2O PD-1 defines a distinct, functional, tissue-adapted state in V-delta-1+ T cells with implications for cancer immunotherapy

Checkpoint inhibition (CPI), particularly that targeting the inhibitory co-receptor, programmed cell death protein (PD-1), has transformed cancer care. Although CPI can de-repress cancer antigen-specific αβ T cells that ordinarily show PD-1-dependent exhaustion, it can also be efficacious against cancers evading αβ T cell recognition through loss of MHC or low mutational burdens. In such settings, tissue-associated Vδ1+ γδ T cells have been implicated as these cells are not classically peptide-MHC dependent. Despite some evidence for association of Vδ1+ T cells to CPI response, the immunological significance of PD-1 expression by human Vδ1+ T cells remains uncharacterised. As Vδ1+ T can operate independent of neoantigen load, we looked for a transcriptomic signature of Vδ1+ T cells in a public dataset of patients with melanoma receiving anti-PD-1 therapy and correlated this signature to response in the context of neoantigen load. Alongside this, we developed a protocol which enabled the extraction of substantial numbers of human skin γδ T cells to permit transcriptomic and functional studies of these rare cells. Skin-derived PD-1+ and PD-1- Vδ1+ T cells were sorted and analysed by NanoString using the nCounter Immune Exhaustion Panel. The cells were also activated in vitro both in the presence and absence of recombinant PD-L1 +/- atezolizumab and assayed for cytotoxic degranulation and production of effector cytokines. We found that a transcriptional signature of intratumoral Vδ1+ T cells predicts response to anti-PD-1 in patients with melanoma, particularly in the context of low neoantigen load. Moreover, we found that skin-derived PD-1+ Vδ1+ T cells display a transcriptomic programme of tissue-residence, survival/self-renewal, and functional competence distinct from the canonical exhaustion programme of co-located PD-1+ CD8+ αβ T cells. Indeed, PD-1+ Vδ1+ T cells retained effector responses to T cell receptor signalling that were inhibitable by PD-1 engagement and partially derepressed by CPI. Our formal demonstration that Vδ1+ T cells can be suppressed by PD-1 engagement and de-repressed by anti-PD-(L)1 CPI therapy supports their utility as a predictive biomarker for therapy.