HLA loss of heterozygosity

In one sentence

Loss of heterozygosity at a class-I human leukocyte antigen locus removes one previously present inherited allele from a tumor-cell population, which can narrow the peptides those cells present.

The intuition

Imagine a cell using several different display cases. Losing one kind can remove some displays while leaving others usable. The analogy has a boundary: human leukocyte antigen (HLA) molecules bind particular peptides, short protein fragments. Their function depends on actual sequence, processing and surface expression, rather than a shelf count alone.

How it works

An allele is an inherited sequence version. A person may carry two different versions at a class-I HLA locus, such as HLA-A. Loss of heterozygosity (LOH) means that the tumor cells no longer retain both versions. Total DNA copy count alone cannot establish which version remains.

Allele-aware analysis compares tumor sequence evidence with a suitable matched normal. HLA regions are highly variable, so a generic reference alignment can obscure the distinction. The original LOHHLA method used patient-specific HLA sequences together with estimated tumor purity, the malignant-cell proportion, and ploidy, the chromosome-copy context. Its primary study was in lung cancer, not a prevalence estimate for every breast cancer. McGranahan 2017.

A loss may occur in only one tumor population. Normal cells and other malignant populations can retain the allele and contribute reads. Thus a weak or indeterminate call is not proof of retention. Likewise, a call from one sampled region does not map every lesion.

DNA loss, reduced allele-specific RNA and reduced surface protein are separate findings. MHC (major histocompatibility complex) Hammer analyzes several such mechanisms separately. Beta-2 microglobulin (B2M) supports classical class-I assembly; its functional loss can affect many alleles at once. That differs from losing one HLA allele. Puttick 2024, Sade-Feldman 2017.

Why it matters in cancer

A neoantigen design may depend on a particular presenting allele. A credible loss call therefore raises a specific question: which proposed targets depend on that allele, and what evidence supports presentation through retained alleles? It does not establish that all vaccine targets fail, or automatically identify a successful alternative therapy.

Assay card

Field	What to retain
Measures and method	Allele-specific DNA retention or loss using HLA typing, informative tumor/normal reads and copy-state inference
Input and tissue cost	Tumor and suitable normal DNA; extraction consumes an aliquot. Reanalysis of adequate existing files uses no new tissue
Output and units	Named allele, estimated allelic copy number, loss call and confidence; possible regional or subclonal scope
Thresholds	Caller-specific coverage, purity, copy-state and statistical criteria; no universal percentage establishes vaccine failure
Failure modes	Ambiguous mapping, incorrect typing, low tumor content, inadequate coverage and mixed populations
Validation limits	Retain specimen types and caller/version validation. DNA loss does not directly measure current peptide display or clinical benefit

Common confusions

LOH versus low expression: losing DNA and making less RNA or protein are different mechanisms.
One allele versus the entire system: other alleles may remain; B2M disruption has a different scope.
Negative versus uninformative: insufficient usable tumor evidence cannot exclude loss.

Try it

A fictional design includes peptide P predicted to bind allele A1. Tumor analysis supports A1 loss while A2 is retained. Does this prove that every peptide in the design is unusable?

Answer: No. It challenges the A1-dependent candidate. Examine other candidates' allele dependencies and presentation evidence. Retained A2 does not itself prove that a useful peptide is displayed.

Explain it back

“HLA typing tells me ______; tumor HLA-LOH analysis asks ______.”

One answer: “which inherited versions exist; which of those versions remain in the sampled tumor population.”

Takeaway

Keep inherited type, tumor DNA retention and current presentation as separate questions.

Sources and scope

Source check: October 10, 2026. Mechanism and assay interpretation; the exercise is fictional. Expert and learner review remain pending. No population prevalence or treatment recommendation is proposed.