Breast Cancer Research (Dec 2019)

Immune gene expression profiling reveals heterogeneity in luminal breast tumors

  • Bin Zhu,
  • Lap Ah Tse,
  • Difei Wang,
  • Hela Koka,
  • Tongwu Zhang,
  • Mustapha Abubakar,
  • Priscilla Lee,
  • Feng Wang,
  • Cherry Wu,
  • Koon Ho Tsang,
  • Wing-cheong Chan,
  • Sze Hong Law,
  • Mengjie Li,
  • Wentao Li,
  • Suyang Wu,
  • Zhiguang Liu,
  • Bixia Huang,
  • Han Zhang,
  • Eric Tang,
  • Zhengyan Kan,
  • Soohyeon Lee,
  • Yeon Hee Park,
  • Seok Jin Nam,
  • Mingyi Wang,
  • Xuezheng Sun,
  • Kristine Jones,
  • Bin Zhu,
  • Amy Hutchinson,
  • Belynda Hicks,
  • Ludmila Prokunina-Olsson,
  • Jianxin Shi,
  • Montserrat Garcia-Closas,
  • Stephen Chanock,
  • Xiaohong R. Yang

DOI
https://doi.org/10.1186/s13058-019-1218-9
Journal volume & issue
Vol. 21, no. 1
pp. 1 – 11

Abstract

Read online

Abstract Background Heterogeneity of immune gene expression patterns of luminal breast cancer (BC), which is clinically heterogeneous and overall considered as low immunogenic, has not been well studied especially in non-European populations. Here, we aimed at characterizing the immune gene expression profile of luminal BC in an Asian population and associating it with patient characteristics and tumor genomic features. Methods We performed immune gene expression profiling of tumor and adjacent normal tissue in 92 luminal BC patients from Hong Kong using RNA-sequencing data and used unsupervised consensus clustering to stratify tumors. We then used luminal patients from The Cancer Genome Atlas (TCGA, N = 564) and a Korean breast cancer study (KBC, N = 112) as replication datasets. Results Based on the expression of 130 immune-related genes, luminal tumors were stratified into three distinct immune subtypes. Tumors in one subtype showed higher level of tumor-infiltrating lymphocytes (TILs), characterized by T cell gene activation, higher expression of immune checkpoint genes, higher nonsynonymous mutation burden, and higher APOBEC-signature mutations, compared with other luminal tumors. The high-TIL subtype was also associated with lower ESR1/ESR2 expression ratio and increasing body mass index. The comparison of the immune profile in tumor and matched normal tissue suggested a tumor-derived activation of specific immune responses, which was only seen in high-TIL patients. Tumors in a second subtype were characterized by increased expression of interferon-stimulated genes and enrichment for TP53 somatic mutations. The presence of three immune subtypes within luminal BC was replicated in TCGA and KBC, although the pattern was more similar in Asian populations. The germline APOBEC3B deletion polymorphism, which is prevalent in East Asian populations and was previously linked to immune activation, was not associated with immune subtypes in our study. This result does not support the hypothesis that the germline APOBEC3B deletion polymorphism is the driving force for immune activation in breast tumors in Asian populations. Conclusion Our findings suggest that immune gene expression and associated genomic features could be useful to further stratify luminal BC beyond the current luminal A/B classification and a subset of luminal BC patients may benefit from checkpoint immunotherapy, at least in Asian populations.

Keywords