m6A-Modified Nucleotide Bases Improve Translation of In Vitro-Transcribed Chimeric Antigen Receptor (CAR) mRNA in T Cells

Lentiviral transduction remains the gold standard in adoptive modified cellular therapy, such as CAR-T; however, genome integration is not always desirable, such as when treating non-fatal autoimmune disease or for additional editing steps using CRISPR to produce allogeneic CAR-modified cells. Delivering in vitro-transcribed (IVT) mRNA represents an alternative solution but the labile nature of mRNA has led to efforts to improve half-life and translation efficiencies using a range of approaches including chemical and structural modifications. In this study, we explore the role of N⁶-methyladenosine (m6A) in a CD19-CAR sequence when delivered to T cells as an IVT mRNA. In silico analysis predicted the presence of four m6A consensus (DRACH) motifs in the CAR coding sequence and treating T cells with an inhibitor of the m6A methyltransferase (METTL3) resulted in a significant reduction in CAR protein expression. RNA analysis confirmed m6A bases at three of the predicted sites, indicating that the modification occurs independently of nuclear transcription. Synonymous mutation of the DRACH sites reduced the levels of CAR protein from 15 to >50% depending on the T cell donor. We also tested a panel of CAR transcripts with different UTRs, some containing m6A consensus motifs, and identified those which further improved protein expression. Furthermore, we found that the methylation of consensus m6A sites seems to be somewhat sequence-context-dependent. These findings demonstrate the importance of the m6A modification in stabilising and enhancing expression from IVT-derived mRNA and that this occurs within the cell, meaning targeted in vitro chemical modification during mRNA manufacturing may not be necessary.