Immune Checkpoints are Predominantly Co-Expressed by Clonally Expanded CD4+FoxP3+ Intratumoral T-cells in Primary Human Cancers

Abstract

Abstract Background. Besides anti-PD(L)1, anti-CTLA-4 and anti-LAG-3, novel immune checkpoint proteins (ICP)-targeted antibodies have recently failed to demonstrate significant efficacy in clinical trials. In those trials, patients were enrolled without screening for drug target expression. Although those novel ICP-targeted antibodies were expected to stimulate anti-tumor CD8 + T-cells, the rationale for their target expression in human tumors relied on pre-clinical IHC stainings and transcriptomic data, which are poorly sensitive/specific techniques to assess membrane protein expression on immune cell subsets. We aimed at describing ICP expression on intratumoral T-cells from primary solid tumors to better design upcoming neoadjuvant cancer immunotherapy trials. Methods. We prospectively performed multiparameter flow cytometry and single-cell RNA (scRNA-Seq) coupled to TCR sequencing on freshly resected human primary tumors of various histological types to determine the actual level of ICP expression at the surface of T-cell subsets. Results. In a given tumor type, we found high inter-individual variability for tumor infiltrating CD45 + cells and for T-cells subsets. The proportions of CD8 + T-cells (~ 40%), CD4 + FoxP3- T-cells (~ 40%) and CD4 + FoxP3 + T-cells (~ 10%) were similar across patients and indications. We found that both stimulatory (CD25, CD28, 4-1BB, ICOS, OX40) and inhibitory (PD-1, CTLA-4, PD-L1, CD39 and TIGIT) checkpoint proteins were predominantly co-expressed by intratumoral CD4+FoxP3+ T-cells. ScRNA-Seq coupled to TCR sequencing revealed that T-cells with high clonality and high ICP expressions were represented by more than 80% of FoxP3+ cells among CD4+ T-cells. Unsupervised clustering of flow cytometry and scRNAseq data identified subsets of CD8 + T-cells and of CD4 + FoxP3- T-cells expressing some checkpoints but those were overall lower then within the CD4 + FoxP3 + T-cells subsets, both in proportions among total T-cells and in terms of ICP expression levels. Conclusions. Tumor histology does not inform on the tumor immune contexture. Assumptions of target expression in clinical trials shall rely on more sensitive and specific techniques than conventional IHC or transcriptomics. Flow cytometry and scRNAseq accurately characterize ICP expression on subsets of immune cells. Like in hematology, flow cytometry could better describe the immune contexture of solid tumors and offer the opportunity to orient and treat patients according to drug target expression rather than tumor histological type.