The next step in the development of an allogeneic-based CAR T cell therapy is engineering or making changes to the genetic makeup of the isolated T cells. These changes ultimately produce T cells that circumvent the life-threatening issues of rejection (for additional background information, see Overview of Cell Isolation, Engineering & Expansion). The changes also introduce the chimeric antigen receptor (CAR) that targets antigens on the surface of tumor cells. Numerous cell engineering approaches exist and this section will describe three gene editing tools currently used: zinc finger nuclease (ZFN), transcription activator-like effector nuclease (TALEN), and clustered regularly interspaced short palindromic repeat (CRISPR)-associated protein 9 (CRISPR-Cas9). Several methods for delivering these tools to T cells, as well a general overview of the engineering workflow will also be discussed.

Knock-ins versus knockouts

The design of the off-the-shelf or universal CAR T cells from third-party, allogeneic healthy donors uses gene editing tools to mutate specific genes by targeting changes in the DNA of the donor cells. These tools facilitate gene editing in allogeneic T cells in two ways (Figure 1):

  1. Knock-in—adds a gene of interest to achieve the desired function in cells (e.g., adding HLA-E to prevent host NK killing of the allogeneic CAR T cells)
  2. Knockout—disrupts unwanted gene functions of the donor T cells, typically through deletions of genetic sequences (e.g., eliminating the TCR αβ chains of the TCR).

To knock in or knock out genes, scientists usually rely on a DNA-specific endonuclease that is directed to a specific cut site using a “guide” protein or nucleic acid sequence. After the double-stranded DNA is cut, cellular repair mechanisms fix the cut region of the gene. These double-stranded DNA repairs can be completed via two ways: nonhomologous end joining (NHEJ) or homology-directed repair (HDR). The imprecise NHEJ is error prone and can lead to small insertions or deletions (indels) at the target site. If the NHEJ repair is made precisely in the coding region of the targeted gene, an indel or knockout of that gene will be produced (Figure 1). With HDR, a donor template sequence (e.g., HLA-E T cell engineering) flanked by the sequences homologous to those surrounding the cut site is added to the reaction for insertion via recombination. This desired sequence insertion results in a precise gene addition known as knock-in mutation (Figure 1).

Figure 1. Type of mutation is guided by the specific cellular repair mechanisms.

Gene editing technologies used to create allogeneic T cells

Numerous approaches to targeted genome modification can generate these permanent mutations, but three gene editing tools have been well-studied for creating allogeneic CAR T cells with either a knockout TCR complex and knock-in HLA-E gene: zinc finger nuclease (ZFN), transcription activator-like effector nuclease (TALEN), and cluster regularly interspaced short palindromic repeat-associated protein 9 (CRISPR-Cas9) (Figure 2) [1–4].

Three commonly used gene editing tools for CAR T cell therapies include zinc finger nucleases (ZFN), Tal effector nucleases (TALEN), and and cluster regularly interspaced short palindromic repeat-associated protein 9 (CRISPR-Cas9).
Figure 2. Gene-editing tools for allogeneic T cells.


ZFN is a transcription factor that can be specially designed as an artificial endonuclease that cuts a specific sequence of double-stranded DNA. ZFN consists of two components: the first part is a zinc-bearing DNA-binding protein consisting of several zinc finger modules, which acts as a guide to bind to a desired DNA sequence through DNA-protein interactions. The second part is the FokI nuclease, which is connected to the DNA-binding protein via a flexible peptide linker, and cuts the DNA creating a double-stranded break (Figure 2A) [5]. Each zinc finger module interacts with three consecutive nucleotides, so two zinc finger modules will interact with 6 consecutive nucleotides and so on. A fully functional zinc finger DNA-binding domain consists of a chain of 3–6 individual zinc finger modules that hybridize to a highly specific target binding site of 9–18 base pairs, which determines the specificity of the cut region of the DNA sequence. In practice, ZFNs are used as a pair with a right-hand and left-hand ZFN, and each FokI must be dimerized for it to function as a nuclease that can cut the desired double-stranded DNA region. A zinc finger DNA-binding protein can be designed to pair with a desired part of the genome, placing these DNA-binding proteins and the nuclease in the exact region to create the desired mutation or change. Designing ZFN is very time-consuming because the DNA-binding protein (a tertiary structure) needs to “fit” exactly to 3 consecutive nucleotides (Table 1). This process takes twice as long because there is a left and right ZFN requirement.  This strict requirement also makes ZFN a highly specific engineering tool that has very few off-target effects.


TALEN is another specially designed DNA nuclease similar to ZFN. TALEN also has a DNA-binding protein (TALE) that is derived from Xanthamonas and a Fok1 nuclease as the cleavage domain. Also like ZFN, TALEN works as a pair of modules (i.e., a right-hand TALEN and left-hand TALEN) and the two FokI nucleases must be dimerized to form a functional nuclease that can cleave the targeted double-stranded DNA (Figure 2B). The DNA-binding guide component is 12–20 individual TALE repeats that are arranged in a chain, where binding to a single DNA base pair is based on the repeat variable di-residues (RVDs) at position 12 and 13 of each TALE units [6]. TALEN offers high target specificity because two TALEN complexes are required for DNA cleavage, but the final TALENS usually take four weeks to synthesize (Table 1).


More recently, CRISPR-Cas9 gene-editing technology derived from Streptococcus pyogenes has become available and is vastly different from both ZFN and TALEN gene-editing technologies [7] (see also the CRISPR genome editing resource guide, 3rd edition). Similar to ZFN and TALEN, there is a desired number of approximately 20 base pairs for the target sequence. CRISPR-Cas9 technology consists of a nuclease component (Cas9) and RNA component that acts as a guide (gRNA). Unlike ZFN and TALEN, which use protein-DNA interactions and require dimerization of the FokI nucleases for specific cleavage, the CRISPR-Cas9 system relies on RNA-DNA hybridization for target specificity. Another difference centers on acceptable target sites. The CRISPR-Cas9 target site for gene editing must have a protospacer adjacent motif (NGG; also known as PAM site) in order for the gRNA to “locate” the specific DNA sequence, somewhat decreasing the flexibility of this system (Figure 2C). Despite this design drawback, CRISPR-Cas9 is highly efficient and takes less time to develop in comparison to TALEN and ZFN. Figure 3 presents data describing the high efficiency of creating knockouts using CRISPR-Cas9.

Figure 3. High efficiency functional knockout in T cells. T cells were isolated from PBMCs (from a healthy donor) using Invitrogen Dynabeads magnetic beads, and then transfected with Invitrogen TrueCut Cas9 Protein v2 and TrueGuide Modified Synthetic sgRNAs targeting T cell receptor alpha (TRAC) or beta (TRBC) regions using the Neon Transfection System. (A) Analysis by flow cytometry following binding with antibody specific to the T cell receptor (TCR) shows >90% functional knockout of the receptor. For both TRAC and TRBC, gRNAs specific for two different genomic DNA targets (T1 and T2) were tested; results are shown only for the T1 target in each case. (B) Summary of next-generation sequencing (NGS)-based analysis of cleavage efficiency at two different genomic targets (T1 and T2) for both TRAC and TRBC loci. For more details on this experiment, see Achieve functional knockout in up to 90% of human primary T cells.

Summary of gene editing nuclease systems

The two genome editing technologies, ZFN and TALEN, (Figures 2A and 2B) are very specific gene-editing tools because two DNA binding proteins-linked to nucleases (left- and right-handed complexes) must be designed for each editing experiment. The genetic sequences for these tools are transferred to a cell to be expressed into functional nucleases. The CRISPR-Cas9 system (Figure 2C) simply requires the in vitro combination of the designed gRNA and the Cas9 protein to create a single CRISPR-Cas9 ribonucleoprotein (RNP) complex that can induce the double-stranded DNA break once it is delivered into a cell. Table 1 summarizes the three gene-editing technologies and offers examples of their use for off-the-shelf CAR T cell generation.

Table 1. Summary of three commonly used gene-editing tools for allogeneic T cells.

Editing toolsTarget recognitionNucleasesNucleotides target lengthNuclease design timeTarget geneDelivery methodsReferences
ZFNsZinc finger binding to 3 base pairFokI18–36~10 weeksTRAC/TRBC
Torikai H, et al. [8]
Provasi E, et al. [9]
Torikai H, et al. [10]
TALENTALE protein binding to 1 base pairFokI30–35~4 weeksTRAC/TRBC
Poirot L, et al. [11]
Knipping F, et al. [12]
Osborn M, et al. [13]
Qasim W, et al. [14]
Eyquem J, et al. [15]
CRISPR-Cas9Hybridization of gRNA with DNACas920–24~1 weekTCR/ß2M
TRAC and ß2M
TRAC ß2M and TCR
Ren J, et al. [16]
Eyquem J, et al. [15]
Georgiadis C, et al. [17]
Ren J, et al. [18]
Knipping F, et al. [12]
Osborn M, et al. [13]

Delivery systems for gene-editing tools

Delivery of gene-editing tools is challenging because these tools are macromolecules. Unlike small molecule reagents that can easily be dissolved in an aqueous solution and enter the cells by diffusion, macromolecules need an active delivery system to get into the cells. Currently numerous delivery systems are available for in vivo and invitro gene editing applications. These methods include electroporation or nucleofection, lipid or polymer nanoparticles, cell-penetrating peptides, and viral vectors such as lentivirus, adenovirus, or adeno-associated virus. The methods are based on one of three basic principles that can facilitate gene-editing tool delivery into the cells:

  1. Charged-surface interaction—a positively charged particle binds to the negatively charged cell membrane and the cell takes up the particle.
  2. Application of an electrical current to the cells—the electrical current destabilizes the cell membrane, creating pores which allow particles to pass through.
  3. Viral vector systems—modified viruses that are not harmful to humans are used to infect cells and deliver cargo of interest.

Here we will discuss the most used delivery systems of each method, they are lipid-based transfection, electrical-pulse-based transfection, and viral transduction (Table 2).

Cationic lipid-based delivery system

This is the most economical and easiest method of delivery for gene-editing tools because this method does not require the use of special equipment or laboratory vessels. This approach relies on lipid molecules to form liposomes to encapsulate the gene-editing tool cargo, aiding transfection into cells. (e.g., Invitrogen Lipofectamine reagents). In this method, the positively charged lipid nanoparticle complex fuses with the negatively charged cell membranes facilitating the delivery of the constructs into the cells through a process called endocytosis. The gene-editing tools are then released into the cytoplasm and ultimately enter the nucleus for gene editing to begin.

Electrical pulsed-based delivery system

This method can be costly because it requires special equipment and cuvettes. The electrical pulsed-based approach involves placing the cells in a cuvette, suspending the cells in conductive buffer, and briefly applying high voltage electrical pulses. The electricity creates temporary pores in the cell and nuclear membranes, which allow the delivery of macromolecules.

There are two electrical-pulsed-based techniques, electroporation and nucleofection (Table 2). These two techniques differ based on the devices used, parameter controls, types of buffer, and where the macromolecules are delivered into the subcellular space. An electroporation device (e.g., Invitrogen Neon Transfection System) is an “open system”, allowing researchers to manually set the parameters for the optimization of each cell type. The gene-editing nuclease components are delivered into the cytoplasm and ultimately enter the nucleus of the cells. In contrast, nucleofection device (e.g., Nucleofector System) is a “closed system” because the electrical pulse parameters and propriety buffer solutions are preset for each cell type by the manufacturer. In this method, most of the gene-editing macromolecules are delivered directly into the nucleus.

Viral-based delivery system

Lentivirus, a single-stranded RNA virus, is a robust vector and one of the most used viral vectors for gene editing. Lentivirus is an excellent vector for both immunotherapy and gene therapy. Studies showed that lentiviral vectors can deliver various CAR constructs into immune cells, deliver gene-editing tools such as CRISPR-Cas 9 into target cells, or deliver a heathy gene to correct diseases such as sickle cell disease, severe combined immunodeficiency, and 18 β-thalassemia [19, 20]. Most noticeable, lentiviral vectors have been approved for the clinical application of delivering CAR constructs into T cells for CAR T therapies. The biggest advantage of lentiviral viral vectors is their high infection (or transduction) efficiency in both dividing and non-dividing cells. The vectors have good safety profiles, have a large carrying capacity, and can maintain long-term transgene expression [21]. The large carrying capacity of single design lentiviral vectors can deliver Cas9, sgRNA, and a puromycin selective marker into a target cell [22].

Table 2. Delivery methods used for gene editing.

Lentivirus vectorDNA/RNA is packaged into the infectious viral particles and introduced into cellsHigh transduction efficiency
Suitable for primary immune cells
CAR T clinical usage precedent
  • Special lab environment needed
  • Requires safety measures
  • Labor intensive
ElectroporationElectrical pulse creates pores in the cell membrane, allowing the entry of DNA/RNA/RNP into the cell cytoplasm the nucleusFast and easy
Large number of cells can be transfected in minutes
  • Requires special equipment
  • Cytotoxic anions can form during the procedure, leading to cell death
Neon Transfection System
MaxCyte GT (Flow electroporation)
NucleofectionSimilar to electroporation except the tools are delivered directly into the cell nucleus using very specific conditions provided by manufacturersEffective for non-dividing cells
High throughput potential–multiple samples can be transfected simultaneously
  • Requires special equipment
  • Less flexible–special protocols and reagents can't be controlled by the user
Lonza Nucleofector System
Cationic lipidsPositively charged liposomes encapsulate protein/RNP and interact with negatively charged cell membrane facilitating entry into the cell cytoplasm via the endocytosis pathwayEasy and versatile (no special equipment), not very toxic and can be a high throughput system.
  • Lower transfection efficiency
  • Not direct delivery into the nucleus
  • Not applicable for all cell types
Lipofectamine CRISPRMAX Cas9 Transfection Reagent

Overview of the gene-editing workflow

The complexity of generating engineered allogeneic T cells for CAR T therapy is quite formidable. With the help of gene editing, numerous immunological hurdles can be overcome to create an allogeneic T cell that will be accepted by the recipient patient without causing life threatening consequences. The workflow can be summarized as follows:

  • Design and creation of chosen gene-editing nuclease
  • Transfection of gene-editing tools into T cells
  • Monitoring cleavage efficiency
  • Analysis and validation of knock-ins and knockouts

A number of preparatory steps need to be performed before the actual engineering of the T cells, beginning with the choice and design of the nuclease components. Each of these gene-editing tools has its own rules and criteria for the design of DNA binding domain; however, the final design for each relies heavily on the use of sequence analysis software tools to determine best sequences surrounding the targeted cut sites that limit undesirable or “off-target” binding sites. Free online software tools specifically developed for use in gene editing are available (e.g., Invitrogen TruDesign Genome Editor), and many gene-editing nuclease vendors offer design help. Once the design is finalized, the gene-editing nuclease components can take from 1–10 weeks to produce (Table 1).

Once T cells are isolated and activated from donors, the cells can be transfected with the chosen gene-editing nuclease and engineered immediately, or if cell numbers are low, they can be expanded and engineered when sufficient cell numbers are reached. The cells can also be frozen and engineered at a later time.

Every step in the gene-editing process needs to be confirmed before moving onto the next step. Typically, these steps can include: determining the efficiency of cleavage at the desired site, checking the sequence of PCR products, monitoring gene and protein expression, and checking the impacts (e.g., toxicity and viability) on the model cell system. Numerous reagents and flow cytometry systems are available to quantify TCR knockout efficiency.

Other cell sources for allogeneic CAR cell therapies

Several other cell sources exist that are compatible with generation of off-the-shelf CAR T cells such as embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs). Both ESCs and iPSCs have the ability to self-renew, allowing potentially limitless in vitro expansion. Human ESCs are derived from embryos and therefore can be problematic for clinical development due to regulatory issues and limited sources [23]. Human iPSCs, on the other hand, are derived from adult somatic cells [24]. For example, fibroblasts can be reprogrammed in vitro, can be transformed into iPSC cells, and behave like embryonic pluripotent stem cells that can be differentiated into red blood cells, T cells, B cells, and NK cells with unlimited proliferation capacity [25] (For additional information, see the Pluripotent Stem Cell Resource Handbook and Manufacturing pluripotent stem cells).

Another iPSC-derived immune cell subset that is gaining momentum are NK cells or natural killer cells, which comprise an important part of the innate immune response. NK cells are responsible for immune surveillance by targeting viral-infected cells and cancer cells that downregulate HLA I presentation and upregulate stress ligands. NK cells are interesting for CAR therapy since they do not have T cell receptors and cannot interact with HLA-I complex, which is the critical contributor for to graft-vs-host disease. Unlike iPSC-derived CAR T cells which only kill tumor cells that express the specific antigen that the CAR recognizes, iPSC-derived CAR NK cells can kill both tumor cells that express the CAR antigen as well as tumor cells that do not express the CAR antigen (Figure 4).

The use of CAR-engineered NK cells can overcome some of the graft vs host rejection issues using allogeneic CAR T cells, and bypass patient complications resulting from cytokine storms.
Figure 4. CAR NK cells can kill a broader range of tumor targets. On the left, CAR NK cells kill tumor cells that are missing HLA I expression without the engagement of CAR. On the right, CAR NK cell-mediated killing is shown to be dependent on the engagement of CAR and the tumor antigen (TA) expressed on the tumor cells.

Another attractive feature of using iPSC-derived CAR NK cells is their cytokine profile; upon activation they secrete a restricted level of IFN-γ, IL-12, and GM-CSF. CAR T cells secrete IL-1 and IL-6 continuously upon activation, and the presence of IL-1 and IL-6 can lead to cytokine-release-syndrome (CRS), a serious adverse event seen in some patients receiving CAR T therapies [26]. To date, adoptive transfer of iPSC-derived NK cells is well tolerated, and neither GvHD nor cytokine toxicity have been induced in patients [27,28].


There are a variety of gene-editing tools and delivery systems available to facilitate the genetic changes necessary for T cells to be used in CAR T therapies. Each system has its own advantages and disadvantages, which should be carefully weighed for each CAR T cell therapy. Tradeoffs such as target specificity versus ease of use or cost versus speed will need to be made. As with many cutting-edge technologies, the tools available are rapidly being improved, and new approaches are appearing to address the current limitations of the field.


  1. Depil S, et al. (2020) “Off-the-shelf”allogeneic CAR T cells: development and challenges. Nat Rev Drug Discov 19:185–199.
  2. Townsend MH, et al. (2020) Paving the way towards universal treatment with allogenic T cells. Immunol Res 68: 63–70.
  3. Zhao J, et al. (2018) Universal CARs, universal T cells, and universal CAR T cells. J Hematol Oncol 11:1–9.
  4. M. Ruella and S. S. Kenderian (2017) Next-generation chimeric antigen receptor T-cell therapy: going off the shelf. BioDrugs 31:473–481.
  5. D. Carroll (2011) Genome engineering with zinc-finger nucleases. Genetics 188:773–782.
  6. Richter A, et al (2016) TAL effector DNA-binding principles and specificity. Methods Mol Biol 1338: 9–25.
  7. Jinek M,et al. (2012) A programmable dual-RNA–guided DNA endonuclease in adaptive bacterial immunity. Science 337:816–821.
  8. Torikai H, et al. (2012) A foundation for universal T-cell based immunotherapy: T cells engineered to express a CD19-specific chimeric-antigen-receptor and eliminate expression of endogenous TCR. Blood 119(24):5697–5705.
  9. Provasi E, et al. (2012) Editing T cell specificity towards leukemia by zinc finger nucleases and lentiviral gene transfer. Nat Med 18(5):807–815.
  10. Torikai H, et al. (2013) Toward eliminating HLA class I expression to generate universal cells from allogeneic donors. Blood 122(8):1341-1349.
  11. Poirot L, et al. (2015) Multiplex Genome-Edited T-cell Manufacturing Platform for "Off-the-Shelf" Adoptive T-cell Immunotherapies. Cancer Res 75:3853–3864.
  12. Knipping F, et al. (2017) Genome-wide Specificity of Highly Efficient TALENs and CRISPR/Cas9 for T Cell Receptor Modification. Mol Ther Methods Clin Dev 4:213–224.
  13. Osborn M, et al. (2016) Evaluation of TCR Gene Editing Achieved by TALENs, CRISPR/Cas9, and megaTAL Nucleases. Mol Ther 24(3):570–581.
  14. Qasim W, et al. (2017) Molecular remission of infant B-ALL after infusion of universal TALEN gene-edited CAR T cells. Sci Transl Med 9(374):eaaj2013.
  15. Eyquem J, et al. (2017) Targeting a CAR to the TRAC locus with CRISPR/Cas9 enhances tumour rejection. Nature 543(7643):113–117.
  16. Ren J, et al. (2017) Multiplex Genome Editing to Generate Universal CAR T Cells Resistant to PD1 Inhibition. Clin Cancer Res 23(9):2255-2266.
  17. Georgiadis C, et al. (2018) Long Terminal Repeat CRISPR-CAR-Coupled "Universal" T Cells Mediate Potent Anti-leukemic Effects. Mol Ther 26(5):1215–1227.
  18. Ren J, et al. (2017) A versatile system for rapid multiplex genome-edited CAR T cell generation. Oncotarget 8(10):17002–17011.
  19. Ghaleh HEG, et al. (2020) Concise review on optimized methods in production and transduction of lentiviral vectors in order to facilitate immunotherapy and gene therapy. Biomed Pharmacother 128:110276.
  20. Bulcha J, et al (2021) Viral vector platforms within the gene therapy landscape. Signal Transduct Target Ther
  21. Cockrell AS and T Kafri T (2007) Gene delivery by lentivirus vectors. Mol Biotechnol 36(3):184–204.
  22. Shalem O. (2014) Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343(6166):84–87.
  23. Cogle CR, et al. (2003) An overview of stem cell research and regulatory issues. Mayo Clin Proc 78(8):993–1003.
  24. Shi Y, et al. (2017) Induced pluripotent stem cell technology: a decade of progress. Nat Rev Drug Discov
  25. Angelos MG and Kaufman DS (2015) Pluripotent stem cell applications for regenerative medicine. Curr Opin Organ Transplant 20(6):663–70.
  26. Liu E, et al. (2020) Use of CAR-Transduced Natural Killer Cells in CD19-Positive Lymphoid Tumors. N Engl J Med 382(6):545–553.
  27. Dolstra H, et al. (2017) Successful Transfer of Umbilical Cord Blood CD34 + Hematopoietic Stem and Progenitor-derived NK Cells in Older Acute Myeloid Leukemia Patients. Clin Cancer Res 23(15):4107–4118.
  28. JS Miller JS and Lanier LL (2019) Natural Killer Cells in Cancer Immunotherapy. Annu Rev Cancer Biol 3:77–103.