Skip to content

Antigen presentation

Confidently known

  • human JAK1/JAK2 loss-of-function drives both primary and acquired anti-PD-1 resistance. Zaretsky 2016 (NEJM) documented acquired JAK1 or JAK2 LOF (with LOH) plus one B2M truncating mutation in 3 of 4 melanoma patients relapsing on pembrolizumab1. Shin 2017 extended the same mechanism to primary non-responders despite high TMB2. Rare (<5% of all failures) but biochemically definitive; the mutant tumors are non-responsive to IFN-γ.
  • mouse PTPN2 is the tumor-intrinsic inverse of JAK/STAT resistance. An in vivo CRISPR screen (Manguso 2017) identified Ptpn2 loss in tumor cells as an ICI sensitizer by amplifying IFN-γ/JAK-STAT signaling3. Now clinically pursued (e.g., ABBV-514 class).
  • human HLA-I allele-specific LOH is a common, positively selected immune-escape event. McGranahan 2017 (TRACERx 100) used LOHHLA to show HLA LOH in ~40% of NSCLCs, enriched at metastatic sites, with subclonal neoantigen association4. Replicated pan-cancer.
  • human dMMR/MSI-H tumors respond broadly to pembrolizumab (Le 2017)5; basis of the first tissue-agnostic FDA approval and of KEYNOTE-177 for 1L dMMR mCRC6. The presumed mechanism is massive neoantigen load from mismatch-repair failure.

Human-study evidence

Study N Feature Effect 95% CI / p Method
Zaretsky 2016 n=4 (melanoma patients with late relapse on pembrolizumab) acquired JAK1/JAK2 LOF or B2M truncation acquired resistance mechanism 3 of 4 relapse tumors harbored a candidate lesion paired WES + functional IFN-γ / MHC-I assays
Shin 2016 n=39 (23 melanoma + 16 MMR-deficient CRC non-responders) biallelic JAK1/JAK2 LOF primary-resistance prevalence ~4% melanoma, ~6% MMR-d CRC non-responders WES + cell-line IFN-γ / ISG induction assays
McGranahan 2017 n=100 (TRACERx 100 NSCLC) allele-specific HLA-I LOH frequency / selection ~40% HLA-I LOH; subclonal, enriched at metastatic sites LOHHLA (allele-specific WES)
Le 2017 n=86 (phase 2 expansion across 12 tumor types, MMR-deficient tumors) MMR deficiency ORR 53% (CR 21%) MMR IHC / MSI testing + pembro phase 2
André 2020 n=307 (KEYNOTE-177 phase 3, treatment-naive dMMR/MSI-H mCRC) 1L pembrolizumab vs 5-FU-based chemotherapy PFS HR mPFS 16.5 vs 8.2 mo; HR 0.60 95% CI 0.45–0.80, p=0.0002 phase 3 RCT

Contradictions / surprises

  • human HLA-I LOH is not deterministic. Landmark biology established the mechanism; clinical implementation has been more nuanced. Some HLA-LOH patients still respond to anti-PD-1, especially when residual heterozygous HLA alleles present the dominant neoantigen(s). The predictive signal depends on which allele is lost, the tumor type, and the neoantigen landscape. So the correct clinical reading is "HLA LOH is a risk marker, not a disqualifier."
  • human β2M loss sometimes doesn't prevent response. The canonical acquired-resistance mechanism in Zaretsky 2016 documents B2M loss causing PD-1 resistance, but rare B2M-low/null tumors still respond — likely via NK-mediated killing or non-classical HLA presentation. β2M/JAK lesions together explain only a minority of acquired failures in larger cohorts.
  • human TMB does not generalize cleanly across tumor types. Rizvi 2015 established the NSCLC association7; Snyder 2014 the melanoma anti-CTLA-4 association8. FDA's pan-tumor pembrolizumab approval for TMB ≥10 mut/Mb has been controversial — certain high-TMB histologies (some gliomas) show limited benefit. The threshold is assay- and context-dependent.
  • human The "shared neoantigen tetrapeptide signature" from Snyder 2014 did not replicate. Only the coarse TMB association survived.
  • human KEYNOTE-177 showed dMMR is not universal within mCRC. ~29% of patients had primary progression on pembrolizumab (worse than chemo in the first 6 months), and final OS did not reach statistical significance (HR ~0.74, crossover confounded)6. dMMR is necessary-but-not-sufficient even in its strongest indication.

Human-study evidence

Study N Feature Effect 95% CI / p Method
Rizvi 2015 NSCLC discovery + validation cohorts nonsynonymous tumor mutational burden DCB / PFS enrichment cutoff ~178 nonsynonymous mutations separates DCB from no-benefit WES
Snyder 2014 n=64 (discovery n=25 + validation n=39 melanoma on anti-CTLA-4) somatic mutational load (and shared-neoantigen tetrapeptide signature) clinical benefit association mutational load associated with benefit P=0.01 WES
André 2020 n=307 (KEYNOTE-177 phase 3, treatment-naive dMMR/MSI-H mCRC) 1L pembrolizumab vs 5-FU-based chemotherapy PFS HR mPFS 16.5 vs 8.2 mo; HR 0.60 95% CI 0.45–0.80, p=0.0002 phase 3 RCT

Suspected but unconfirmed

  • human deltaHED — a metric combining germline HLA heterozygosity with somatic HLA-I alterations — predicts worse PD-1 outcomes despite higher TMB/neoantigen load across POLARIS-02 (NPC), melanoma, and JUPITER-06 (ESCC) cohorts9. Inverts the naïve "more divergent HLA → more neoantigens → better response" prediction. Retrospective; needs prospective validation in a single-agent PD-1 setting.

Human-study evidence

Study N Feature Effect 95% CI / p Method
Xu 2026 n=729 (3 cohorts: 164 NPC (POLARIS-02), 88 melanoma, 477 ESCC (JUPITER-06)) deltaHED (germline + somatic HLA-I evolutionary divergence) PFS/OS direction worse on PD-1 blockade despite higher TMB/neoantigen load HLA typing + WES (deltaHED metric)

Emerging

  • mouse PAR-2 (F2RL1) activation suppresses DC antigen presentation in lung cancer. A selective negative allosteric modulator (I-117) restores presentation and synergizes with anti-PD-1 in preclinical models — novel GPCR-to-APC axis with a druggable lead (this period's literature).
  • human Tumor B2M expression as a TMB/PD-L1-independent response biomarker in R/M HNSCC progressing on anti-PD-1 (multi-omics 2026). Moves B2M from "LOF = resistance" framing toward a graded expression biomarker.

Human-study evidence

No human-study citations in this section.

Practical takeaways

  • HLA LOH detection (LOHHLA, companion tools) is not yet a clinical-grade decision aid; report it as risk context, not a disqualification.
  • dMMR / MSI-H remains the strongest single biomarker, but KEYNOTE-177's biphasic PFS warns that a subset will fail immediately; early radiologic reassessment at ~2 months is worth considering.
  • TMB thresholds vary by assay and histology; treat the FDA pan-tumor cutoff as a heuristic, not a universal predictor.


  1. Zaretsky 2016 NEJM. Link

  2. Shin 2017 Cancer Discov. Link

  3. Manguso 2017 Nature. Link

  4. McGranahan 2017 Cell. Link

  5. Le 2017 Science. Link

  6. André 2020 NEJM KEYNOTE-177. Link

  7. Rizvi 2015 Science. Link

  8. Snyder 2014 NEJM. Link

  9. deltaHED 2026. Link