Abstract
Relative to donor CD8+ T cells, the recovery of CD4+ T cell populations is severely delayed following allogeneic hematopoietic stem cell transplantation (allo-HCT). We reasoned CD8+ T cell differentiation would therefore take place under ‘helpless’ conditions and impair anti-tumour immunity. To test whether accelerating CD4+ T cell recovery would improve CD8+ T cell fitness and improve anti-tumour immunity, we conducted a randomized phase II, multicentre trial to evaluate the efficacy of prophylactic transfer of CD4+ T lymphocytes after alemtuzumab-based, reduced intensity HLA-matched related donor transplant for haematological cancers (PROT4; NCT01240525).
106 patients were registered for the trial and tapered cyclosporine from day 40 to day 70. 54 patients without GVHD, disease progression and normal graft function were randomised at a 2:1 ratio to receive 1 x 106 CD4+ T cells/kg between day 100-120. Blood samples were taken at day 90 (pre-infusion) and at 12-week intervals over the next 12 months. Immune population architecture and function were analysed using a combination of high-parameter flow cytometry, effector molecule expression, bulk/ single cell (sc) RNAseq and TCRseq. Although the trial did not meet its primary endpoint of increasing 1-year progression-free survival, CD4+ T cell addback induced greater rates of full donor chimerism without a greater risk of severe graft-versus-host disease.
The randomised intervention permitted unbiased measurement of the effects of CD4+ T cell infusion upon T cell population architecture and function. Compared to controls, CD4+ T cell infusion significantly delayed CD8+T cell reconstitution both in terms of frequency and absolute numbers. Although per-cell cytokine generation was similar upon ex vivo stimulation, gene expression profiling showed that CD4+ T cell infusion induced signatures of CD8+ T cellular ‘fitness’ (increased DNA replication, TCR activation and ATP generation). These qualitative changes following CD4+ T cell infusion were matched by a dramatic shift in population architecture towards less differentiated CD8+ T cells with a marked reduction in effector cell accumulation as evaluated by unbiased clustering of flow cytometry and scRNAseq data. By pairing trajectory analysis and TCRseq, we found that CD4+T cell addback restricted CD8+ T cell clonal expansion and redirected differentiation away from terminal effector-like cellular states. In parallel, CD4+ T cell addback remodelled the CD4+ T cell compartment by enriching for circulating PD-1+CXCR5+ T follicular helper (TFH)-like cells with features of central memory cells but antagonized clonal expansion of cytotoxic CD4+ T cells (TCTX).
To test whether TFH cells could provide direct help to CD8+ T cells, we cultured human CD8+ T cells in vitro under conditions of repetitive anti-CD3/CD28 stimulation with or without addition of FACS-isolated CD4+ TFH cells or CD4+ TCTXcells. Whereas helpless or CD4+ TCTX-helped CD8+ T cells underwent terminal differentiation, co-incubation with TFH cells promoted early differentiation-like cellular states and proliferative fitness, mirroring the phenotypic effects identified in trial participants receiving CD4+ T cells. To identify molecules which provide helper signals from CD4+ TFHto CD8+ T cells, we performed scRNAseq and Olink proteomics of CD4+ and CD8+ T cell co-cultures to facilitate interactome analyses searching for differential interactions between CD8+ T cells and CD4+ TFH versus CD4+TCTX. Consistent with pre-clinical models of CD4+ TFH help, IL-21-IL21R signalling was found to be the major candidate axis upon which human TFH help to CD8+ T cells was mediated, with IL-21R neutralisation reducing the frequency and absolute number of CD8+ T cells retaining a less differentiated, memory phenotype.
Thus, CD4+ T cell infusion after allo-SCT in human patients shifts CD8+T cell clonal expansion and differentiation trajectory to less differentiated cellular states. Our finding that human TFH cells use IL-21 to directly promote fitness in human CD8+ T cells provides a potential mechanism involved and is consistent with recent data from pre-clinical models demonstrating TFH-mediated help for anti-tumour and anti-pathogen CD8+ T cells. These data provide further impetus to explore how the properties of CD4+ T cells can be best exploited in immunotherapeutic targeting of cancer.