CRISPR-Mediated Non-Viral Site-Specific Gene Integration and Expression in T Cells: Protocol and Application for T-Cell Therapy

T cells engineered with chimeric antigen receptors (CARs) show great promise in the treatment of some cancers. Modifying T cells to express CARs generally relies on T-cell transduction using viral vectors carrying a transgene, resulting in semi-random DNA integration within the T-cell genome. While this approach has proven successful and is used in generating the Food and Drug Administration (FDA, USA) approved B-lymphocyte antigen CD19-specific CAR T cells, it is possible the transgene could integrate into a locus that would lead to malignant transformation of the engineered T cells. In addition, manufacturing viral vectors is time-consuming and expensive. One way to overcome these challenges is site-specific gene integration, which can be achieved through clustered regularly interspaced short palindromic repeat (CRISPR) mediated editing and non-viral DNA, which serves as a template for homology-directed repair (HDR). This non-viral gene editing approach provides a rapid, highly specific, and inexpensive way to engineer T cells. Here, we describe an optimized protocol for the site-specific knock-in of a large transgene in primary human T cells using non-viral double stranded DNA as a repair template. As proof-of-principle, we targeted the T-cell receptor alpha constant (TRAC) locus for insertion of a large transgene containing green fluorescence protein (GFP) and interleukin-15 (IL-15). To optimize the knock-in conditions we tested template DNA concentration, homology arm length, cell number, and knock-in efficiency over time. We then applied these established guidelines to target the TRAC or interleukin-13 (IL-13) locus for the knock-in of synthetic molecules, such as a CAR, bispecific T-cell engager (BiTE), and other transgenes. While integration efficiency depends on the targeted gene locus and selected transgene, this optimized protocol reliably generates the desired insertion at rates upwards of 20%. Thus, it should serve as a good starting point for investigators who are interested in knocking in transgenes into specific loci.


Donor DNA Design
Suggestions for donor DNA design are illustrated in Figure 2. It is recommended to design the homology arms right next to the cut site of the sgRNA of the targeted gene locus. Based on our protocol optimization we recommend using homology arms of 400 bp for longer insert sizes. A 2A or IRES sequence should be implemented in front of the transgene to ensure proper separation of the product from the native gene product. When two genes are cloned together in one construct, it is recommended to add a 2A sequence in between those genes to avoid fusion genes during translation. At the end of the transgene, before the 5′ end of the right homology arm, a poly A sequence is beneficial for appropriate gene termination. Lastly, mutating the PAM sequence in the construct inhibits the Cas9 enzyme from repeatedly cutting the DNA in this location.

PBMC Preparation
For our studies we isolated PBMCs by Lymphoprep (Abbott Laboratories, Chicago, IL, USA) gradient centrifugation. Generally, any PBMC isolation method is appropriate for this protocol as long as it produces healthy and viable PBMCs. While we used cryopreserved PBMCs, using fresh PBMCs would result in a higher viability of T cells.

T Cell Enrichment
This part of the protocol is a modified version of the MidiMACS kit protocol.
1. Thaw a vial of PBMCs (prepared as described above) in a 37 C water bath. Wash the thawed PBMCs with T cell media (RPMI-1640 media including 10% FBS and 1% GlutaMAX-I). Resuspend the cell-pellet with MACS buffer (500 mL PBS, 25 mL MACS BSA stock solution, and 2 mL 0.5 M EDTA pH 8.0). 2. Count the cells and spin down at 400g for 5 minutes at room temperature. Aspirate the MACS buffer and add 250 L of MACS buffer to the cell pellet. 3. Add 20 L of CD4 and 20 L of CD8 MicroBeads per 10 7 cells. Ensure proper mixing and incubate for 15 minutes at 4 C. 4. After incubating, wash the cells with 2 mL of MACS Buffer and spin down at 400 g for 5 minutes at room temperatures. During the wash step, place the MidiMACS Separator on the MACS MultiStand and a MACS LS column in the MidiMACS Separator and equilibrate the column with 3 mL MACS buffer. 5. After washing, aspirate the supernatant and resuspend the stained cells with 500 L of MACS Buffer. Run the cell suspension through a 40 M cell strainer to remove dead cell clumps and apply it onto the column. 6. Wash the column with 3 mL of MACS buffer. Repeat this step for a total of 3 washes. 7. Remove the column from the MidiMACS Separator and place on a 15 mL collection tube.
Add 5 mL of MACS buffer and insert the syringe plunger (comes with the MACS LS column) to flush the column and release the cells into the collection tube. Spin at 400 g for 5 minutes at room temperature. 8. Add 5 mL of T cell media and count the cells. Plate 10 6 selected T cells in 2 mL T cell media in one well of a 24 well tissue culture treated plate. 9. Rest the plated T cells overnight at 37 °C and 5% CO2.

T Cell Activation
Here we tested 3 approaches to activate T cells: plate bound CD3/CD28, Dynabeads and Transact. While we did not observe any differences between Dynabeads and Transact, plate bound CD3/CD28 resulted in lower knock-in efficiencies as well as T cell viability. Thus, below we describe two methods for T cell activation. Dynabeads 1. The next day, transfer Dynabeads Human T-Activator CD3/CD28 (25 L per well of plated T cells) to a 1.5 mL collection tube. 2. Add 1mL MACS buffer to the tube with Dynabeads and place the tube on the back side of a MiniMACS Separator. With the tube pressed against the magnet, remove the MACS buffer. 3. Re-suspend the beads in T cell media in the same volume as the volume of Dynabeads in step 1. 4. Add 25 L Dynabeads to each well of plated T cells. Add IL-7 at 10 ng/mL and IL-15 at 5 ng/mL. 5. Rest the activated T cells for two days at 37 °C and 5% CO2 before electroporation.

TransAct Media
1. The next day, add 28.5 L T cell TransAct to each 24 well tissue plate with T cells. Add IL-7 at 10 ng/mL and IL-15 at 5 ng/mL. 2. Rest the activated T cells for two days at 37 °C and 5% CO2 before electroporation. Image the gel and cut out bands with the appropriate band size. Perform gel extraction with NucleoSpin Gel and PCR Cleanup kit according manufacturer's protocol. Elute all gel pieces in 2 × 30 L nuclease-free water total (use 2 columns and one collection tube).

HDR Template Preparation
Concentrate DNA further with Agencourt AMPure XP beads according to manufacturer's protocol. Elute the final solution from step 5.7. in 10 L nuclease-free water (recommended final concentration 1-2 g/L).
Measure the final DNA concentration on NanoDrop.

Electroporation
Prepare RNP complexes using a 4.5:1 sgRNA:Cas9 molar ratio (carried out in a RNA-free environment): sgRNA working stock preparation: spin down sgRNA tube and add 10 L 1× TE to make a 150 M stock. Dilute in 15 L RNase-free water to generate a 60 M working stock solution. Transfer 3 L TRAC exon 1 sgRNA from the 60 M stock to a PCR tube.
Add 1 L cas9 from 40 M stock to PCR tube containing the sgRNA and incubate for 10 minutes at room temperature (store at −20 C until use).
Set up recovery plates for the T cells after electroporation (Note: 2 electroporation reactions of 1 × 10 6 cells each combined in one recovery well is recommended for optimal viability and knock-in): 48 well tissue culture treated plate including 550 L recovery media (RPMI-1640 media including 20% FBS, 1% GlutaMAX-I, IL-7 at 10 ng/mL and IL-15 at 5 ng/mL). The total final volume per well should be 750 L. (Note: if knock-out efficiency is low, use 0.6 × 10 6 cells per electroporation reaction).
Prepare P3 Primary Cell Nucleofector Solution (17 L total volume per electroporation reaction): 13.94 L P3 Primary Cell Nucleofector Solution and 3.06 L Supplement 1.
Collect the T cells in a 1.5 mL collection tube and remove the Dynabeads Human T-Activator CD3/CD28 from T cells by pressing the tube against a MiniMACS Separator. Take out the cell solution without the Dynabeads. (Note: if T cells were activated with T cell TransAct, transfer the activated T cells to a collection tube).
Count the T cells and take 10 6 cells per electroporation reaction (do not spin cells prior to counting). Spin down the T cells at 200 g for 10 minutes at room temperature.
Remove all the media from the cell pellet and re-suspend the cell pellet in P3 Primary Cell Nucleofector Solution (1 × 10 6 T cells per 17 L P3 Primary Cell Nucleofector Solution) Add 2 g of HDR template (obtained in step 4.9) in no more than 3 L nuclease-free water together with 4 µl of RNP (obtained in step 5.1) in a new collection tube. Incubate for 10 minutes at room temperature. (Note: always prepare the following controls: No HDR template + RNP; no HDR template, no RNP; and HDR template, no RNP).
Add 17 µl of the T cells in P3 Primary Cell Nucleofector Solution to the tube with HDR template and RNP (and to the control tubes).
Transfer 23 µl of T cells with HDR template and RNP (and of the controls) to one well of the Nucleocuvette Strip and nucleofect the cells using program: EH-115.
After nucleofection, add 80 µl of recovery media to the T cells in the Nucleocuvette Strip (as per manufacturer's suggestions, do not add the media directly into the well but slowly pipet at the rim of the well).
Let T cells to recover at 37 °C and 5% CO2 for 30 minutes. Then transfer the cells to the recovery plate prepared in step 5.2 (2 electroporation reactions into 1 well of the recovery plate).

T Cell Expansion
Incubate the electroporated T cells at 37  C and 5% CO2. Split the cells when the media yellows and the cells are at high density (this can take a few days as the cell will grow slowly right after electroporation).

Determine Knock-in
Determine knock-in efficiency 8-10 days post-electroporation. Detection method is based on the transgene. Detection should be done at a protein level using flow cytometry, ELISA or WB to confirm that the protein is expressed. In addition, targeted next-generation sequencing (NGS) is highly recommended to determine editing efficiency.