# Antigen-processing machinery (APM)

Source: https://onco.cc/terms/antigen-processing-machinery/  
OnCo record `antigen-processing-machinery` (Term). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

Cells cut up proteins, move some fragments to a loading site and place them on display molecules that T cells can inspect. Having a protein inside a cell does not guarantee its fragment is shown.

## Summary

Antigen-processing machinery (APM) is the set of proteins that generate, transport and load peptides onto HLA. A common class I route has five steps. (1) The proteasome cuts intracellular proteins into fragments, which can create a useful peptide or destroy it. (2) The transporter associated with antigen processing, TAP1 with TAP2, moves peptides into the endoplasmic reticulum; Shepherd et al. (1993) showed in a cell-free system that TAP1 is part of an ATP-dependent, sequence-selective peptide translocator. (3) ERAP1 and ERAP2 trim amino acids from the front end. York et al. (2002) found that ERAP1 can enhance or limit presentation by trimming epitopes to 8 to 9 residues, and Saveanu et al. (2005) found complementary human ERAP1 and ERAP2 activity, so more expression does not universally improve display. A class I HLA heavy chain assembles with beta-2 microglobulin (B2M) and a peptide, with helpers such as tapasin. (4) Stable peptide-HLA complexes reach the surface; the peptide stays held by HLA. (5) A compatible T-cell receptor on a conventional CD8 T cell may recognise the pair; recognition is a separate gate and display alone does not show killing.

Class II HLA commonly loads peptides in vesicles that process acquired proteins. An invariant chain protects the new groove and directs it there; after cutting, a remaining fragment is exchanged for other peptides with help from HLA-DM, and conventional CD4 T cells inspect these displays. Cross-presentation lets externally acquired material reach class I. The scheme is a common route, not a rule for every peptide.

A cancer cell may keep an altered protein and still show little of its peptide, and a shared processing defect can affect several targets at once, so adding candidate peptides does not bypass the bottleneck. Low expression, gene loss and a functional defect are different findings: a weak RNA signal does not prove deletion. Interferon gamma can increase presentation machinery when its signalling pathway and the required genes are intact, and cannot replace a deleted B2M or HLA gene. Sucker et al. (2017) described melanoma lesions in which JAK1 or JAK2 defects gave interferon-gamma resistance and evolved into HLA class I negative, T-cell-resistant lesions with antigen-presentation genes silenced and no longer inducible. Zaretsky et al. (2016) found, in biopsies from patients with acquired resistance to PD-1 blockade, JAK1 or JAK2 loss of function in two of four patients and a truncating B2M mutation with lost surface class I in a third. These are melanoma observations, not treatment-benefit evidence.

Evidence layers differ. DNA sequence and copy number show gene retention or alteration but not expression or function; RNA shows sampled gene activity but not protein, assembly or display; protein staining or flow cytometry detects protein but not which peptide is shown; HLA-associated peptide recovery shows a peptide was recovered with HLA in that preparation but not its cell of origin or recognition; controlled T-cell recognition shows function against tested targets. Peptide loading can bypass processing, so rescue by supplied peptide supports a bottleneck under those conditions only. Binding prediction is not presentation, and CD4 help is not proof of CD8 tumour recognition.

## Fields

- Kind: Term
- Last checked: 2026-11-12
- Also known as: APM; antigen processing
- Tags: education-immunology-2026-11

## Notes

- Abstract-depth reads on 9 November 2026. The class II chapter cited by the teaching source (Janeway, NBK26926) refused scripted access, so the invariant chain and HLA-DM description is textbook-level and not checked against the chapter text here.
- The HLA-DM step is independently supported by Vogt and colleagues: in vitro kinetics and binding measurements demonstrated catalytic exchange of invariant-chain-derived CLIP for peptides on HLA-DR. This verifies the exchange mechanism in that system; the inaccessible textbook chapter remains an access failure.
- TAP1, TAP2, ERAP1 and ERAP2 now have separate reviewed-human gene owners. PSMB8, PSMB9 and CIITA retain their existing catalogue owners; this audit does not re-verify every catalogue claim.
- Accessible MHC review: selected invariant-chain, CLIP and HLA-DM sections were read; this is review context in addition to the primary HLA-DM experiment, not an opened Bookshelf chapter.
- In the 14-participant TNBC-MERIT vaccine study, one recurrent tumour had downregulated antigen-presentation transcripts, including B2M, and nearly complete class-I protein loss compared with the earlier specimen. Sahin and colleagues considered B2M downregulation the most probable driver. This was an expression and protein observation in that recurrence, not a demonstrated new B2M DNA mutation, a recurrence probability or proof of a rescue treatment. Allele loss, RNA downregulation and a shared class-I assembly defect require different evidence; adding more source windows does not itself repair the failed step.
- The common proteasome-TAP route is not a rule for every class-I peptide. A TAP or proteasome-component alteration, reduced RNA or one weak stain does not by itself establish complete loss of every class-I display; establish the functional defect and its scope. B2M-dependent assembly, loss of a restricting HLA allele and an inducible signalling defect are different dependencies. The accessible Roche and Cresswell review supports the specified common processing and class-II exchange steps; it does not turn the unread Bookshelf chapters into verified sources.

## Sources

- Shepherd et al., 1993, Cell (PMID 8348620): https://doi.org/10.1016/0092-8674(93)80058-m
- York et al., 2002, Nature Immunology (PMID 12436110): https://doi.org/10.1038/ni860
- Saveanu et al., 2005, Nature Immunology (PMID 15908954): https://doi.org/10.1038/ni1208
- Sucker et al., 2017, Nature Communications (PMID 28561041): https://doi.org/10.1038/ncomms15440
- Zaretsky et al., 2016, New England Journal of Medicine (PMID 27433843): https://doi.org/10.1056/nejmoa1604958
- Vogt et al. 1996: HLA-DM-catalysed peptide exchange on HLA-DR in vitro: https://pubmed.ncbi.nlm.nih.gov/8790398/
- Review: Major Histocompatibility Complex (MHC) Class I and MHC Class II Proteins: Conformational Plasticity in Antigen Presentation.: https://pubmed.ncbi.nlm.nih.gov/28367149/
- Sahin et al.: TNBC-MERIT immune responses and scoped presentation-escape observations (2026): https://pubmed.ncbi.nlm.nih.gov/41708868/
- Roche and Cresswell: antigen processing and presentation mechanisms (2016 review): https://pubmed.ncbi.nlm.nih.gov/27337472/

## Connected records

- terms: [Antigen](https://onco.cc/terms/antigen/), [Dendritic-cell cross-presentation (cDC1)](https://onco.cc/terms/dendritic-cell-cross-presentation/), [Epitope](https://onco.cc/terms/epitope/), [HLA (human leukocyte antigen) and MHC](https://onco.cc/terms/hla/), [HLA allele names and typing reports](https://onco.cc/terms/hla-typing-and-nomenclature/), [Neoantigen candidate selection](https://onco.cc/terms/neoantigen-candidate-selection/)
- targets: [B2M](https://onco.cc/targets/b2m/), [CIITA](https://onco.cc/targets/ciita/), [ERAP1](https://onco.cc/targets/erap1/), [ERAP2](https://onco.cc/targets/erap2/), [JAK1](https://onco.cc/targets/jak1/), [JAK2](https://onco.cc/targets/jak2/), [PSMB8](https://onco.cc/targets/psmb8/), [PSMB9](https://onco.cc/targets/psmb9/), [TAP1](https://onco.cc/targets/tap1/), [TAP2](https://onco.cc/targets/tap2/)

---
JSON: https://onco.cc/api/v1/entities/antigen-processing-machinery.json