Background: Retention of a subset of introns in spliced polyadenylated mRNA is emerging as a frequent, unexplained finding from RNA deep sequencing in mammalian cells. Results: Here we analyze intron retention in T lymphocytes by deep sequencing polyadenylated RNA. We show a developmentally regulated RNA-binding protein, hnRNPLL, induces retention of specific introns by sequencing RNA from T cells with an inactivating Hnrpll mutation and from B lymphocytes that physiologically downregulate Hnrpll during their differentiation. In Ptprc mRNA encoding the tyrosine phosphatase CD45, hnRNPLL induces selective retention of introns flanking exons 4 to 6; these correspond to the cassette exons containing hnRNPLL binding sites that are skipped in cells with normal, but not mutant or low, hnRNPLL. We identify similar patterns of hnRNPLL-induced differential intron retention flanking alternative exons in 14 other genes, representing novel elements of the hnRNPLL-induced splicing program in T cells. Retroviral expression of a normally spliced cDNA for one of these targets, Senp2, partially corrects the survival defect of Hnrpll-mutant T cells. We find that integrating a number of computational methods to detect genes with differentially retained introns provides a strategy to enrich for alternatively spliced exons in mammalian RNA-seq data, when complemented by RNA-seq analysis of purified cells with experimentally perturbed RNA-binding proteins. Conclusions: Our findings demonstrate that intron retention in mRNA is induced by specific RNA-binding proteins and suggest a biological significance for this process in marking exons that are poised for alternative splicing.
Immune checkpoint blockade (ICB) therapy involves the inhibition of immune checkpoint regulators which reverses their limitation of T cell anti-tumor responses and results in long-lasting tumor regression. However, poor clinical response or tumor relapse was observed in some patients receiving such therapy administered via antibodies blocking the cytotoxic T lymphocyte-associated protein 4 (CTLA-4) or the programmed cell death 1 (PD-1) pathway alone or in combination, suggesting the involvement of additional immune checkpoints. CD96, a possible immune checkpoint, was previously shown to suppress natural killer (NK) cell anti-tumor activity but its role in human T cells remains controversial. Here, we demonstrate that CRISPR/Cas9-based deletion of CD96 in human T cells enhanced their killing of leukemia cells in vitro. T cells engineered with a chimeric antigen receptor (CAR) comprising human epidermal growth factor receptor 2 (EGFR2/HER2)-binding extracellular region and intracellular regions of CD96 and CD3ζ (4D5-96z CAR-T cells) were less effective in suppressing the growth of HER2-expressing tumor cells in vitro and in vivo compared with counterparts bearing CAR that lacked CD96 endodomain (4D5-z CAR-T cells). Together, our findings implicate a role for CD96 endodomain in attenuating T cell cytotoxicity and support combination tumor immunotherapy targeting multiple rather than single immune checkpoints.