Histone marks identify novel transcription factors that parse CAR-T cell subset-of-origin, clinical potential and expansion [RNA-seq]

Chimeric antigen receptor-modified T cell (CAR-T) immunotherapy has revolutionized the treatment of blood cancers. Parsing the genetic underpinnings of T cell precursor quality and subsequent CAR-T efficacy is challenging. RNA-seq informs infused CAR-T state, but the nature of dynamic transcription during activation hinders identification of transiently or minimally expressed genes, such as transcription factors, and over-emphasizes effector and metabolism genes. We investigated whether analyses of transcriptionally repressive and permissive histone methylation marks reveal associations with CAR-T potential beyond what is seen by transcriptomic analysis. We assessed human CD8+ T cell naïve, central and effector memory subsets that form the substrate of CAR-T cell products, and CAR-T cells derived from these subsets. We extended these observations into the clinic, by examining CAR-T products from a clinical trial of lymphoma patients (NCT01865617). We report that histone marks provide a rich dataset for identification of genes not apparent by conventional transcriptomics. Histone marks improved identification of T cell subsets, CAR-T manufactured from these subsets, and CAR-T manufactured from central memory cells from healthy donors and patients. Using this discriminative approach, we controlled for clinical factors and identified a factor, KLF7, associated with CAR-T cell expansion in patients. Epigenomic methods are an orthogonal, robust and wide-reaching approach for the assessment of T cell immunotherapeutic quality. To investigate the relationship of transcriptomic output and histone marks in CD8+ T cell subsets we performed RNA-seq analyses and CUT&RUN analyses of transcriptionally repressive H3K27me3 and permissive H3K4me2 histone marks on sorted healthy donor (HD) CD8+ naïve, central memory (CM), effector memory (EM) and CM-enriched CD8+ CAR-T cells from a clinical trial (NCT01865617). Quality check (QC) metrics were satisfied for RNA-seq samples and data (6 HD, 3 T cell subsets/HD) and CUT&RUN data (4 HD, 3 T cell subsets/HD) and all 15 patients from NCT01865617. We compare and contrast RNA-seq and CUT&RUN using a custom bioinformatic platform (https://github.com/yezhengSTAT/CAR-T_Histone).