Potent and Timed Epigenetic Repression of PD-1 in Human T Cells

T cell exhaustion mediated by checkpoint genes - most prominently PD-1 - is believed to play a major role in limiting the efficacy of chimeric antigen receptor (CAR) T cell therapies of hematologic malignancies. Clinically durable responses to CAR-T therapy of B cell malignancies almost invariably occur in patients who achieve a complete remission by day 28 post infusion. Accumulating data indicate that the impact of T cell exhaustion is most prominent during the early post-infusion interval. PD-1 has recently been found to function as a tumor suppressor in T cells; as such permanent disablement of PD-1 expression via genetic knockout may thus increase long term adverse events from cell therapy.

We sought to develop an approach for increasing the potency and efficacy of CAR-T cell therapy of B cell and other hematologic malignancies by timed epigenetic abrogation of PD-1 expression that was limited to the critical early (21-28d) therapeutic interval. To accomplish this we engineered synthetic transcription factors coupled to a modified KRAB repressive domain that densely tiled the transcriptional regulatory regions of PD-1 at up to single-base resolution.

To identify TF-repressors that were both potent and specific, we transiently transduced each engineered TF-repressors separately into primary total CD3+, CD4+, or CD8+ human T-cells. These experiments revealed striking position dependence of TF-repressor activity, indicating that spatial and rotational positioning of the repressive domain was critical for function. We identified several TF-repressors that exhibited highly potent repression (>90%), and screened these for gene selectivity using transcriptome-wide deep RNA-seq. These studies yielded a TF-repressor with both high potency (>90%) and true single-gene (PD-1) specificity on a genomic scale.

Next we examined the durability of expression. Transduction of TF-repressor mRNA resulted in rapid expression and activity, with PD-1 repression observable beginning at 12 hours and peaking by 48 hours post transduction. Experiments with tagged TF-repressors revealed that substantially all translated protein had dissipated within 72 hours post transduction. Remarkably, however, potent PD-1 repression reliably persisted for >3 weeks post-transduction.

Collectively, our studies show the feasibility of generating highly potent and single-gene-selective epigenome editing reagents capable of reprogramming clinically relevant genes over a defined temporal interval. This approach should be broadly applicable to other key regulators of clinical response to produce genome-targeted medicines without incurring the genotoxic consequences of double-stranded DNA breaks.