OnCo

Immune checkpoints

The brakes and accelerators on immune cells, drawn where they sit: receptors on the T cell, NK cell and macrophage, their ligands on the tumour cell or antigen-presenting cell, and the enzymes and secreted signals in between. Pink means an approved drug exists against it.

46 members7 classes4 with an approved drug92 productsJSONThe other meaning: cell-cycle checkpoints

The synapse

Hover or focus a node for where it is expressed and its best drug; click for its row. Lines join partners.

T cellor NK cellMacrophageTumour cellor antigen-presenting cellSoluble signalsInside the cellPD-1CTLA-4LAG-3TIM-3TIGITBTLAVISTANKG2AKIR (KIR2DL1/2/3)CD96PVRIG (CD112R)PD-L1PD-L2CD80 (B7-1)CD86 (B7-2)MHC class II (HLA-DR)FGL1Galectin-9CEACAM1CD155 (PVR)CD112 (nectin-2)HVEMHLA-ECD47SIRP-alphaCD24Siglec-10LILRB1 (ILT2)LILRB2 (ILT4)CD28ICOSOX404-1BB (CD137)GITRCD27CD40IDO1TDO2CD73CD39Adenosine A2A receptorB7-H3 (CD276)B7-H4 (VTCN1)HHLA2 (B7-H7)TGF-beta 1IL-10

approved drug against itInhibitory brakeCo-stimulatory acceleratorLigand on the other cellDon't eat me pairMetabolic brakeSoluble signal

Inhibitory brake · 11 members

Inhibitory receptors on T and NK cells

Brakes carried by the immune cell itself. When the receptor meets its ligand on a tumour cell or an antigen-presenting cell, the T cell or NK cell stands down. Antibodies that block the receptor release the brake.

“PD-1 and CTLA-4 antibodies shaped treatment, most patients still show primary or adaptive resistance, and LAG-3, TIM-3, TIGIT and VISTA are the next checkpoints characterised.” Qin et al. 2019, Mol Cancer: novel immune checkpoint targets beyond PD-1 and CTLA-4

Ligand on the other cell · 12 members

Their ligands on tumour and antigen-presenting cells

The other half of each pair. Tumours borrow these molecules from normal tissue so that the brakes on nearby T cells are pressed. Blocking the ligand works as well as blocking the receptor for PD-1.

Don't eat me pair · 6 members

Innate 'don't eat me' checkpoints

Signals a cell shows to macrophages to say it is self and should not be eaten. Tumours over-express them; blocking the pair lets macrophages engulf the cancer cell.

“The CD24-Siglec-10 interaction inhibits macrophage-mediated phagocytosis as well as NK-cell cytotoxicity.” Panagiotou et al. 2022, Cancers: CD24 as a target for cancer immunotherapy

  • Partner
    Expressed on

    Very broadly distributed on normal adult tissues as well as ovarian tumours, especially abundant in some epithelia and the brain. UniProt Q08722

    Best drug status
    Phase 3Evorpacept
    Antibody · Fusion protein · Bispecific
    2 stopped
    Approvals by cancer
    None
  • Partner
    Expressed on

    Myeloid cells but not T cells; ubiquitous at lower levels, highly expressed in brain. UniProt P78324

    Best drug status
    Phase 3Evorpacept
    Fusion protein
    Approvals by cancer
    None
  • Partner
    Expressed on

    B cells and the T-cell surface; erythroleukaemia and small-cell lung carcinoma cell lines. UniProt P25063

    Best drug status
    No product with a status
    Approvals by cancer
    None
    Drugs
    No product in the corpus
  • Partner
    Expressed on

    Peripheral blood leukocytes: eosinophils, monocytes and an NK-cell subpopulation; an inhibitory receptor that recruits phosphatases on ligand binding. UniProt Q96LC7

    Best drug status
    No product with a status
    Approvals by cancer
    None
    Drugs
    No product in the corpus
  • Partner
    Classical and non-classical HLA class I (HLA-A, B, C, G a…
    Expressed on

    B cells, monocytes and myeloid, plasmacytoid and tolerogenic dendritic cells; decidual macrophages and NK cells (protein level). UniProt Q8NHL6

    Best drug status
    No product with a status
    Approvals by cancer
    None
    Drugs
    No product in the corpus
  • Partner
    HLA class I including HLA-G; LILRB receptors negatively r…
    Expressed on

    Monocytes, at lower levels myeloid and plasmacytoid dendritic cells; tolerogenic IL-10-producing dendritic cells, myeloid-derived suppressor cells, B cells and low levels in NK cells. UniProt Q8N423

    Best drug status
    No product with a status
    Approvals by cancer
    None
    Drugs
    No product in the corpus
    • MK-4830 (anti-LILRB2) is in phase 2 with pembrolizumab, including a completed study with chemotherapy in ovarian cancer. ClinicalTrials.gov NCT05446870
Co-stimulatory accelerator · 7 members

Co-stimulatory receptors and agonist targets

Accelerators rather than brakes. These receptors tell a T cell, or the dendritic cell that primes it, to go harder; the drugs are agonists that press them, or bispecifics that press them only where the other arm has found the tumour.

“CD40 activation licenses dendritic cells to promote antitumour T-cell activation and re-educates macrophages to destroy tumour stroma.” Vonderheide 2020, Annu Rev Med: CD40 agonist antibodies in cancer immunotherapy

Metabolic brake · 5 members

Metabolic checkpoints

Enzymes and receptors that change the chemistry around the tumour: they use up tryptophan or turn spilt ATP into adenosine, and either change starves or sedates T cells.

“CD39 and CD73 are cell-surface enzymes that catabolise extracellular ATP into adenosine; ectonucleotidases and adenosine receptors have emerged as therapeutic targets.” Allard et al. 2019, Immunol Rev: targeting the CD73-adenosine axis

Ligand on the other cell · 3 members

Other B7 family members

Relatives of PD-L1 and CD80 whose receptors are only partly known. Because tumours over-express them, most drugs treat them as an address for an antibody-drug conjugate rather than as a brake to release.

“B7-H3, B7x (B7-H4) and HHLA2 form the third phylogenetic group of the B7-CD28 family, with antagonistic antibodies and agonistic fusion proteins emerging for cancer.” Janakiram et al. 2017, Immunol Rev: the third group of the B7-CD28 family

Soluble signal · 2 members

Soluble brakes of the microenvironment

Not receptors but secreted signals that quieten immune cells across a whole tumour. The cited review frames them as immunosuppressive cytokines rather than checkpoints in the strict sense; they are listed here because the drugs against TGF-beta are fused to checkpoint antibodies.

“HPV up-regulates IL-10 and TGF-beta1 to produce a local immunosuppressive environment that inhibits the antitumour immune response.” Torres-Poveda et al. 2014, Infect Agent Cancer: IL-10 and TGF-beta1 in local immunosuppression

Sources. Identifiers from the HGNC REST API; expression from the UniProt tissue-specificity or function comment for the accession named on each row, or from the review cited; drug statuses from the drug records linked (each carries its own approvals and sources); phases for agents without a record here, and every stopped programme, from the ClinicalTrials.gov study cited, quoting the reason the registry gives. Where nothing was found the row says so. Taxonomy in src/data/checkpoint-map.ts.