Prediction of non-muscle-invasive bladder cancer recurrence by measurement of checkpoint HLAG's receptor ILT2 on peripheral CD8+ T cells

Oncotarget. 2018 Sep 4;9(69):33160-33169. doi: 10.18632/oncotarget.26036.

Abstract

Background and objective: Recurrence of non-muscle invasive bladder cancer (NMIBC) after initial management occurs in 60-70% of patients. Predictive criteria for recurrence remain only clinical and pathological. The aim of this study was to investigate the prognostic significance of the proportion of checkpoint HLA-G's receptor ILT2-expressing peripheral CD8+ T cells.

Results: The proportion of CD4+ILT2+and CD8+ILT2+ T cells was not increased in NMIBC compared to controls. However, a strong association was found between recurrence and CD8+ILT2+ T cell population levels (p = 0.0006). Two-year recurrence-free survival was 83% in patients with less than 18% CD8+ILT2+ T cells, 39% in the intermediary group, and 12% in patients with more than 46% CD8+ILT2+ T cells. Multivariate analyses demonstrated that the proportion of CD8+ILT2+ T cells was an independent predictive factor for recurrence. Adding CD8+ILT2+ T cells population level to clinical variables increased the predictive accuracy of the model by 4.5%.

Materials and methods: All patients treated for NMIBC between 2012 and 2014 were included prospectively. Blood samples, tumor and clinico-pathological characteristics were collected. HLA-G expression was measured using IHC, and CD8+ILT2+ T cell levels using flow cytometry. Association between HLA-G and CD8+ILT2+ T cell population levels with NMIBC risk of recurrence was investigated using Cox regression analyses. Prediction was measured using the concordance index statistic.

Conclusions: We demonstrated a strong association between the proportion of circulating CD8+ILT2+ T cells and NMIBC risk of recurrence. Gain in prediction was substantial. If externally validated, such immunological marker could be integrated to predict NMIBC recurrence.

Keywords: HLA-G; ILT2; bladder; cancer; immune checkpoint.