checkpoint-map

The A2A receptor is where adenosine lands on a T cell and tells it to rest. Oral antagonists, cousins of caffeine, are in phase 2 combinations with PD-1 blockade and chemotherapy.

BTLA is a brake on T and B cells that is pressed by HVEM, a molecule many tumours carry. An antibody that blocks it is being tested with PD-1 blockade in small-cell lung cancer.

BUB1 is one of the kinases that stops a dividing cell pulling its chromosomes apart before every one is attached. Inhibitors exist in the laboratory; none has reached patients.

CD112 is a widespread cell-adhesion molecule that can either encourage a T cell (through CD226) or hold it back (through PVRIG and TIGIT), depending on which receptor grabs it first.

CD155 is the molecule on tumour cells that TIGIT and CD96 read; normal tissue barely shows it, many tumours show a lot. Every TIGIT antibody works by breaking the CD155 handshake.

CD24 is a second 'don't eat me' badge after CD47: tumours show it to Siglec-10 on macrophages, which then leave them alone. Blocking antibodies are in first-in-human trials.

CD27 is an accelerator on most T cells that is pressed by CD70. Agonist antibodies against it reached phase 2 in lymphoma; its partner CD70 is a separate target for antibodies and CAR-T cells.

CD39 is the first of two enzymes that turn the ATP spilt by dying tumour cells into adenosine, which sedates T cells. Blocking it keeps the ATP alarm ringing; one antibody has completed phase 2 in pancreatic cancer.

CD40 sits on the cells that teach T cells what to attack. Agonist antibodies press it to turn cold tumours hot; the strongest signals so far are in pancreatic cancer with chemotherapy.

CD86 is one of the two B7 molecules on antigen-presenting cells that either wake a T cell up (through CD28) or calm it down (through CTLA-4). Ipilimumab works by stopping CTLA-4 from hogging CD86 and CD80.

CD96 sits on T and NK cells and binds CD155, the same molecule TIGIT reads on tumour cells. It is the least tested of that family; one blocking antibody is in phase 2 platform trials in lung cancer.

FGL1 is a liver protein that some tumours secrete; it binds LAG-3 on T cells and switches them off, separately from the MHC class II route. It explains part of how LAG-3 blockade works.

Galectin-9 is a sugar-binding protein that presses TIM-3 on T cells and can kill the helper T cells that carry it. Tumour blood vessels make more of it than normal ones.

GITR is an accelerator on T cells that agonist antibodies tried to press to boost immunotherapy. Several reached early trials; one programme was dropped for reasons unrelated to safety.

HHLA2 is a B7 relative found on gut, kidney and lung tissue and on many tumours; it can both encourage and restrain T cells depending on the receptor. No drug against it is in the corpus yet.

HLA-E is the self badge that NKG2A on NK cells reads. Tumours raise it to escape NK cells; monalizumab blocks the reader rather than the badge.

IL-10 is a calming signal that many immune cells release; tumours and the viruses behind them raise it to keep the immune system quiet. Oddly, the main cancer drug built on it gave more IL-10 rather than less, and failed.

Inhibitory KIRs are the receptors NK cells use to recognise a person's own HLA-C and stand down. Lirilumab, an antibody that blocks three of them, reached phase 2 but was dropped for blood cancers.

LILRB1 is a brake on macrophages, dendritic cells and some NK and T cells that reads ordinary HLA class I, so any cell showing HLA is protected. Antibodies that block it, with or without LILRB2, are in early trials.

LILRB2 is a brake on monocytes and dendritic cells that reads HLA-G and other HLA class I molecules and keeps them in a tolerant, tumour-friendly state. The antibody MK-4830 blocks it and is in phase 2 with pembrolizumab.

MPS1 arms the checkpoint that holds a dividing cell until its chromosomes are attached. Blocking it lets cancer cells divide carelessly and die; one inhibitor is in phase 1/2 in breast cancer.

NKG2A is the brake on NK cells and some killer T cells that is pressed by HLA-E, a molecule tumours raise to look like self. Monalizumab blocks it and is in phase 3 with durvalumab in lung cancer and with cetuximab in head and neck cancer.

OX40 is an accelerator that appears on T cells once they are switched on; pressing it makes them multiply and survive longer. Agonist antibodies against it are in trials, including a phase 3 in head and neck cancer.

PVRIG is a brake on killer T cells and NK cells that reads CD112 on tumour cells, a cousin of the TIGIT pair. The antibody COM701 blocks it and is being tested in ovarian cancer.

Siglec-10 is the receptor on macrophages and other blood cells that reads the CD24 'don't eat me' badge on tumour cells and calls off the attack.

TDO2 does the same job as IDO1, breaking down tryptophan into kynurenine, which quietens T cells. It is one reason IDO1 inhibitors alone may not have been enough; no TDO2 drug has reached late trials.