How tumors escape a treatment
Resistance can arise when tumor cells lose a target, bypass a pathway, prevent immune recognition, or tolerate the damage a treatment causes. A mechanism suggests experiments; it does not automatically select the next drug.
Before you start: HLA displays peptides, and CAR-T and TCR-T recognize different kinds of targets.
Selection and new changes both matter
A tumor is a mixed population. Treatment can select cells that already resist it. New changes can also arise during growth. Finding a change in a resistant biopsy does not establish that it caused resistance; comparison with earlier tissue and functional evidence help test that claim.
Recognition has several failure points
Human leukocyte antigen (HLA) class I molecules display peptides to many CD8 T cells. Beta-2 microglobulin (B2M) supports the conventional class I complex. Loss of a presenting HLA allele, B2M disruption or altered antigen processing can weaken recognition.
Allele-specific loss is different from complete loss of presentation. A receptor restricted to the lost allele may fail while receptors using retained alleles remain active. More targets do not overcome a shared failure of their presentation system. Reduced expression can also be reversible; a DNA deletion and an expression change are different findings.
The small, single-arm TNBC-MERIT vaccine study includes an instructive B2M-loss relapse. It does not establish that this is the dominant TNBC resistance mechanism. The earlier “18% before treatment to 50% after six cycles” population claim was not supported by the cited evidence and has been removed. Evidence and assay questions belong in the HLA/B2M assessment.
Match the broken step to the therapy
| Therapy | Required step | Possible resistance mechanism | What would need testing |
|---|---|---|---|
| Vaccine or engineered TCR (T-cell receptor)-T | Intended peptide displayed by the matching HLA | Antigen loss, allele loss, processing failure | Target expression, retained alleles, presentation and recognition |
| Surface-antigen CAR (chimeric antigen receptor)-T or antibody-based engager | Accessible surface target and competent effector cells | Target loss, poor access, suppressive environment | Target distribution, immune context and off-tumor binding |
| TCR-based engager | Its specific peptide–HLA complex | Same display failures as other TCR-based approaches | HLA restriction and peptide presentation |
| Antibody–drug conjugate | Binding, delivery and payload activity | Target changes, trafficking changes or payload resistance | Which step failed; a different payload is not automatically effective |
| Targeted kinase inhibitor | Continued dependence on the inhibited pathway | Target mutation or bypass signaling | Pathway activity and causal dependence |
| Radioligand | Adequate target uptake and absorbed dose | Heterogeneous uptake or repair/tolerance | Dosimetry, distribution and normal-tissue limits |
HLA-independent recognition does not mean resistance-free therapy. Surface targets can disappear, T cells can be excluded, and normal tissue can express the same target. Natural killer cells can respond to missing-self signals, but inhibitory and activating inputs determine their response; HLA loss does not guarantee NK killing.
Why a combination needs its own evidence
Two mechanisms may cover different vulnerabilities. They can also share resistance or add toxicity. A diagram showing complementary biology is a hypothesis, not proof of benefit. No universal rule supports a vendor “stack,” automatic checkpoint switch, or ablation for any blood-positive lesion.
For a confirmed recurrence, clinicians first establish the extent and setting of disease. Biopsy and molecular results can inform an evidence-based treatment or a trial discussion. A ctDNA (circulating tumor DNA) result alone does not show which escape mechanism is present.
Try it
A TCR-T target peptide remains expressed, but the tumor has lost that receptor’s presenting HLA allele. Would adding more copies of the same TCR solve the problem?
Answer: It would not restore the missing display complex. Another retained allele might support another target, but that requires new validation.
Explain it back
“Before calling a finding a rescue strategy, I need to show which step broke, whether the alternative bypasses it, and what clinical evidence supports the alternative.”
Takeaway
Identify and test the failed step; do not turn a plausible bypass into a treatment instruction.
Sources and scope
General mechanisms, source-checked October 8, 2026; expert and learner review pending.