a weakness — the usable antigens have to be shared across patients, and outside B-cell lineage markers there are very few.

TIL therapy starts from the opposite observation. The tumour is already infiltrated by T lymphocytes, and some of them have been selected in vivo against this particular patient's neoantigens — peptides arising from the somatic mutations of this particular tumour. Specificity is not created; it is recovered.

## What the manufacturing process is actually doing

The resected tumour is cut into fragments and cultured in high-dose interleukin-2. That step is biological rather than technical: inside the tumour, T cells are suppressed — by regulatory cells, by TGF-beta, by cytokine deprivation — and an excess of IL-2 pulls the surviving clones out of that state. A rapid expansion protocol follows: an anti-CD3 antibody as a polyclonal stimulus, IL-2, and irradiated allogeneic mononuclear cells acting as a feeder layer that supplies costimulation. Over two to three weeks a handful of cells becomes tens of billions.

The product is polyclonal by nature. It contains CD8 and CD4 cells recognising many different peptides on the patient's own HLA, so it has no single escape route: loss of one antigen by the tumour does not abolish the response. The same property makes it uncharacterised in the conventional sense — the fraction of genuinely tumour-reactive cells in an unselected preparation is often modest, and is estimated through recent-activation markers such as CD137 or PD-1.

## Why lymphodepletion is not optional

Before infusion the patient receives conditioning with cyclophosphamide and fludarabine. This is not preparation for engrafting foreign tissue — the cells are autologous. The mechanism is different: removing endogenous lymphocytes frees the homeostatic cytokines IL-7 and IL-15, for which all of the body's T cells normally compete. The infused TIL land in an environment saturated with growth signals and undergo homeostatic proliferation. Conditioning also depletes regulatory T cells and myeloid-derived suppressor cells. Without this step the same cells at the same dose perform markedly worse, and the difference is attributed to the freed niche.

## The conflict that does not go away

Persistence of the infused cells correlates not with their number, nor with their cytotoxicity in vitro, but with their differentiation state: less-differentiated memory CD8 cells retaining CD27 and CD28, with longer telomeres, survive longest. But reaching a clinical dose requires many divisions, and every division pushes the cell further down the effector pathway and shortens its telomeres. Dose and quality are hard to obtain together — a structural conflict in the platform, not a flaw in a protocol. Hence "young TIL" approaches with shortened expansion, and attempts to define product composition by selection rather than by bulk growth.

The second constraint is tumour mutational burden. Neoantigen count scales roughly with the number of non-synonymous mutations, so melanoma and tumours driven by ultraviolet or tobacco exposure yield a rich repertoire while low-burden tumours yield almost none. The requirements for a resectable lesion, several weeks of manufacture, and tolerance of high-dose IL-2 toxicity sit on top of that.