Identification and characterization of neoantigen-reactive CD8+ T cells following checkpoint blockade therapy in a pan-cancer setting

Abstract

Background Immune checkpoint blockade (ICB) has been approved as first-line or second-line therapies for an expanding list of malignancies. T cells recognizing mutation-derived neoantigens are hypothesized to play a major role in tumor elimination. However, the dynamics and characteristics of such neoantigen-reactive T cells (NARTs) in the context of ICB are still limitedly understood. Methods To explore this, tumor biopsies and peripheral blood were obtained pre- and post-treatment from 20 patients with solid metastatic tumors, in a Phase I basket trial. From whole-exome sequencing and RNA-seq data, patient-specific libraries of neopeptides were predicted and screened with DNA barcode-labeled MHC multimers for CD8+ T cell reactivity, in conjunction with the evaluation of T cell phenotype. Results We were able to detect NARTs in the peripheral blood and tumor biopsies for the majority of the patients; however, we did not observe any significant difference between the disease control and progressive disease patient groups, in terms of the breadth and magnitude of the detected NARTs. We also observed that the hydrophobicity of the peptide played a role in defining neopeptides resulting in NARTs response. A trend towards a treatment-induced phenotype signature was observed in the NARTs post-treatment, with the appearance of Ki67+ CD27+ PD-1+ subsets in the PBMCs and CD39+ Ki67+ TCF-1+ subsets in the TILs. Finally, the estimation of T cells from RNAseq was increasing post versus pre-treatment for disease control patients. Conclusion Our data demonstrates the possibility of monitoring the characteristics of NARTs from tumor biopsies and peripheral blood, and that such characteristics could potentially be incorporated with other immune predictors to understand further the complexity governing clinical success for ICB therapy.

IMPROVE: a feature model to predict neoepitope immunogenicity through broad-scale validation of T-cell recognition

“…The second is a cohort of metastatic urothelial carcinoma (mUC) patients who received PD-L1 CPI ( 7 ). The third is a basket trial cohort with different cancer types and different CPI treatments ( 20 ) ( Figure 1A ). Their percentage of immunogenic neoepitopes found in the different cohorts was 3.45%, 2.36%, and 2.16%, respectively ( Supplementary Figure 1A ).…”

“…The cancer patients included in this study were enrolled in three different studies: a melanoma cohort that received TIL-ACT ( 5 , 62 ), a cohort of mUC patients who received PD-L1 checkpoint inhibition ( 6 , 63 ), and a basket trial cohort with different cancer types and CPI treatments ( 20 ). Only patients with assessable patient material were included, which resulted in 70 patients in total.…”

“…In total, 17,520 neoepitope candidates were screened for their potential to activate a T-cell immune response with MHC multimer barcoding as described in References ( 5 , 6 , 13 , 20 ). Predicted neopeptides and viral control peptides were synthesized using Pepscan (Pepscan Presto, Lelystad, The Netherlands).…”