# Self-similarity and immune-recognition specificity

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

## TL;DR

A receptor may recognize more than one related peptide. Target-selection evidence needs to address recognition of healthy tissue as well as the intended target.

## Summary

A primary investigation of an affinity-enhanced MAGE-A3-directed receptor identified cross-recognition of a peptide from titin after unexpected cardiac toxicity. This was a particular engineered-receptor program, not a toxicity rate for all T-cell therapies or neoantigen vaccines. It illustrates why target-gene expression and a sequence match alone are insufficient descriptions of receptor specificity. Proposed-target recognition and off-target recognition need evidence in relevant systems.

## Fields

- Kind: Term
- Last checked: 2026-11-09
- Also known as: peptide self-similarity and immune specificity; TCR cross-recognition

## Notes

- Central and peripheral tolerance include distinct mechanisms. Kisielow and colleagues demonstrated deletion of self-reactive developing cells in an H-Y-receptor transgenic mouse model. Jenkins and Schwartz demonstrated antigen-related unresponsiveness after altered antigen presentation in mouse experiments, and Sakaguchi and colleagues linked a CD4/CD25 regulatory population to prevention of autoimmunity in mice. These are mechanistic examples, rather than universal human response or safety thresholds. A tumour-specific mutation can make a peptide distinguishable from its normal counterpart, but display and a responsive receptor are still needed, and there is no universal percentage of identity that makes a target safe. Higher receptor affinity is not automatically more selective. Receptors can cross-react with peptides from unrelated proteins, as in the titin peptide recognised by the MAGE-A3-directed receptor (Cameron et al., 2013, PMID 23926201; Linette et al., 2013, PMID 23770775). Luksza et al. (Nature 2022, PMID 35589842) modelled neoantigen quality, using non-selfness and selfness, in pancreatic cancer survivors; the model is not a clinical safety rule. In a loaded-peptide concentration series, compare mutant and normal peptides at the same concentration with the same HLA and responder; artificial loading leaves natural processing open and a response threshold is not a patient safety threshold.

## Sources

- Identification of a Titin-derived HLA-A1-presented peptide as a cross-reactive target for engineered MAGE A3-directed T cells.: https://doi.org/10.1126/scitranslmed.3006034
- Primary evidence for the antigen-presentation teaching distinctions: https://doi.org/10.1182/blood-2013-03-490565
- Primary evidence for the antigen-presentation teaching distinctions: https://doi.org/10.1038/s41586-022-04735-9
- Primary tolerance mechanism experiment (PubMed 3260350): https://pubmed.ncbi.nlm.nih.gov/3260350/
- Primary tolerance mechanism experiment (PubMed 3029267): https://pubmed.ncbi.nlm.nih.gov/3029267/
- Primary tolerance mechanism experiment (PubMed 7636184): https://pubmed.ncbi.nlm.nih.gov/7636184/

## Connected records

- terms: [Binder affinity, epitope and developability](https://onco.cc/terms/binder-affinity-epitope-developability/), [Immune measurements and product evidence](https://onco.cc/terms/immune-product-evidence/), [Neoantigen](https://onco.cc/terms/neoantigen/), [Peptide binding, presentation and immune recognition](https://onco.cc/terms/binding-presentation-recognition/)
- technologies: [TCR-T cell therapy](https://onco.cc/technologies/tcr-t/)
- pathways: [Antigen presentation & immune editing](https://onco.cc/pathways/antigen-presentation-immunoediting/)

---
JSON: https://onco.cc/api/v1/entities/self-similarity-immune-recognition.json