Back

HLA allele-specific expression loss in tumors can shorten survival and hinder immunotherapy

Filip, I.; Orenbuch, R.; Zhao, J.; Manji, G.; Lopez de Maturana, E.; Malats, N.; Olive, K. P.; Rabadan, R.

2020-10-04 oncology
10.1101/2020.09.30.20204875 medRxiv
Show abstract

Efficient presentation of aberrant peptide fragments by the human leukocyte antigen class I (HLA-I) genes is necessary for immune detection and killing of cancer cells. Patient HLA-I genotypes are known to impact the efficacy of cancer immunotherapy, and the somatic loss of HLA-I heterozygosity has been established as a factor in immune evasion. While global deregulated expression of HLA-I has been reported in different tumor types, the role of HLA-I allele-specific expression loss - that is, the preferential RNA expression loss of specific HLA-I alleles - has not been fully characterized in cancer. In the present study, we quantified HLA-I allele-specific expression (ASE) across eleven TCGA tumor types using a novel method from input RNA and whole-exome sequencing data. Allele-specific loss in at least one of the three HLA-I genes (ASE loss) was pervasive and associated to worse overall survival across tumor types, including pancreatic adenocarcinomas, prostate carcinomas and glioblastomas, among others. In particular, our analysis shows that detection of neoantigens with binding affinity to the specific HLA-I genes subject to ASE loss was a top prognostic indicator of overall survival. Additionally, we found that ASE loss hindered immunotherapy in retrospective analyses. Together, these results highlight the prevalence of HLA-I ASE loss - a previously uncharacterized phenomenon in cancer - and provide initial evidence of its clinical significance in cancer prognosis and immunotherapy treatment.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.