17P - Continued expansion of long-lived effect CD8+ T cells associates with durable response post-PD-1 blockade

Background

Early CD8⁺ T cell responses to PD-1 blockade are well-characterised. However, an understanding of the long-term clonal and transcriptomic signatures of complete response (CR) to Immune Checkpoint Blockade (ICB) is currently lacking. We aimed to identify and describe CD8⁺ T cell clonal and transcriptomic changes post complete response to PD-1 blockade in metastatic melanoma (MM).

Methods

We performed single-cell RNA sequencing (scRNAseq) of both the T cell receptor (TCR) and transcriptomes of PBMCs and sorted peripheral CD8⁺ T cells in patients with MM with a complete response to PD-1 blockade (n=10), patients with progressive disease (PD, n=6) and healthy donors (HD, n=2). Blood was sampled at multiple time points, including a median of 3.2 years following cessation of PD-1 blockade for CR.

Results

Following a CR and cessation of ICB, there is increased abundance of a subtype of CD8⁺ T cell characterised by high expression of NK-associated genes, termed a killer-like subset. Using the top 25 genes upregulated by this subset to generate a ‘Killer Score’, we show that patients with CR to anti-PD-1 demonstrate a broad increase in killer gene expression over time absent in those with PD and HDs. By tracking key clones through TCR usage, we identify those which are stable/expanding from pre-treatment to post treatment, occupy >0.5% of the repertoire and are persistent post-cessation of ICB, terming these large expanding persistent clones. These clones have a predominantly cytotoxic, effector, gene expression profile and display surface markers consistent with a recently-described ‘Long-lived Effector Cell’ subtype.

Conclusions

CR to PD-1 blockade is associated with persistence of highly-cytotoxic effector CD8⁺ T cells that circulate years after cessation of treatment. Whilst similar subsets have been implicated with chronic viral infection, these observations are novel in the cancer setting and have important implications for mechanisms of durable disease control.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

The Wellcome Trust, Cancer Research UK, National Institute for Health Research.

Disclosure

All authors have declared no conflicts of interest.