T cells are a key component of the adaptive immune system, possessing direct cytolytic ability, the ability to modulate the immune response through the production of cytokines, licensing of B cells and dendritic cells, and establishment of immunological memory1. They play critical roles in immune development, homeostasis and surveillance, protection from pathogens, and prevention and defense from cancer, as well as allergy and autoimmunity1. T cells possess a massive diversity of T cell receptors (TCRs) that are generated through V(D)J recombination, allowing T cells to recognize a vast array of antigens and mount effective immune responses against various pathogens1,2. T cells can be generally categorized into two categories, CD4 T cells, also known as helper T cells, that primarily assist other immune cells, such as B cells, in coordinating the immune response, and CD8 T cells, or cytotoxic T cells, that directly kill infected or cancerous cells by recognizing specific antigens presented on their surfaces1.
The development of chimeric antigen receptors (CAR) has led to a massive increase in interest in the genome engineering of T cells for immunotherapies. CARs are engineered proteins that merge antibody-derived antigen-binding domains with T cell signaling domains, allowing T cells to identify and target cells that express the specific epitope recognized by the antibody portion of the CAR3. These receptors have been used for a variety of immunotherapies, including infectious disease and autoimmunity, but the technology is most advanced for cancer immunotherapies.
CAR-T cells have been extremely successful in the treatment of leukemias and lymphoma, but have shown limited efficacy for the treatment of solid tumors4,5. This has led to a wave of further development seeking to improve the effectiveness of CAR-T cells for solid tumor indications. Multiple approaches have been developed, including cytokine armoring, checkpoint gene knockout, dominant negative receptors, chemokine receptors, expressing multiple CARs in one cell, modification of the CAR to enhance intracellular signaling, and integration into predetermined loci, e.g., the TRAC locus, to exploit host regulatory mechanisms to prevent exhaustion6,7,8. Many of these approaches require either larger genetic cargo and/or site-specific integration. Alternative approaches also include using transgenic TCRs to allow T cells to target intracellular neoantigens9,10. However, this has the significant drawback of requiring the TCR to have specificity to both the neoantigen epitope and the HLA molecule, restricting the use of the eventual therapeutic product to patients expressing the cognate HLA. Furthermore, many tumors alter or reduce HLA expression in response to immunotherapy, greatly reducing the effectiveness of T cells expressing transgenic TCRs11.
Most CAR-T or TCR-T cell therapies in clinical trials are manufactured using retroviral vectors, such as lentivirus or gammaretrovirus, achieving high integration frequency with moderately sized cargo. However, viral vectors suffer from long manufacturing timelines due to current good manufacturing practices (cGMP) requirements and non-specific integration profiles that create a risk of insertional mutagenesis12,13. Furthermore, it can be difficult to produce transgenic retrovirus at high titers if the cargo exceeds 5 kb14. Other vectors, such as those derived from recombinant adeno-associated virus (rAAV), do not integrate naturally but can shuttle DNA donor template to the nucleus and can be used in combination with CRISPR/Cas9 to facilitate traditional homology-directed recombination (HDR) mediated genome engineering. However, these viruses also have long and complicated production workflows and are limited by cargo size (<4.7 kb) and the need to include long homology arms (500-1000 bp)15,16,17,18.
Non-viral genome engineering using transposons or a combination of targeted nucleases and a DNA donor template has been reported in primary human lymphocytes8,19,20. However, these approaches are limited by the toxic response to naked DNA molecules in the cytoplasm following recognition by cytoplasmic DNA sensors expressed in lymphocytes21. Attempts have been made to use small molecule inhibitors of these DNA sensing pathways during transfection, but the redundancy of these pathways may complicate their use in cGMP protocols22. Notably, transposon vectors, such as Sleeping Beauty, PiggyBac, and Tc Buster, allow for the integration of large genetic cargo with high efficiencies, but have a non-specific integration profile23,24. Non-viral, targeted transgene integration using plasmid, linear, or single-stranded DNA templates in combination with a targeted nuclease for HDR is an appealing alternative, but has been limited by poor efficiency, especially with increasingly large genetic cargos, with less than 10% efficiency reported when using cargo over 1.5 kb8,19.
Here, we present the step-by-step protocol for non-viral, homology-mediated end joining (HMEJ) insertion of large DNA payloads in primary human T cells, as described in Webber, Johnson et al.25. HMEJ utilizes short 48 bp homology arms flanked by Cas9 gRNA target sites to allow for high-efficacy targeted integration of large DNA cargo when compared with traditional HDR. One method for reducing the cytotoxicity of plasmid DNA in primary T cells is to employ plasmids with minimized backbones, such as minicircles or nanoplasmids25. Minicircles are miniaturized plasmid vectors produced by excision of the origin of replication and antibiotic resistance gene through recombination after plasmid amplification; they have been shown to improve non-viral engineering of T cells and reduce cell toxicity23,24. Nanoplasmids also have a reduced overall size accomplished through the use of a minimal origin of replication and a non-traditional selection marker26. In our experience, minicircle and nanoplasmid vector platforms offer comparable improvement in efficiency and reduced toxicity over traditional plasmids25.
Here, we present a detailed protocol synergizing temporal optimization of reagent delivery and reagent composition as well as using HMEJ and CRISPR/Cas9 to achieve high efficacy, site-specific genome engineering of primary human T cells with large (>6.3 kb), multicistronic DNA templates for use in immunotherapies and a variety of other applications25. We achieve higher integration with HMEJ and 48 bp homology arms than with traditional HR using 1 kb homology arms, particularly with genetic cargos >1.5 kb25,27. Importantly, T cells engineered through HMEJ repair retain excellent cell expansion, cytotoxicity, and cytokine production, while retaining a non-exhausted phenotype25. This protocol is readily adaptable to cGMP standards and is scalable to clinically relevant cell numbers, enabling a rapid transition to future use in a variety of clinical trials25.