EPAS1 adaptive loss-of-function variants as germline determinants of primary antiangiogenic TKI resistance in high-altitude hepatocellular carcinoma: a translational pharmacogenomic study.

Adaptive loss-of-function variants in EPAS1, which encodes HIF-2, are enriched in high-altitude-adapted populations and may determine response to antiangiogenic tyrosine kinase inhibitors through the HIF-2/STC2 axis.

The authors test the hypothesis that host germline genetic variation, not only acquired tumour alterations, determines primary resistance to antiangiogenic tyrosine kinase inhibitors in hepatocellular carcinoma. Five independent sources were integrated: the retrospective QHRCH-HCC cohort (n = 1,396), a multi-ancestry iPSC-derived endothelial cell transcriptome dataset (GSE160906), TCGA pan-cancer data, GDSC2 pharmacogenomics (951 cell lines, 11 antiangiogenic TKIs) and DepMap dependency data. In the cohort, altitude correlated with PIVKA-II (rho = +0.244; p = 0.0003) and with an altitude-adaptive genetic background score (rho = +0.517; p = 5.59 × 10⁻⁴⁹); under hypoxia, EPAS1 expression in high-altitude-adapted endothelial cells was reduced to 61.4% of controls (p = 0.0006). The EPAS1-STC2 correlation proved weak in TCGA-LIHC (rho = 0.092) compared with HIF1A-STC2 (rho = 0.379; p = 2.21 × 10⁻¹⁴) but strong in clear cell renal cell carcinoma (rho = 0.320; p = 3.47 × 10⁻¹⁴), and in GDSC2 EPAS1 expression correlated positively with the IC50 of all 11 inhibitors tested (sign test p = 0.0005).

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The main weakness is that the link between genotype and treatment response is never established in the same patient: the cohort is tied to the hypothesis through a composite altitude-adaptation score, not through EPAS1 genotyping matched to outcome on treatment, leaving ancestry, altitude and genetic background confounded. The most awkward point comes from the authors themselves, since the EPAS1-STC2 axis they postulate is strong in renal carcinoma but weak in liver, that is, in the target tumour. File it as hypothesis generation, with no consequence for the cancer genetics clinic until a prospective genotype-stratified cohort has been run.

Analysis by Dr Thibaut Benquey